E - D O K U M E N T I S J M Three Roads to Comparative Research: Analytical, Visual, and Morphological Edited by Niko Toš | Karl H. Müller | Zoltán Fábián Jindřich Krejčí | Marcel Zielinski Three Roads to Comparative Research: Analytical, Visual, and Morphological Edited by Niko Toš | Karl H. Müller | Zoltán Fábián Jindřich Krejčí | Marcel Zielinski E-DOKUMENTI SJM/Ljubljana, 2022 edition echoraum | Vienna Niko Toš/Karl H. Müller/Zoltán Fábián Jindřich Krejčí/Marcel Zielinski (edited) Three Roads to Comparative Research: Analytical, Visual and Morphological First electronic edition Published by: Univerza v Ljubljani, Fakulteta za družbene vede, IDV, CJMMK Kardeljeva ploščad 5, Ljubljana Ljubljana, 2022 Book Series: E-Dokumenti SJM 9 Editor: Slavko Kurdija Layout: Werner Korn Text edition: Gertrud Hafner Figures and Graphs: Michael Eigner Proofreading: Hannah Elmer This work is licensed under a Creative Commons AttributionShareAlike 4.0 International License. URL: https://knjigarna.fdv.si/ in www.cjm.si DOI: 10.51936/9789612950170 Kataložni zapis o publikaciji (CIP) pripravili v Narodni in univerzitetni knjižnici v Ljubljani COBISS.SI-ID 102277635 ISBN 978-961-295-017-0 (PDF) Printed edition was published by Edition Echoraum, Wien, 2009 ISBN 978-3-901941-17-7 Table of Contents Preface 7 Rudolf Hundstorfer | Federal Minister of Labour, Social Affairs and Consumer Protection, Austria Ivan Svetlik | Minister of Labour, Family and Social Affairs, Slovenia Acknowledgements 11 Abstracts 13 General Introduction 25 Karl H. Müller | Niko Toš Part I – The Analytical Road to Comparative Research 27 Section I Working Conditions, Household Work, Social Activities and Migration 31 1 Patterns of Work Flexibility in Typical Jobs in Eastern and Western Europe 33 Claire Wallace | Endre Sik | Bori Simonovits 2 Division of Housework in European Societies through the Analysis of Macro Level Factors 55 Christian Bischof 3 Attitudes towards Immigrants and the Effect of Migration in European Countries 75 Jana Chaloupková | Petra Šalamounová 4 Cure for Loneliness? Social Activity and Its Role in Changing European Societies 99 Anna Domaradzka Section II Family, Gender, Life-Satisfaction 127 5 Life Satisfaction, Happiness and Marital Status in Four Central European Countries: International Comparison Based on the European Social Survey 129 Dana Hamplová 6 Women’s Role in the Family: A Comparative Study 143 Filip Raciborski 7 Children – To Have or Not to Have. A Question in the Context of Cross–National Differences 171 Marcin W. Zieliński Section III Citizenship, Institutions, Political Participation 193 8 An Outline for a New Concept of Citizenship in a Trans-National Context 195 Mitja Hafner-Fink | Samo Uhan 9 Trust in Institutions of the System – European Comparisons – and Slovenia 221 Niko Toš | Vlado Miheljak | Slavko Kurdija 10 Types of Non-Electoral Political Participation in Europe 269 Klára Vlachová | Tomáš Lebeda Part II – The Visual Road to Comparative Research 293 The Online-Program WISDOMIZE in a Nutshell 297 Armin Reautschnig 11 Static Patterns 301 12 Dynamic Patterns 375 13 Complex Patterns 407 Part III – The Morphological Road to Comparative Research 419 14 The Third Road for Comparative Research: Morphological Designs 423 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš 15 Towards New Frontiers in Comparative Survey Research 459 Karl H. Müller | Niko Toš Bibliography 507 Authors 527 Index 533 Preface Rudolf Hundstorfer | Federal Minister of Labour, Social Affairs and Consumer Protection, Austria The processes of European integration and co-operation occur at multiple levels and in multiple dimensions and create new opportunities and new challenges for practically all relevant segments of the Austrian society. One of the biggest challenges of these complex integration processes lies in an adequate observation and monitoring which must be performed, more and more, in an internationally comparable manner. For political decision making, the results of these observation and monitoring processes are urgently needed since an increasing number of decisions at the political level has to take into account that these decisions do not only affect the Austrian side alone, but have also important implications and ramifications for neighbouring countries and for the processes of European integration as a whole. For this reason, the Austrian Ministry of Labour, Social Affairs and Consumer Protection has initiated a number of programs which, on the one hand, create new infrastructures and platforms for bilateral information exchange and which, on the other hand, make significant contributions for improving the observation and monitoring processes of integration and co-operation in Europe. For example, the so-called expert academies with Austria’s neighbouring countries have become a necessary and widely used platform for information exchanges and discussions of best practice-models in the field of labour market policies. Likewise, the program EECO-LAB (Eastern European Co-operation on Labour) has developed into a dense information platform which has improved our understanding of the subjective working and living conditions in Central and Eastern Europe. Furthermore, EECO-LAB is responsible for the Austrian data production of the European Social Survey (ESS) which has become the most important data source for high quality comparative micro-data on living, working and social conditions across Europe. This new volume presents a timely summary of different ways and designs for comparative research on Central and Eastern Europe. In its first part, it gives comparative summaries on a wide variety of topics, ranging from different patterns of work flexibility to the role of women in different European societies. The second and the third parts have been developed over the course of the EECO-LAB projects themselves and point to new ways for comparative research, one with the help of new visualization tools and the 8 Preface other with new forms of analyses which are able to utilize large quantities of currently unused but available data sets. It is our hope that this volume marks an important contribution for significantly improving the observation and monitoring processes of European integration and co-operation. Likewise, this volume shows that the Austrian Ministry of Labour, Social Affairs and Consumer Protection plays and will continue to play an active and innovative part in enabling these very much needed comprehensive observation and monitoring processes. Preface Ivan Svetlik | Minister of Labour, Family and Social Affairs, Slovenia In the course of the next two years the European Union with its 27 member states will open its borders for free and unrestricted movement of employees across all EU-countries. The new era of a European labour market will generate a new wave of challenges for nationally oriented labour market policies, for the national institutional arrangements of social partnership and cooperatist regimes and for employment and labour market policies at the European level. Somewhat surprisingly, the new era of European labour markets implies dramatic changes for the social sciences across Europe too, which, so far, have followed national agendas as well. For European labour market policies, the primary challenges lie in an adequate coping with the multiplicity of levels and in intelligent designs for European and national activities. More specifically, the next years and decades will bring the formation of European employment programs, the implementation of common European minimal standards, the harmonization of unemployment systems and a quick transfer of national best practice policies in the field of labour market policies to other European areas. Coming from the social sciences myself, it is quite remarkable that only in recent decades the social sciences have faced and taken up the dual challenges of becoming European and, equally important, becoming global at the same time. The global challenges are manifold and highly complex and will not be discussed further in the present context. But the European challenges turn out to be highly complex and diversified, too. One of these challenges lies, for example, in the current information systems for employment and professional qualifications which, so far, were clearly restricted by their national boundaries and which were developed for decades and even centuries within national confines. Harmonization across Europe turns out to be one of the deep social science problems, too, because it leads to a re-thinking of what constitutes professions, qualifications and what can be used as a frame of reference for comparisons. Likewise, comparative research on the different levels of regional, national and supranational trajectories of employment systems and of labour markets becomes another highly intriguing issue which will require a large number of new approaches as well. Seen from this perspective I highly welcome the edition of this new book because it shows some of the traditional and some of the new ways how to deal with the process of Europeanization in the area of labour markets 10 Preface and employment systems. Moreover, the three roads to comparative research, which were used in this volume, offer a variety of comparative results and patterns for Central and Eastern Europe. I hope that this book contributes to an improved understanding of the complexities of the European project and to a widening of our current approaches in comparative research. Finally, amidst the current turmoil in the financial markets and in industrial production, I hope that this volume will be used not only by social scientists but by policy makers as well in order to keep them aware of the difficult paths and of the enormous challenges ahead. Acknowledgements Working on comparative research across national boundaries and across different languages requires a coordinated effort by a large number of persons. In our case, this co-operation included several data archives in Central Europe which were linked with the help of a research program under the name of EECO-LAB (Eastern European Co-operation on Labour), financed by the Federal Ministry of Labour, Social Affairs and Consumer Protection (BMASK). Thus, thanks go to – Gertrud Hafner in Vienna who was confronted with the difficult tasks of transforming a very heterogeneous manuscript into a homogeneus format and into the new publishing program of the book series – Ivi Kecman in Ljubljana who served as a vital interface between Ljubljana and Vienna – Armin Reautschnig who worked very hard on the visualization tool WISDOMIZE which is at the core of Part II of this book – Christian Bischof with whom we explored the potentials inherent in the third road to comparative research – Michael Eigner who was mainly responsible for the design and the redesign of the diagrams, figures and graphs in the book – Stefan Potmesil, Richard Fuchsbichler, Roland Hanak and Susanne Schlögl from the Federal Ministry of Labour, Social Affairs and Consumer Protection (BMASK) who provide a stable support for EECO-LAB – Werner Korn who acted as an unmoved prime mover behind this book project and behind the entire series on ”Complexity, Design, Society” which has reached already its eighth volume – a remarkably good spirit of stable cooperation and friendship between the editors which has overcome many obstacles and barriers and which will continue to last well-beyond the publication of this book. It should be emphasized that the present book in its final design fits very well into the overall context of the book series with its emphasis on complexity research or on new research designs, new methodologies or, as an essential element, on new information designs. It is hoped that the two rather unconventional ways for comparative research, namely data visualization as an exploratory method and the morphological approach as a new road on the basis of heterogeneous data sets, will meet the interest of social scientists across Europe who search for novel ways and methods in comparative analysis. In a variation of Peter Berger’s “Invitation to Sociology” our invitation to comparative research can be phrased in the following way: 12 Comparative researchers who like to avoid shocking discoveries, who prefer to believe that comparative research is just what they were taught in standard textbooks, who like the safety of the rules and the maxims of what Alfred Schütz has called the ‘world taken for granted’, should stay away from the ‘Three Roads to Comparative Research’. We sincerely hope that the three roads to comparative research, presented in this volume, enable researchers in the field to widen their current tool-box significantly and to open up new ways for comparative research with exciting and innovative results. Vienna, July 2009 Niko Toš | Karl H. Müller | Zoltán Fábián Jindřich Krejčí | Marcel Zielinski Abstracts Patterns of Work Flexibility in Typical Jobs in Eastern and Western Europe Endre Sik | Claire Wallace | Bori Simonovits Flexibility has long been a debate in Western Europe, and this debate is now also being introduced to the New Member States of the European Union. In this study, we use a representative sample survey carried out in 2001 (N=11194) of the following countries: the UK, the Netherlands, Sweden, Slovenia, Hungary, the Czech Republic, Bulgaria and Romania. We consider how flexible the workforce is by looking at the extent of time, place, contract and income flexibility, and the ways in which these are combined. The paper goes on to examine the social characteristics of flexible workers in different countries. We find that time and contract flexibility are associated with women in the Western EU countries, but that these types of flexibility are more often linked to men in the Eastern European countries. The nature of flexibility is related to the regime of regulation in a country, with strongly regulated countries tending to have flexibility concentrated among a narrow range of workers who are young and less educated, while less regulated flexibility leads to more of the population being flexible. The feminization of flexibility has to be seen in the context of the gendered forms of family and work, and it is not uniform across Europe. Division of Housework in European Societies through the Analysis of Macro Level Factors Christian Bischof The purpose of this paper is to identify the factors that can help explain the gender disparity of the division of housework, a majority of the domestic work is still done by women. This is done through an international comparison using the data for 22 European countries. Factors from both the individual level and the macro level (in the form of context variables for the countries) will be incorporated in the analysis. The explanation for the strongly differing share of housework for women might thus be found primarily on the individual level. The great differences on the national level emerged only through individual characteristics and not through macro factors. Naturally, 14 Abstracts this still does not mean that the respondent exhibits this behavior simply from him or herself, that is, independent from the social context to which he or she is connected on many levels. Attitudes Towards Immigrants and the Effect of Migration in European Countries1 Jana Chaloupková | Petra Šalamounová This article focuses on a comparison of attitudes on migration, based on data from the European Social Survey 2002. It looks at the question of whether attitudes towards immigrants are related to the numbers and structure of immigrants within a country. For these purposes, the past and the current state of migration in Europe are briefly outlined. Three thematic areas of attitudes towards migration are examined: 1) the host population’s willingness to accept immigrants; 2) perceptions of the impact of immigration on the host country; 3) attitudes towards different forms of integration of immigrants. The findings indicate that Europeans are more willing to accept migrants from of the same race and from Europe than migrants of different races and from states outside Europe. The greatest unwillingness to accept people from other states and the strongest emphasis on the negative impact of immigration was observed in Greece and Hungary, while the greatest willingness to accept immigrants was found in Sweden and Switzerland, which was also connected to a more positive perception regarding the impact of immigration. Cure for Loneliness? Social Activity and its Role in Changing European Societies Anna Domaradzka The starting point of the paper is Ulrich Beck’s thesis that nowadays close family relations are disappearing and individuals are left alone with their problems and emotional needs. It can be stated that a new pattern of social networks is emerging; less connected with family and more with various self-development groups, voluntary associations and other forms of social activity. It especially concerns elderly people, who are no longer active in the 1 The research for this paper was supported by a grant from the Czech Science Foundation (No. 403/04/1219) as a part of the European Social Survey II project and by a grant from the Ministry of Education, Youth and Sports (No. 1N04192). Abstracts 15 labour market, but still are able to lead an active life and are often affected by this “civilization disease”, which is loneliness. The question is whether loneliness is indeed growing among surveyed countries and if yes, then is it a general change or are there specific groups affected by this process? The second issue is the substitute for close family relations. If the family ties are weakened, are new forms of close social relation established instead of traditional family networks? One of the main focuses of the paper is the subpopulation of elderly people. It is often stated that nowadays more and more of them suffer loneliness. What is the cure? Or is there none? Do lonely elders close themselves at homes and withdraw from social life? Or, to the contrary, do they try to establish new forms of activity and substitutes for vanishing family networks. The analysis will be based on the International Social Survey Program data derived from “Social Networks” and “Family and Changing Gender Roles” modules as well as WVS data. Among the researched countries are those with high as well as low social participation level, which should enrich international comparisons. The research question will be investigated using multivariate analysis. The goal of the analysis is to test the hypothesis of a weakening of family relations and establishing new forms of social networks instead. It seems probable that a high level of social activity is correlated with disintegration of family bonds, but it may also be that inclination for social activity is a form of extended family activity. That is why the difference between countries with high and low levels of social activity and participation will be closely examined. Life Satisfaction, Happiness and Marital Status in 4 Central European Countries: International Comparison Based on the European Social Survey2 Dana Hamplová The text will focus on the relationship between life satisfaction and marital status in selected European countries. While up to date research indicates a positive effect of marriage on life satisfaction, it is less clear whether this effect is mediated by the level of divorce within the society. Theoretically, divorce could increase or decrease life satisfaction of married couples, depending on 2 The research for this paper was supported by a grant from the Czech Science Foundation (No. 403/04/1219) as a part of the European Social Survey II project and by a grant from the Ministry of Education, Youth and Sports (No. 1N04192). 16 Abstracts whether the selection effect or the effect of investment in the relationship is more important. The research has been conducted using OLS regression and multilevel modeling. Preliminary results show that married people are happier than singles, even though the strength of this effect varies. Analysis does not indicate that the effect of the marital status depends on the divorce rate in the society. The effect of cohabitation is much less clear: it resembles marriage in some countries and is more similar to a single lifestyle in others. Women’s Role in the Family: A Comparative Study Filip Raciborski In the contemporary world we constantly experience changes. These changes affect all spheres of our lives. Some of them happen suddenly, while others occur so slowly that we hardly notice them. The changes in the family that took place at the turn of the 20th century are undeniable. I would especially like to emphasize the shift from the extended family model to the nuclear family model, a process that is still continuing. Moreover, it is quite possible that in the 21st century, the very existence of the family as the basic social unit will be threatened by the new social phenomenon of the so-called “singles”. In my research I would like to focus on the effects of the changes regarding attitudes of women’s roles in the family. I am especially interested in the issue of combining workplace duties with family responsibilities in selected EU countries. I want to verify whether views on the roles of women in the family are reflected in everyday life. The analysis will be based on the results of the International Social Survey (2002) for selected countries. The research conducted in 2002 addresses precisely the changes in family life and the division of roles. Through this research, I intend to examine not only Poland, the Czech Republic, Slovakia, Hungary and Austria, but also Spain, Ireland, France and Great Britain. Ireland and Spain lend themselves especially well to a comparative study with Poland because of essential similarities between them (fast economic growth, the important role of Catholicism). In contrast, France and Great Britain exemplify multi-cultural countries in which many cultures come together. France, moreover, is considered to be the country in which postmodern values are the most widespread. I am also interested to see if this accompanies a different outlook on women’s role. Apart from the assessment of the present situation, I will refer back to the mid-1990s, when the earlier International Social Survey was conducted. Unfortunately, in 1994 far fewer countries were included in the survey, so the Abstracts 17 comparison will not be complete. Nonetheless, I hope that the analysis of the discussed material will make it possible for me to answer the questions concerning the scale and speed of the changes that we currently observe. Children – To Have Or Not To Have. A Question in the Context of Cross–National Differences Marcin W. Zieliński Childlessness is an important demographic problem for European countries. It may result in difficulties for the social security systems and threaten social needs satisfaction for the next generations of Europeans unless immigration policies are changed. Low birth rates in most European countries have also begun to affect the new EU members. This article focuses on the changes in the perception of the importance of children, treating it as a mediator in the decision to have babies. Moreover, it identifies the factors that are responsible for the differences across selected Central European countries. The analysis is based on the International Social Survey Program data (ISSP) derived from the “Family and Changing Gender Roles” module. It shows the main differences between selected Central European and Scandinavian countries regarding the agreement with two opposing statements that address the importance of children. To investigate the relationships of selected socioeconomic factors and the differences between countries, a multilevel regression analysis is used. The main goal is to explain these differences on the two levels of observation – the respondent (Level 1) and the country (Level 2). An Outline for a New Concept of Citizenship in a Trans-National Context Mitja Hafner-Fink | Samo Uhan While the modern nation-state (along with the concepts of sovereignty, nationhood, nation membership and national identity) is understood as both the foundations and framework for the modern conception of citizenship or ‘national citizenship’, today’s processes of globalisation have led to a shift to the new concept of so-called ‘postnational citizenship’ which ‘ignores’ the nation-state’s borders ( e.g. Soysal 1996). In the first part of the chapter the traditional concept of citizenship is briefly presented through a discussion of the ‘conflict’ between the liberal and republican models of citizenship. 18 Abstracts A possibility to resolve this ‘conflict’ by introducing a new concept of ‘post-modern’ citizenship is indicated. In the second part the results of (comparative) analyses of survey data from the International Social Survey Programme (ISSP) module 2004 ‘Citizenship’ and from the European Social Survey – Round 1 (ESS 2002) are presented. In general, the analysis reveals the following pattern: in highly developed countries with longer democratic tradition (particularly Scandinavian countries), we can find more elements of post-modern citizenship. Empirical analysis prepared additional elements for discussion in the third part, where authors propose a possible direction for modelling a new (‘post-modern’) paradigm of citizenship. They share the view of Joe Painter (Painter, 2002), who sees the post-modern paradigm of citizenship as a kind of synthesis of classical conceptions of citizenship, where the main characteristic of this synthesis should be the idea of multilevel citizenship.. Trust in Institutions of the System – European Comparisons – and Slovenia Niko Toš | Vlado Miheljak | Slavko Kurdija Trust in institutions of political and social system is a basis of modern democracies. The article is based on research of trust in state and civil society institutions, that was carried out within the program Slovenian public opinion in the period from 1991–2006. Higher levels of trust in state and lower civil society institutions in the beginning (1991) were followed by a decline and stagnation of trust in political institution on a low level in the first decade (1991–1999) – and at the same time a rise of trust in civil society institutions. These trends characterize the course of democratic institutionalization in Slovenia in this period. Short-term change of rulers and elections in 2000 are characterized by an increase of level of trust in institutions to the upper limit, followed by another decrease in political institutions – in spite of changes in power positions in 2004 – on the lower level in comparison to 1999. The analysis confirms that expression of trust in political institutions is under the strong influence of political views of the evaluators. The authors call attention to Slovenians’ low level of trust in political institutions in the recent period, which limits participation of Slovenians in the democratic processes and the success of their control over legislative and executive power. In the last part the analysis is based on the results of the European Social Survey. The observation of attitudes towards institutions was limited to six or seven institutions, in three consecutive surveys. Differences in the Abstracts 19 expression of trust between countries are quite significant. The clustering method demonstrated the relationship between the ranking of countries and their development of democratic and economic conditions. Two large groups of countries emerged. In one group are the developed European countries with a developed democratic nature, whilst in the other are those which are currently on the way towards it. Types of Non-Electoral Political Participation in Europe3 Klára Vlachová | Tomáš Lebeda In this article, the authors describe the forms of political participation outside the electoral process that the populations of twenty-one European countries tend to employ and the degree to which this is done. The article begins with a general discussion of the types of political participation (conventional or unconventional) and their changes (for example, some forms conventional participation in contemporary democracies have become unconventional and vice versa, and new forms of participation have been identified in democracies). Differences between political participation in post-communist democracies and “old” democracies are discussed as well. The core of the article is based on analyses of the data from the European Social Survey, Round I. Using factor analysis, the authors identify three types of non-electoral political participation in the societies under consideration: active-conventional, active-protest, and passive participation. The overall non-electoral political participation is considerably lower in the post-communist and Mediterranean countries than in the Western European and Scandinavian countries. In the latter countries, the passive type of political participation is clearly a much stronger form of participation than the other two types. Conversely, in the Mediterranean countries, passive participation is weaker and is exceeded in places by the active-conventional type of participation. The Mediterranean area is also notable for the unusually strong presence of the active-protest type of participation. The authors try to explain the differences in non-electoral political participation between the countries by examining micro and macro variables. 3 This paper was written as a part of the European Social Survey II project supported by a grant from the Czech Science Foundation (No. 403/04/1219). 20 Abstracts The Visual Road: Static Patterns, Dynamic Patterns and Complex Patterns Armin Reautschnig The second road for comparative research covers a set of research paths which, so far, were not considered as scientific research trajectories altogether, but as a necessary side-step within the first road. Already along the first road for comparative research diagrams were widely used to display the results of statistical analyses, but they served mainly the purpose of representing comparative data. In sharp contrast to a mere auxiliary instrument for displaying statistical information, the second road for comparative research builds on data visualization as an instrument and as an object for research. Thus, the second approach transforms a set of comparable data into visual patterns with which a competent observer interacts and, moreover, which serves as a heuristic basis for the formation of new hypotheses. Here, the part of statistical analyses is substituted with a complex interaction between a visual data pattern on the one hand and a competent social scientist on the other hand. It must be noted that the first and the second road to comparative research are not strictly independent from each other. In many instances it is useful to use the second road first in order to get quick insights into interesting or significant characteristics of a data set. Afterwards, it might be advisable to use statistical techniques in order to substantiate the results of visual interactions further. In the three chapters of Part II, different types of patterns are produced, namely static patterns (Chapter 11), dynamic patterns (Chapter 12) and complex patterns (Chapter 13). At the outset, a short introduction is presented on a new visualization program under the name of WISDOMIZE. In a nutshell, the second road for comparative research can best be utilized as an exploratory way of data analysis while the first road is the appropriate form for testing the statistical viability and significance of micro-social hypotheses or theories. The Third Road for Comparative Research: Morphological Designs Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš This article presents the outlines for a new road for comparative research which could become of special relevance for large quantities of unused data Abstracts 21 outside the current data bases for comparative analyses. These datasets have been produced and analyzed within a single region only and, so far, were of no use for comparative research due to the absence of functionally identical datasets in other regions or countries. Subsequently, these surveys which have been designed and assembled for the purpose of analyzing a single population only will be characterized as atomic survey-datasets. Essentially, atomic survey datasets are, due to the missing links to similar survey datasets in other areas, lost for comparative research. At this point, the morphological approach sets in, tries to create bridges between these different atomic survey datasets and attempts to bring them into the arena of comparative research. With a common class of data transformations which can be labelled as formation and as aggregation, the morphological approach generates a set of morphological constructs which, despite the heterogeneity of the underlying atomic survey datasets, can be used for comparative analyses. Towards the end of this article, several designs within the third road for comparative research will be described in greater detail. But the main thrust of this paper lies in a high-risk test of the morphological approach in order to establish its viability for comparative research. Towards New Frontiers in Comparative Survey Research Karl H. Müller | Niko Toš This chapter turns its focus on the widening horizons and on the emerging cognitive landscapes for new ways and directions in comparative survey research. The article starts with some general remarks on the increasing pressures for comparative survey research and will continue with a basic phase transition in the cognitive organization of science, namely the transition from Science I to Science II. The subsequent parts of this article deal with the repercussions and consequences of this phase transition for the social sciences in general and for comparative survey research in particular. Adapting and accommodating to this new background knowledge, comparative survey research should and will create new survey designs and will leave behind its established core routines and its standard procedures as special cases, very much as Newtonian physics has become a special terrain within contemporary physics. In its final parts, the article offers a distinction between under-learned and over-learned responses and evolutionary architectures as a model background which should contribute to a deeper understanding of the complex cognitive 22 Abstracts processes inherent in survey interactions. Moreover, the re-combination of comparative survey research with the cognitive neuro-sciences and with evolutionary learning models should lead to a fundamental change in the core approaches to identify attitudes, to determine subjective as well as objective living conditions or to explore individual life styles. The paths for comparative survey research of the 20th century and its trajectories in the 21st century will be situated in significantly different cognitive territories. Overview Mor 1 13 pholo 4 1 12 5 11 gy VisualizationAnalyses 8 5 1 9 6 2 10 7 3 4 General Introduction Karl H. Müller | Niko Toš The Three Roads to Comparative Research: Analytical, Visual, and Morphological This book had an unusual history of its own, because it started as a project on comparative analyses, using mostly data from the European Social Survey (ESS). However, in the course of the compilation of articles, this primary goal could not be attained sufficiently. The data base, underlying the articles, was not only ESS-based and the articles themselves covered rather heterogeneous fields and could not be compared with one another in a satisfactory manner. As a consequence, a new goal for the book was needed. Fortunately, it turned out that we at the Wiener Institute for Social Science Methodology and Development (WISDOM) were working on ESS-based data visualizations and on the methodology of survey research. So we changed the focus of the book into a general overview on comparative research and on three very different ways to achieve comparative results. The first road is widely established and well-known, the second and especially the third one are non-standard and in the process of becoming viable alternatives to the normal research trajectories for comparative research. The first road to comparative research is centered on statistical analyses and on comparative descriptions of the obtained results. This road is most widely used and can be characterized as the traditional way for comparative investigations on the basis of micro-social data. Moreover, the first road was the one that was planned for the book in its initial stage. The second road to comparative research transforms a set of comparable data into a visual pattern with which a competent observer interacts. Additionally, these patterns should be created in a way that they can as a heuristic basis for the formation of new hypotheses, assumptions or theory components. Here, the sequential design of the first road is substituted with a complex interactive design between a visual data pattern on the one hand and a competent social scientist on the other hand. Thus, for the second road one needs, first, a transformation from data sets into visual patterns and, second, competent observers who are capable of interacting with these information-rich visual patterns. Finally, the third road to comparative research is based on an unusual assumption, namely on strictly incomparable data. Normally, incomparable 26 Karl H. Müller | Niko Toš data sets are not used in comparative research, simply for the reason of their incomparability. But Part III will open up some surprising possibilities within the context of strictly incomparable data sets. Figure 01 summarizes the main ingredients for the three roads to comparative research which, in conjunction, cover the main content for this volume. Figure 01 The Three Roads to Comparative Research Statistical Transformations Comparable Pattern Comparative Micro-Data Recognition Results Incomparable Morphological Micro-Data Transformations Part I – The Analytical Road to Comparative Research Part I: The Analytical Road to Comparative Research Karl H. Müller | Niko Toš The first road to comparative research comprises the methods and techniques which are usually covered under the term statistical analyses in the social sciences and can be separated in broad domains like descriptive statistics, variance analytical methods and multivariate methods1. Over the last decades, a broad range of methods, tools and software-support has become available which offer a wide array of analytical designs in varying degrees of complexity for social science studies in general and for comparative research in particular. Recently, the scope of analytical methods has been widened and deepened considerably. In terms of widening, the last decades experienced a massive increase especially in complex forms of statistical methods as well as in complex modelling of survey data. With respect to deepening, a very interesting field of research has emerged which is devoted to the measurement quality and the construct validity in comparative data sets. Here, new methods like the MTMM-approach (MultiTrait-MultiMethod) have shed new light on the data quality, the construct validity and on the range of measurement errors within comparative data sets.2 In this volume, the analytical methods are used throughout Part I are of low and medium complexity and consist of methods of descriptive statistics like cross-tabulations or correlations, of variance analyses or of multivariate methods like factor analysis. In fact, the analytical methods in Part I of the book are the most frequently utilized techniques in comparative research. In this sense, Part I shows the strength as well as the limitations inherent in the analytical road to comparative research. Figure I.01 summarizes the basic design for the first road to comparative research. 1 See, for example, the basic distinctions in Bortz, 1999. 2 See, for example, Saris and Sniderman, 2004 or Saris and Gallhofer, 2007. 30 Karl H. Müller | Niko Toš Figure PArT i.01 The First Road to Comparative Research Section I Working Conditions, Household Work, Social Activities and Migration 1 Patterns of Work Flexibility in Typical Jobs in Eastern and Western Europe Claire Wallace | Endre Sik | Bori Simonovits 1.1 Introduction Europe is often compared unfavorably to the USA in terms of flexibility: while the US labour market is flexible, Europe is alleged to suffer from “Eurosclerosis” [Ganßman, 2000]. The assumption is that all European countries more or less suffer from the same kinds of inflexibility – that flexibility is a uniform phenomenon. However, flexibility can actually mean a range of things [Felstead and Jewson, 1999]. In addition to the distinction between functional and numerical flexibility [Pollert, 1988], flexibility can mean the removal of regulations and institutions protecting workers [Riboud, Silva-Jauregui, et al. 2001] or flexibility can be narrowly defined in terms of the extent of part-time work, fixed term contracts and self-employment. However, in most cases, flexibility is assumed from external variables: that is, it is assumed that if de-regulation is introduced, labour markets will become more flexible, like in the US. We have criticized these narrow models elsewhere (author ref). Here we set out to show the real extent of flexibility, based not upon assumptions about people’s behavior nor upon a small range of activities but rather upon the actual behavior of the population in organizing their work and adjusting people to jobs or jobs to people. Furthermore, we seek to ascertain whether there is really a “European” model of flexibility or rather a number of models of flexibility in different European countries. Flexibility is associated with labour market segmentation. Deriving from the “flexible firm” hypothesis [Atkinson and Meager, 1986] and earlier from the dual labour market theories [Doeringer and Piore, 1971], it is argued that in a segmented labour market there is a core protected group of employees and a more peripheral group of workers that can be more easily dismissed following fluctuations in demand. In this secondary labour market, we are more likely to find more marginal and less powerful groups of workers: immigrants, ethnic minorities, women, and younger workers. However, since the 1980s, more and more workers have become flexible [Standing, 1999; Beck, 2000]. Hence, the conditions that used to be associated with the secondary labour market have become more common for larger groups of workers, including managerial and professional workers [Sennet, 1998]. Nevertheless, there is substantial literature arguing that flexibility is seen as affecting the po[][sition of weaker groups in the labour market, such as women and young people [Dex, 1997; Perrons, 1998; Standing 1999; Bradley, Erikson, et al. 2000]. In this respect, we can ask the question: is flexibility evenly spread around the workforce or is it concentrated in particular groups? The flexibility debate began in Western Europe and has only recently been imported into the post-communist countries of Eastern and Central Europe. 36 Claire Wallace | endre Sik | Bori Simonovits The assumptions upon which it is based are those derived from Western labour markets, in which cultures of work and care are implicitwHaas, 2003]. In these labour markets, women form a flexible group because they have the major responsibility of caring for the family; since in many countries they are not fully employed, they constitute a “reserve army of labour”. In the ECE countries, by contrast, women were traditionally fully employed, and flexibility was organized within the context of a full-time, permanent job [Corrin, 1992; Medgyesi, 2002]. However, the introduction of de-regulatory measures and the rising unemployment among women and young people mean that perhaps some of the same conditions now apply in those countries too. Flexibility takes many forms. While time flexibility has been rather well documented [European Foundation for the Improvement of Living and Working Conditions, 2002; Dex, 1997; O’Reilly and Fagan, 1998; Conditions, 2002], the emphasis has been mostly on the increasingly important role of part-time work and a variety of flexible hours (annualized hours, shift working, evening and weekend working, time sharing, term-time working etc.), which have enabled employees to meet the demands of longer opening hours, round-the-clock demand and just-in-time production. Part-time work need not be precarious, and it has been the policy goal in countries such as Sweden and the Netherlands to introduce security for part-time workers with comparable conditions to those of full-time workers [Boje and Strandh, 2003; Jager, 2003]. Contract flexibility has also been rather well discussed in terms of jobs often with fixed contract duration. However, flexibility of place has enjoyed much less attention, except in the analysis of telework and other IT professionals [Huws, 1996; Hochgerner, 1998]. Nevertheless, we can see this as another area in which the needs of the labour market and the availability of the workers come together in different ways. Inflexibility of place is seen as one of the barriers to meeting labour market needs and reducing joblessness [Vecernik and Stepankova, 2002]. These are all sources of flexibility within a job. However, another source of flexibility, which is seldom considered, is the extent to which people might combine several jobs or several sources of income. This kind of additional flexibility can provide new opportunities for some (for example, it can be way of venturing into self-employment) or a source of hyper-exploitation, as people undertake several jobs with declining wages to make ends meet [Nelson and Smith, 1999]. Additional job holding (whether in the formal or informal economy) has been a common source of economic activity in Eastern and Southern Europe in order to make up for low or declining wages. For this reason, we have looked at flexibility of time, including a range of different working arrangements, flexibility of place, or where a person lives Patterns of Work Flexibility in Typical Jobs in eastern and Western europe 37 and works and flexibility of conditions, or the sort of contract arrangements that exist. In addition, we have looked at the extent to which people have more than one source of income. Finally, we have looked at the extent to which these different kinds of flexibility relate to one another. Flexibility is normally discussed in terms of “atypical” or “non-standard” jobs, assuming that regular, full-time employment is standard or typical work [Keller and Seifert, 1995; Zilian and Flecker, 1998]. Atypical employment implies the introduction of new work forms. In fact, we could argue that these work forms are not really so new, but that they were regarded as atypical throughout much of the post-war period, when the standard working week and standard working day were seen as the norm (at least for men). Such jobs were traditionally most often protected by trade unions. This is, in fact, still the normal working pattern for the majority of Europeans, and it was certainly the typical pattern in Eastern and Central Europe for most of the post-war period under Communist regulation, although it has been changing rapidly since 1989. Hence, standard typical employment is not usually regarded as flexible. We question this assumption. The countries under consideration exhibit different “regimes of regulation” which affect flexibility through the combination of labour market developments, state and other regulation, as well as social dialogue [Lodovici, 1999; Regini, 2000]. The UK has introduced more liberal, US-style flexibility policies that allow people to be hired and fired under a variety of different conditions relatively easily, with weak unions and decentralized collective bargaining. This has been achieved mainly by taking away regulations that protect workers. Even though this has been mitigated by the policies of the “New Labour” since 1997, we might still regard this as more “de-regulated flexibility”[Cousins and Tang, 2003]. The Netherlands and Sweden, on the other hand, have introduced flexibility policies during the 1990s to counteract unemployment. These have been introduced in the context of strong regulations to protect the working conditions of flexible workers and with strong trade-union and state intervention. We can see this as “regulated flexibility”. However, we should be aware that Sweden and the Netherlands have been regulated in very different ways. In Sweden, flexibility has been introduced so that people can take time to raise families or study while still maintaining full time employment, and gender equality has been an explicit policy goal [Boje and Strandh, 2003]. In the Netherlands, by contrast, regulation takes place by introducing part-time employment or creating various flexible time options within employment in order to draw women into the workforce and to ensure that time flexible workers receive the same social benefits full-time workers [Jager, 2003]. Of all the post-communist countries, Hungary has embraced 38 Claire Wallace | endre Sik | Bori Simonovits flexibility most enthusiastically and also from early on, even though much legislation was not successfully implemented [Medgyesi, 2002]. We might term this “partially regulated flexibility”. The Czech Republic and Slovenia have tried to resist flexibility, seeing it as a threat to conditions, but nevertheless have introduced a range of legislation allowing for different kinds of con-tractual arrangements [Vecernik and Stepankova, 2002; Sik, 2003]. However, Slovenia has the strongest labour market protection in Europe [Outlook, 1999; Riboud, Silva-Jauregui, et al. 2001; Cazes and Nesporova 2003]. Finally, Romania and Bulgaria have not introduced legislation to encourage flexibility to any extent, but the populations have been forced to become flexible due to the economic problems faced in the labour market [Stanculescu and Berevoescu, 2003]. This kind of flexibility tends to bypass state regulation or not be regulated at all. We might term this “unregulated flexibility”, where workers are forced to become flexible, even though most would prefer a regular job (author ref). We are, therefore, interested in considering what kinds of flexibility actually occur under these different regulation regimes. The types of labour market regulation regimes outlined above could be taken as predictors of flexibility. However, rather than assuming that labour market regulation or de-regulation leads to flexibility, we should look instead at the actual patterns of flexibility, which we can then map back onto the different regulatory regimes. It is often the case, for example, that regulations are introduced and are not implemented, or are introduced but have no effect, as is the case with some of the flexibility legislation in ECE [Medgyesi, 2002]. Indeed, regulations can have entirely unintended consequences, and as the transformation of the ECE countries has been a big “experiment”, we might expect this to be the case. Research Questions Summing up the various arguments presented here, we will investigate the following research questions: – What is the real extent of flexibility in Europe? Is it actually so little? – Is there a general “European model” of flexibility or are there rather country specific patterns? In particular, do ECE countries fit the model developed for Western countries? – Is flexibility concentrated among disadvantaged social groups such as women, young people and those in the lower parts of the labour market as the dual labour market theory would claim? – How does the regulation of flexibility in different countries affect the actual way in which flexibility takes place? Patterns of Work Flexibility in Typical Jobs in eastern and Western europe 39 For this purpose, we use the data set from the project “Households, Work and Flexibility” (HWF) (http://www.hwf.at). The survey is based upon a representative cross section of people between 18 and 65, numbering around 1000 in each country, and was carried out in 2001. The total sample is 11194 respondents. In this survey, rather than beginning with arbitrary classifications of different forms of work (part-time, full-time etc.), we began by asking people about all the kinds of work they do and the various hours and places that they work in order to try to define flexibility inductively rather than deductively. In this paper we will carry out the following analysis. First we will define the different kinds of flexibility. Then we will look at how they relate to one another to see if there are patterns of multiple flexibility. Next, we will consider how these different kinds of flexibility differ between countries. Finally, we will look at the socio-demographic characteristics of these different kinds of flexibility and consider whether there is one general model of flexibility or in fact a range of different models. 1.2 The Extent of Flexibility in Eastern and Western Europe We will begin with our first research question: what is the extent of flexibility in Europe? In the usual approach to flexibility, fairly simple indicators of non-standard employment are used to compare flexibility: part-time work, self-employment, fixed contracts (despite the problem of defining each of these indicators) [Bastelaer, Lemaitre, et al. 1997]. If we apply this to the HWF survey, we come out with striking differences between countries. Starting with part-time work, there is a very different pattern in Eastern and in Western Europe. While in Sweden, the Netherlands an the UK there are many part-time workers, since this is the way that women combine child-rearing with labour market activities, this form of flexibility is mainly absent in the post-communist countries. Women work either full-time or are housekeepers. Those who work part-time are usually men who are either pensioners or nearing retirement or who have individual reasons for not working full-time, such as disability or sickness [Sicherl, Stanovnik, et al. 2003; Wallace, 2003]. The conditions governing part-time work also discourage employers from introducing this kind of work. Secondly, shift work is found most often in those countries heavily dependent upon large industrial enterprises, as in the post-Communist countries. Thirdly, self-employment is found more in certain countries than others and tends to be a product of the legislation 40 Claire Wallace | endre Sik | Bori Simonovits that encourages self-employment. The UK, the Czech Republic, Sweden and Hungary have the highest number of self-employed workers, the majority of whom are men. Additionally, the role of farmers, while being fairly marginal in EU countries, is substantial in ECE countries. In particular, there is a very large number of farmers in Romania (20%); these are people who are excluded from the conventional labour market because of restructuring and have returned to a peasant-style agriculture, which is mostly subsistence in nature [Stanculescu and Berevoescu, 2003]. Thus farm work, self-employed, part-time work and shift work tend to muddy the analysis because their distribution reflects rather the structure of industry and employment in different countries; we are not comparing like with like. Here, in our concept of flexibility, we want to go beyond standard definitions. We want to understand the real extent of flexibility in Europe in order to see if European labour markets are really as inflexible as they are sometimes portrayed. We have defined flexibility as either holding multiple jobs or as a multidimensional phenomenon within a certain job. Thus, we have selected typical jobs, since these are be less likely to reflect the idiosyncrasies in the structure of the labour market. We have defined the following forms of flexibility: Time flexibility is defined as people on a non-regular or irregular working schedule. Place flexibility is defined as people working at home either the whole time or part of the time, working abroad or having an irregular place of work (commuters were excluded). Contract flexibility is defined as people having anything but a permanent regular contract ( i.e. no contract, fixed term contract, on call, with a temporary work agency, on a fee only basis, subject to performance or on a work experience project). Income flexibility includes all those with more than one income source. As to the more complex flexibility measures, while combined flexibility covers those with time and/or place- and/or contract-flexibility, cumulative flexibility covers those characterized by all forms of flexibility simultaneously. The first aim of this analysis is to answer the following questions: to what extent are the four basic and the two combined forms of flexibility present in the eight European countries? What are the characteristic relationships among the four dimensions of the flexibility phenomenon, and to what extent are they similar or different in the eight countries? As already mentioned, we know that certain patterns of flexibility are strongly associated with certain labour market patterns. Therefore, in order to compare like with like, we have excluded several kinds of flexibility that are only country-specific and have thus attempted to homogenize the sample. In the first step of the analysis, we discounted shift work. Next we removed Patterns of Work Flexibility in Typical Jobs in eastern and Western europe 41 farmers and self-employed from the sample. The reason for this is that these groups are flexible by definition and are likely to display all aspects of flexibility simultaneously. Thirdly, we excluded part-time workers, as their presence in the data would have clouded the more subtle associations among the four forms of flexibility. With these exclusions we can claim that our paper concentrates upon the theoretically most homogeneous forms of flexible work in contemporary Europe, i.e. employees working in full-time employment, who are not usually designated as “flexible”. This part of the labour force is still the majority of those between 18 and 65 from which our sample is drawn. If we find flexibility in this group, we can assume that this is a strong test of the spread of flexibility. We are able to create a more nuanced picture of flexibility and of the stratification of the different forms of flexibility in Europe by focusing upon typical rather than atypical jobs. 1.3 General Model of Flexibility or Particular Ones? Now we turn to our second research question: is there a general model of flexibility or are there strong variations between countries? Table 1.01 shows that even in the homogenized approach, nearly one third of the workers are employed in a job with more than one type of flexibility (combined flexibility). TABle 1.01 Flexibility Patterns by Countries, Within Full-Time Workers, Excluding Farmers, Shift Workers, Self- Employed and Part-Time Workers (%) income- Time- Place- Contract- Combined Cumulative N flexibility flexibility flexibility flexibility flexibility flexibility (part-time excluded) united Kingdom 4 18 13 15 34 2 404 The Netherlands 5 16 8 11 29 1 416 Sweden 6 14 9 10 27 0 732 Slovenia 3 19 11 6 30 0 385 Czech republic 10 20 7 13 34 1 762 Hungary 2 28 7 13 37 1 537 romania 4 17 4 4 21 0 524 Bulgaria 1 12 2 26 35 0 679 Total 5 18 7 13 31 1 4438 42 Claire Wallace | endre Sik | Bori Simonovits However, it is not very common for workers to combine all four aspects of flexibility (cumulative flexibility). In terms of income flexibility, the Czechs stand out, with 10% having more than one source of income; this is followed by those in Sweden. In the Czech Republic, holding a second job is quite widespread. In Sweden, this flexibility is due to the labour market and welfare regulations that allow people such as students to take on additional jobs to supplement their grants. The Hungarians are the most time flexible. This may be the last remnant of the former second economy under liberal Kadarism, where flexi-time schedules were the essential element for giving underpaid employed the opportunity to earn some extra “on the side” [Sik, 1993]. However, it may also be reflecting the fact that Hungary was the first and most enthusiastic of the ECE countries to embrace flexibility. This is followed by the Czechs and the Slovenians and only thereafter the UK. Therefore, it seems that the more successful transition countries have managed to introduce a range of flexible working practices, even if they have not always explicitly embraced flexibility. Western countries no longer seem so time flexible once the effects of part-time work are removed. Place flexibility is common in Slovenia, partly because employment is provided in a few industrial and urban centers, while the population lives in more scattered communities. In the UK there has been an effort to build up home-working and teleworking [Huws, 1996]. Contract flexibility is most common in Bulgaria, where 26% of the employees in full-time employment do not have a permanent working contract, showing a high degree of precariousness in the Bulgarian labour market. However, this is also the case for 15% of full-time employees in the UK, where encouraging this kind of precariousness was a specific governmental goal during the 1980s. In Bulgaria and the UK, it is common for people to have “no contracts”, and thus not even receive the kind of security that a fixed term contract might provide. Looking at the combined forms of flexibility, we can see that combining different kinds of flexibility is found among about one third of full-time workers. Combined flexibility is most common in Bulgaria, Hungary, the Czech Republic and the UK. The Romanian full-time employed work mostly in jobs that are non-flexible in terms of place and contract; consequently, the level of combined flexibility there is also very low – which may be a remnant of the inflexible socialist full-time economy. Hence, we could say that there are several dimensions to flexibility in Romania: there is the difference between the very flexible peasant farmers and the normally employed workforce, but also within the normally employed workforce there is a strong separation Patterns of Work Flexibility in Typical Jobs in eastern and Western europe 43 between the traditionally employed sector, which is rather inflexible, and the flexible sector. This is rather reminiscent of the situation in Southern Europe with labour markets segmented between “protected” and “unprotected” workers [Regini, 2000]. Cumulative flexibility (that is all forms of flexibility) is rather rare, so we have excluded it from further analysis (while noting that 2% of full-time workers in the UK exhibit this kind of flexibility). The overall association among the four forms of flexibility is a low level of positive correlation (Table 1.02). This means that the different kinds of flexibility tend to be associated with one another. However, contract flexibility and time flexibility are the most strongly correlated, followed by place flexibility and time flexibility. So there are two dimensions to the association of different kinds of flexibility. On the one hand, we have time and contract flexible workers and on the other hand we have time and place flexible workers. TABle 1.02 Correlation Between Different Kinds of Flexibility (Persons) Income flexible Time flexible Place flexible Contract flexible Income-flexible 1 0.088 0.079 0.080 Time-flexible 1 0.196 0.243 Place-flexible 1 0.150 All correlations are significant at the p=0.01 level To understand the importance of having low level and positive association among the four dimensions of flexibility, we should try to imagine a world with strong positive and negative associations: – The former would mean a dichotomized world of work where there were either inflexible jobs or jobs that were flexible in all four dimensions. Such a job market would be similar to what segmented labour market theory suggests. – The world with a negative association between the four dimensions of flexibility would be a world with balanced and counter-balanced flexible jobs. If one aspect of the job were flexible, the other aspects would be inflexible. From the Table 1.03, we can see that, although there is general y a positive correlation between the different kinds of flexibility in different countries, there are also important differences between countries. In the UK, there are general y strong correlations between contract flexibility and al other kinds of flexibility. 44 Claire Wallace | endre Sik | Bori Simonovits TABle 1.03 Correlation Between Flexibility Types by Country UK Time flexibility Place flexibility Contract flexibility Income flexibility 0.136 0.153 Time flexibility 1 0.260 Place flexibility 1 0.255 NL Time flexibility Place flexibility Contract flexibility Income flexibility 0.138 Time flexibility 1 0.260 Place flexibility 1 Sweden Time flexibility Place flexibility Contract flexibility Income flexibility Time flexibility 1 0.268 Place flexibility 1 Slovenia Time flexibility Place flexibility Contract flexibility Income flexibility Time flexibility 1 0.236 0.291 Place flexibility 1 Czech Republic Time flexibility Place flexibility Contract flexibility Income flexibility 0.138 0.148 0.133 Time flexibility 1 0.217 0.182 Place flexibility 1 0.199 Hungary Time flexibility Place flexibility Contract flexibility Income flexibility Time flexibility 1 0.264 0.244 Place flexibility 1 0.123 Romania Time flexibility Place flexibility Contract flexibility Income flexibility 0.119 0.144 Time flexibility 1 0.381 0.403 Place flexibility 1 0.408 Bulgaria Time flexibility Place flexibility Contract flexibility Income flexibility Time flexibility 1 0.241 0.159 Place flexibility 1 0.172 Only associations significant at the p=0.01 level (2 tailed) are shown Patterns of Work Flexibility in Typical Jobs in eastern and Western europe 45 This implies a rather dichotomized labour market driven by contract flexibility. In the Netherlands, the only significant correlations are those between time and contract flexibility and between income and contract flexibility. Hence, it would seem that in the Netherlands, flexibility is concentrated in particular types of work. In Sweden, it is even more concentrated, with only time and contract flexibility being significant. Thus, the range of flexible work is even more restricted. In Slovenia there are positive associations only between time and place flexibility and between time and contract flexibility. Thus, in the most regulated labour markets, flexibility is concentrated in particular work forms. In most ECE countries we see a different pattern. In all of these countries, time and contract flexibility are rather strongly correlated as is time and place flexibility: temporary jobs are associated with irregular working hours and irregular working hours are also associated with irregular working places. In Romania, we find the strongest correlations. They are especially strong in the relationships between precarious contracts and time flexibility and also with place flexibility. This suggests a very segmented labour market in Romania, even after we have excluded the farmers and the self-employed. However, this is not the case in Bulgaria, where contract flexibility is more weakly associated with place and time, but place flexibility and contract flexibility are also associated. In sum, all countries fall into a low-positive pattern – the general European pattern. Sweden and Slovenia are the closest to the no-correlation situation, meaning that the various aspects of flexibility are unrelated, while Romania is closest to the dual labour market structure. 1.4 Labour Market Stratification and Flexibility Now we turn to our third research question: is flexibility concentrated among the predictable groups of workers (women, young people, less skilled) as much literature would suggest? To understand how multiple flexibility relates to labour market stratification, we have developed a simple, multilevel model of flexibility by combining all the countries. In the following table, the dependent variables are the various forms of flexibility, narrowed down to time, place and contract flexibility. The independent variables are the basic socio-demographic dimensions of the individuals. The dimensions of age and gender would indicate if multiple flexibility, as we have measured it, is found in accordance with the general literature of flexibility using other measurements. The dimension of education helps us to understand how multiple flexibility 46 Claire Wallace | endre Sik | Bori Simonovits might relate to the stratification system (using education as a proxy for social level). The dimension of urban-rural differences reflects our understanding of the processes taking place in some post-communist societies, including that of re-ruralization. First we ran our logistic regression models on the pooled sample, and then we ran them country by country. Below we compare the social characteristics behind the different forms of flexibility in all eight countries. TABle 1.04 The Odd Ratios1 of the Types of Flexibility by Predictor Time-flexibility Place-flexibility Contract-flexibility 18–28 years old (1) 1.284 1.116 3.396 29–36 years old (2) 0.904 0.949 1.246 46–54 years old (4) 0.798 0.810 1.054 55–65years old (5) 1.106 1.175 1.664 1=male, 0=female 1.120 2.049 0.820 Primary (1) 1.499 1.302 1.598 Tertiary (3) 1.304 0.796 0.589 urbanized area (1) 1.023 0.995 1.027 rural area (3) 1.168 1.413 1.448 All forms of flexibility affect older and younger people the most. Young people are especially strongly affected by contract flexibility, as are older people – though to a lesser extent. Males are more likely to be affected by all forms of flexibility (except contract flexibility) than females. Being male significantly increases the possibilities of being flexible in regards to place and slightly decreases the chance of being contract flexible. Living in a rural area also increases the chances of being in flexible work. Education has a definite impact on flexibility. The least educated are the most flexible in every dimension, while the higher educated are mostly negatively associated with flexibility, except in the case of time flexibility. The role of higher education, therefore, increases the chances of time flexibility but decreases the probability of contract and place flexibility, and consequently, 1 Odds ratios are interpreted by looking at how far above or below 1 they are. Approximately 1.6 and above can be seen as having a strong positive, 0.7 and below a strong negative impact (in relation to the reference group stated in the table) on the dependent variable. Patterns of Work Flexibility in Typical Jobs in eastern and Western europe 47 the contradictory influence has no significant influence on combined-flexibility. From this, we can tentatively suggest that there are two divergent types of flexibility: high status flexibility of the better educated, which is associated more with having flexibility of time, and low status flexibility, which is associated with a lower level of education, being male, being younger or older, and living in a rural area. Low status flexibility is associated with contract, place and time flexibility, as well as with the combination of all of these. In the next section, we will compare the influence of the same basic social determinants on the various forms of flexibility country by country, posing the question: are there country specific differences in explaining flexibility? If the answer is no, then – from a sociological perspective – we can speak of a single model of flexibility. However, if the answer is yes, then we have to focus on individual countries (or groups of them) if we want to understand the social consequences of flexibility. TABle 1.05 The Odd Ratios of Time-Flexibility by Country and Predictor UK NL S SI CZ HU RO BG Age (reference group 3=37-45 years old) 18–28 years old (1) 1.146 1.178 0.841 1.901 1.266 1.479 1.186 0.909 29–36 years old (2) 0.795 1.253 0.481 0.879 0.924 0.913 0.489 1.334 46–54 years old (4) 0.892 0.789 0.487 0.632 0.761 1.106 0.738 1.086 55–65 years old (5) 1.745 1.455 0.477 1.382 1.229 1.132 1.430 1.147 gender 1=male. 0=female 0.668 0.554 0.908 0.746 1.440 1.977 1.761 1.341 Education (reference group: 2=secondary education) Primary (1) 0.734 2.528 1.090 1.456 1.688 1.509 3.406 2.361 Tertiary (3) 0.787 0.861 1.744 1.905 1.359 1.294 2.552 1.881 Place of residence (reference group: 2=intermediate area) urbanized area (1) 1.220 0.684 0.739 1.185 1.374 1.391 0.549 1.438 rural area (3) 1.015 0.516 1.084 1.317 0.869 1.322 1.702 1.444 As far as time-flexibility is concerned (Table 1.05), we see differences according to gender. In all three western countries (especially in the Netherlands), 48 Claire Wallace | endre Sik | Bori Simonovits females are the most likely to become time-flexible, as is also the case in Slovenia. However, in the remaining four ECE countries, being male increases the chances of being time-flexible. Except for Slovenia, there is a clear East-West cluster in this case – gender and time. Being at either end of the age range (young or old) increases time flexibility in the UK, in the Netherlands somewhat, in the Czech Republic, in Hungary and in Romania. In Bulgaria, age makes less difference for time flexibility. The anomaly is Sweden, where the middle-aged group of 37– 45 is the most flexible, presumably because they are able to take advantage of the various leave arrangements for combining family and work. In most countries, apart from the UK, time flexibility is associated with having a lower level of education, which is especially true in Romania, Bulgaria and the Netherlands. There is, therefore, a general pattern. However, in Sweden, Slovenia, Romania and Bulgaria (and to a lesser extent Hungary and the Czech Republic), time flexibility is also increased by having more education. In these countries, we therefore find high status flexibility, but not in the others. The urban-rural dimension shows different patterns in each country, with rurality increasing the chances of time flexibility, especially in Romania. TABle 1.06 The Odd Ratios of Place-Flexibility by Country and Predictor UK NL S SI CZ HU RO BG Age (reference group 3=37-45 years old) 18–28 years old (1) 0.792 1.349 0.851 1.115 1.556 0.639 2.778 0.643 29–36 years old (2) 0.945 0.940 0.777 0.801 0.940 0.881 0.830 1.343 46–54 years old (4) 0.434 0.326 1.161 0.935 0.740 0.981 1.153 0.874 55–65years old (5) 1.924 0.721 0.681 1.207 1.526 1.037 1.295 1.531 gender 1=male. 0=female 1.111 6.175 2.259 2.485 1.850 1.845 3.919 1.422 Education (reference group: 2=secondary education) Primary (1) 0.349 4.233 2.833 0.782 1.274 0.775 4.953 1.327 Tertiary (3) 0.752 0.820 1.296 0.593 0.399 1.480 0.371 0.757 Place of residence (reference group: 2=intermediate area) urbanized area (1) 0.715 0.405 1.163 1.025 0.932 1.210 0.850 1.373 rural area (3) 1.204 0.791 0.705 0.846 0.902 2.197 3.312 1.429 Patterns of Work Flexibility in Typical Jobs in eastern and Western europe 49 Therefore, we can see that there are important differences between European countries in terms of time flexibility. There is a general tendency for the less educated to be time-flexible, but we can identify a certain high status flexibility for the better educated in some countries. Place flexibility seems to have a different dynamic in the various countries than the other dimensions of flexibility. It most often affects men and it is mainly associated with prime-aged men with low or intermediate education. We can assume perhaps that these men are not so much teleworkers (who are too small a specialized a group to have been captured by our survey), but most likely work in particular kinds of jobs (lorry driver, plumber, builder etc.). The fact that in some countries contract and place flexibility are strongly associated also implies that perhaps some of these men are contractors. TABle 1.07 The Odd Ratios of Contract-Flexibility by Country and Predictor UK NL S SI CZ HU RO BG Age (reference group 3=37-45 years old) 18–28 years old (1) 2.921 3.274 9.977 13.451 3.013 4.893 2.816 1.659 29–36 years old (2) 0.828 1.029 3.992 1.914 1.400 1.791 1.018 0.778 46–54 years old (4) 1.109 0.959 0.432 1.237 1.119 1.283 0.929 0.934 55–65years old (5) 2.092 1.209 2.240 5.712 2.109 1.932 2.342 1.162 gender 1=male. 0=female 0.657 0.665 0.475 0.474 0.841 1.135 1.705 0.965 Education (reference group: 2=secondary education) Primary (1) 0.770 3.288 1.427 1.626 2.851 2.041 9.924 23.644 Tertiary (3) 0.633 0.552 1.006 0.768 0.666 0.583 0.294 0.504 Place of residence (reference group: 2=intermediate area) urbanized area (1) 1.459 0.937 1.042 1.230 0.865 0.937 0.487 1.042 rural area (3) 1.261 1.264 2.647 0.997 0.714 2.112 1.249 1.133 As to contract flexibility, we see that it is overwhelmingly found in the youngest and oldest age cohorts in each country. This is especially the case in Slovenia and in Sweden. In almost all countries, being female increases the chances of contract flexibility, although in Hungary and Romania it is being male. Having less education strongly increases the chances of contract 50 Claire Wallace | endre Sik | Bori Simonovits flexibility in all countries, except for the UK, and this tendency is especially strong in Bulgaria, Romania and the Netherlands. In most countries, having more education decreases the chances of contract flexibility. In Sweden and in Hungary, the chances of having contract flexibility are much greater in rural areas, but in most other countries (apart from Czech Republic and Slovenia) rurality increases the probability of contract employment. Thus contract flexibility displays one similar feature across the different countries: it affects younger people and older people the most. It is also associated with the less educated. Contract flexibility has the strongest associations of all the different kinds of flexibility, and it is the only one that can really be described as “precarious employment”. Hence, the socio-demographic composition of flexibility is varied. Not only are the most vulnerable in the labour market affected in each country, but it is not necessarily a feminized phenomenon, as the Western literature generally suggests. When we concentrate on time and contract flexibility (which we know are strongly associated with each other – see Table 1.02), we find that in all western EU countries plus Slovenia, it is a feminine phenomenon, albeit more weakly so in Sweden and in Slovenia. In all the remaining ECE countries, flexibility is more masculine. Hence, the cultural differences surrounding work and family in the various European countries account for the vulnerability of women to flexible and precarious work. If we count the least educated as a marginalized group, then we can say that they are most likely to experience flexibility of time and contract in all countries apart from the UK. However, time flexibility is also associated with the better educated in Sweden, Slovenia, Romania and Bulgaria. In these countries, flexibility can have favorable as well as unfavorable dimensions, as both a marginal group and a privileged group are affected. Age is generally an important dimension. It is always young people who are most likely to experience flexibility (except the UK); contract flexibility also affects older workers in many countries. Only in the UK do the young seem less affected by time flexibility than in other countries. Thus we can say that time and contract flexibility do seem to affect marginalized groups in most countries. These marginalized groups are mostly younger and sometimes older workers, but in some countries, mostly Western ones (the UK, the Netherlands), such workers are mostly female, while in other, Eastern countries (Romania, Hungary, Czech Republic and Bulgaria) they are more likely to be male. In Romania and Hungary, marginal flexible workers are also rural. 1.5 How Does the Regulation of Flexibility Affect the Patterns of Flexibility? If we try to match the countries’ governing regimes with the effects of flexibility, we find that the most de-regulated regime (the UK) seems to spread flexibility around a wider group of workers, though mostly women are affected. However, the strongly regulated regimes of Sweden and the Netherlands protect the main group of workers but produce a marginalized sub-group of less educated, younger and female workers who are flexible. The gender dimension and the strength of this exclusion was higher in the Netherlands than in Sweden – the egalitarian nature of the labour market regulations regarding gender in Sweden do seem to have had some influence. In ECE, we find a group of favorably flexible workers in Slovenia, as well as a group of unfavorably flexible workers who tend to be male, less educated and either younger or older. In the mainly unregulated regimes, we find strong labour market segregation in Romania and a flexibility more spread across the workforce in Bulgaria. In these two countries, there are also favorable and unfavorable forms of flexibility to be found. It would seem, therefore, that in the more de-regulated regimes of Bulgaria and the UK, the risk of flexibility is spread around many different kinds of workers, while in the most regulated, it is concentrated among a small very specific group. In Sweden, the regulation has also had the effect of spreading flexibility around different groups, but not so much in the Netherlands. The gendered nature of flexibility seems to reflect not only the regulatory regime, but also the division of labour in the home and the labour market that is traditional in different parts of Europe. 1.6 Conclusions The first conclusion is that, seen from this “bottom up” perspective, there is actual y a great deal of flexibility in Europe. Even if we exclude the usual “atypical” forms of employment and concentrate on mainstream, ful -time, regular workers, we find a substantial amount of flexibility within these jobs, with about one-third of ful -time employees flexible in more than one dimension. Hence, the idea that European labour markets are inflexible can be rejected if we look at the actual patterns of flexibility in terms of what people do, rather than making assumptions based on labour market regulations. Even in ECE countries, where flexibility is rather recent, there are large numbers of 52 Claire Wallace | endre Sik | Bori Simonovits people in flexible jobs, defined in terms of time, place, contract and income flexibility. We can now return to our second research question: is there a general model of flexibility? In general, we can say that there is, since the different kinds of flexibility are weakly associated with one another, meaning that if a worker is flexible in one dimension, he or she is also likely to be flexible in others. The weak correlation, however, means that there are many variations on the model of flexibility. While in some countries flexibility is concentrated heavily in certain kinds of work, in other countries it is more evenly spread. However, even in the group of countries where strong labour market protection helps to protect many workers from precarious work, there is a precarious group of young contract workers with less education, who appear to be excluded. Therefore, it is necessary to also look more closely at the different patterns that flexibility assumes across Europe. We might ask: why are there these differences between countries? One reason is the different traditions of gendered labour market participation [Lewis, 2002; Pfau-Effinger, 2003]. Another is the kind of labour market regulation that takes place in different parts of Europe, regulation that does not necessarily have the intended effects. Another might be the regime legacy, as we see in Hungary. Our third research question is whether flexibility is concentrated among disadvantaged groups. Contract flexibility is strongly associated with time flexibility and also associated with marginalized groups. These are usually younger workers but can also be older workers at the beginning or end of their careers and they are also less educated. They could be said to be the victims of flexibility. However, time flexibility and income flexibility are also associated with more privileged groups – those with higher incomes and better education, although their numbers are small compared to the disadvantaged groups, and they are found mainly in certain countries. The role of gender is not uniform across Europe. In Western European countries (to a lesser extent Sweden) and in the Czech Republic, women are most at risk from flexibility because of their domestic roles, which lead them to have a worse position in the labour market. However, in the other ECE countries, it is often men rather than women who are the most vulnerable to flexibility. Moreover, women’s domestic roles do not seem to affect their labour market position – in this respect at any rate – because they are either in the labour market (and inflexible) or out of it altogether. The Western literature on flexibility has, therefore, created a bias, and cannot be so easily applied in post-communist countries with their different traditions regarding gender. The assumption that women are most flexible is based upon the gender-work models of Western Europe. In Eastern and Central Europe we see rather the legacy of Patterns of Work Flexibility in Typical Jobs in eastern and Western europe 53 the older system where women worked ful -time. However, two factors could be changing this pattern: on the one hand, the loss of women’s jobs in Eastern and Central Europe pushes those countries more in the direction of the kind of gender-divided labour market seen in Western Europe. On the other hand, the pressure to increase the labour force participation of women in Western Europe through the European Employment Strategy might lead to different patterns of women’s employment in those countries too. Finally, to go back to our fourth question: what might the role of regulatory regimes be? The high levels of flexibility in full-time employment seem to exist despite or because of the strong labour market regulation of different regulatory regimes. In those countries where there is strong labour market regulation (Sweden, the Netherlands, Slovenia), flexibility is concentrated in particular kinds of work and particular demographic groups, especially among young people and women. In countries where there is much less job protection (the UK and Bulgaria), flexibility is much more widespread. Hence, regulation does not necessarily result in its intended consequences. Strong labour market protection does not altogether prevent the disadvantageous forms of flexibility; rather it concentrates such work among particular groups. De-regulation, by contrast, tends to spread the effects of flexibility around wider areas of the labour market, affecting more social groups, rather than just females, the young and the less educated. Thus, although there is some similarity in the kinds of flexibility experienced by regular, ful -time workers in Europe, there are also some important differences in the social composition of the flexible groups in different countries. This seems to reflect regimes of regulation as wel as the gender dynamics between the home and the labour market. It also seems to be the case that strong regulation of various kinds leads to flexibility being concentrated among a smal range of workers such as the young and the less wel educated. On the other hand, de-regulation and lack of regulation lead to flexibility being spread to larger groups in the population, even if age and education stil play a part. The other side of the strong protection of certain workers is the creation of a social y excluded group of people who cannot enter the regular labour market and are disadvantaged in terms of age, gender or lack of education. Acknowledgements Our thanks to Sergej Nagaev, Christine Cousins, Siyka Kovtacheva and Ray Pahl 2 Division of Housework in European Societies through the Analysis of Macro Level Factors Christian Bischof 2.1 Introduction Over the past decades, the participation of women in the workplace has clearly increased; however, despite this change, a majority of the domestic work is still done by women. The purpose of this paper is to identify the factors that can help explain this gender disparity. This is done through an international comparison using the data of the European Social Survey for 22 European countries. Factors from both the individual level and the macro level (in the form of context variables for the countries) will be incorporated in the analysis. 2.2 Problems and Theoretical Background Makiko Fuwa believes that the explanatory factors for the current disparity in the division of household labour are not only localized on the individual level. She sees differences on the national level regarding gender differences in income levels, career possibilities and political participation as influences that affect the individual level. The leaders are the Scandinavian countries, having become egalitarian societies to a large extent, which has allowed women to negotiate the division of household labour within a partnership more effectively, thereby achieving a higher level of fairness [Fuwa, 2004]. Thus, one can see a connection between contextual factors and influences on the individual level. Since the latter are embedded in the macro level, the disparities can really only be explained by analyzing the connections between the two levels. The micro-level unit of the household is embedded in a series of macro-level units – from the community, to the social classes, to the state, and finally to the world economy. The status of women varies from level to level. The economic power of women, which can be related to the negotiation process with their partners, depends both on the individual power of the woman as well as on the male-dominated power found in the macro-level, economic, political and ideological realms [Blumberg, 1984]. The study “Premarital Cohabitation and Housework” can be seen as empirical evidence for this, thus adding gender equitable social conditions to the macro level, that partners can also divide the household labour fairly, independent of the individual female and male characteristics [Batalova and Cohen, 2002]. Concerning the question of female employment, two approaches have developed along the contextual level. In the structural approach, the direct governmental effects are seen as significant – either in a typological form of gendered welfare state regimes [ e.g. Lewis, 1992; Sainsbury, 1996] or through 58 Christian Bischof general political conditions in the form of childcare services, for example, parental leave, childcare provision, public childcare, tax benefits [Fuwa, 2007]. The weakness of this approach lies in the assumption that the behavior of women in the labour market is directly influenced through such contexts. In the culture-based approach, however, the differences between the political contextual influences and the actual participation in the labour market do not play a determining role. This perspective addresses the historical development of social conceptions regarding the gender specific division of work, which determines political decisions and labour market practices [Haas, 2005]. In an analysis of eight European countries [Haas, et al. 2006] arrived at the following results: attitudes towards gender roles are more strongly traditional in Eastern European countries than in Western European ones, which stands in contrast to a higher female participation in the labour market; this high employment of Eastern European women can be explained by financial necessity. In the Eastern European states during the post-war period, a dual-earner model developed under the influence of Socialism. The employment of women was one of the desired conditions of the state ideology, thus, the state made sufficient childcare available [Scott, 1974]. On the individual level, three factors that influence the division of household labour between the genders can be identified in the investigations that have been undertaken so far. First, the relative income resources: in societies in which women have a higher average income in relation to men’s income levels, the women also have more power in negotiating the division of household duties. In these countries, women also have better career possibilities, which then leads to a higher income, and thus again to a stronger negotiation position. Second, time availability: the women who are employed to a greater degree also divide the household work more fairly with their partners. Since the women lend their working abilities to the labour market as wel , there are fewer resources left for the household work, and thus, the men must help more domestical y. Third, men’s positions towards egalitarian gender roles also influence the division of domestic work: husbands who advocate greater gender equality also share the household responsibilities to a higher degree [Fuwa, 2007]. 2.3 Data and Research Design For the analysis, the data set from the European Social Survey (ESS) was utilized. The survey was conducted and repeated in 22 European countries, and its purpose was to identify and explain the changing characteristics, behavior patterns, values and social norms from various European populations. Division of Housework in european Societies 59 The questionnaire for the main part included socio-economic, socio-political, socio-psychological and socio-demographic variables, supplemented by a rotating selection of additional questions. The ESS was financed by the European Commission’s Fifth Framework Program, by the European Science Foundation and national funding bodies in each country. The second round from 2004 was used in the analysis. Only cases in which the respondent lived with a partner were used. Additionally, the respondent was required to be at least 18 years old. The percentage of women involved in household work was used as the dependent variable. This share was determined by the following: the respondents were asked to give the number of hours spent on housework, excluding that used for childcare. This question was posed for both weekdays and weekends, from which the total number of hours per week was determined. Additionally, the respondents were asked to specify how much of this time was spent by them or by their partners; the possible answers were: none or almost none; up to a quarter of the time; more than a quarter; up to half of the time; all or nearly all of the time. For better intelligibility, these categories were then calculated as percentages along the following scale: 0, 20, 40, 60, 80, 100. Since these values were presented for both partners, cases in which the total percentage for the pair are over 120% or under 80% were eliminated. Also, since these questions were answered by respondents, who could be either male or female, the values were calculated so that they always present the woman’s share of the housework. The dummy variable “respondent is male” was introduced as a control variable; it is known from previous studies that results vary depending on the provider of the information [Fuwa, 2004:756]. As explanatory factors for the individual level, variables come into play that were identified in previous research as the main factors affecting the division of household work. First is the relative income contribution, which was divided into the shares provided by the man and woman. This variable was based on the question: “Around how large a proportion of the household income do you provide yourself?” The possible answers were: none; very small; under half; about half; over half; very large; all. This variable was coded in case of a male response to show what percentage of the household income the woman contributed. As the second factor, the availability of time turns out to be relevant, and came into the analysis through the specification of work-time per week, for both the man and the woman. The stance regarding egalitarian gender roles was incorporated in the form of an index. This was based on five questions addressing the topic of male and female roles in the family: “A woman should be prepared to cut down on her paid work for the sake of her family”, “Men should take as much responsibility as women for 60 Christian Bischof the home and children”, “When jobs are scarce, men should have more right to a job than women”, “When there are children in the home, parents should stay together even if they don’t get along”, “A person’s family ought to be his or her main priority in life.” The possible answers were: agree strongly; agree; neither agree nor disagree; disagree; disagree strongly. Negatively formulated questions were converted so that the index could assume a value between one and five, where five represents an egalitarian conception of roles. The Cronbach Alpha value of the index amounts to 0.61. Additional variables were used as control variables on the individual level. Household income was used to control for the influence of material resources. Since gender roles change only gradually, age becomes an influential factor. Thus younger couples should demonstrate a more egalitarian division of household work than older ones. Additionally, the duration of the cohabitation could be expected to be influential; couples initially displaying an egalitarian arrangement can lose this over the course of their cohabitation. One factor might be the woman’s occasional withdrawal from professional life due to childbirth, in which case she takes over more of the household responsibilities and often retains these responsibilities afterwards, even when returning to the professional world. The following indicators were used as dummy variables: if the statements came from a man, this factor was also controlled for. Furthermore, if a couple is married, it could be expected that they might not maintain such an egalitarian division of labour because marriage presupposes more conservative values and, thus, a less egalitarian view of women’s roles. The level of education was seen in the response concerning the completion of high school. A more progressive image of gender roles is associated with a higher level of education. Here only the education level of the respondent was used; as it could be assumed that the relationship is rather homogenous regarding education, this response should be sufficiently valid (the measure of association yields a value of 0.58, according to Cramer’s V). The final control variable entered the analysis in the form of religion, with the four most common faiths coded as dummy variables – these are: Roman Catholicism, Protestantism, Christian Orthodoxy and Islam. It was assumed that all religions engender a less egalitarian view of roles, though the influence of Protestantism could be assumed to have the least effect. On the national level, four factors come into play, the most important for the survey being the Gender Empowerment Measure (GEM), which is used by the United Nations Development Program (UNDP) to measure gender imbalances in connection with female participation. GEM is based on three dimensions: political participation and decision-making, economic participation and decision-making, and the availability of economic resources. The Division of Housework in european Societies 61 first dimension is measured by the gender balance in the national parliament; the second by the relationship between the executive positions of the public administration and those in the economic sector; the third by the income structure. The GEM ranges in value from zero to one, with a higher value indicating a more equal gender balance. As an indicator of economic development, the gross national product per capita (GNP) was used; to account for the differences in purchasing power in the different currency areas, the adjusted purchasing power value was used. The GNP entered the analysis due to the following reason: modern societies are per se interested in gender equality because progressive political and economic institutions allocate corporate positions according to rational considerations, and their development likely reduces inequalities [Fuwa, 2004:757]. As an additional macro indicator, the female employment rate was used, which specifies the percentage of women participating in the labour market. This rate indicates the share of employed women between the ages of 15 and 64 within the total female population of the same age range. Professional occupation, in addition to other functions, allows one to achieve a higher social status, and is, therefore, a means of attaining economic power, and thus also more equality. As the final factor, the GINI-Coefficient was used. This is a measure used to indicate inequality regarding capital. It shows the distribution of capital within a country, and it ranges between zero and one hundred, with zero representing a complete equality in capital distribution and one hundred indicating complete inequality. 2.4 Results Differences by Country Table 2.01 shows the average female participation in household work, with the lowest being 58% in Sweden. Sweden is relatively clearly ahead of the other countries. The next two positions are also filled by Scandinavian countries, Denmark and Finland, and more Northern European countries follow with Norway ranked sixth and Iceland seventh. Here, the influence of the Scandinavian model of the egalitarian social state can be seen. The next group comprises countries that were under Communist control after the Second World War and can be found between places four and 13 in the ranking; the former East Germany can also be found here. The reason for this closed group of countries may stem from the relatively egalitarian gender roles of Communism. 62 Christian Bischof TABle 2.01 Share of Housework Completed by Women in % and Relevant Indicators in 22 European Countries Welfare state regime ☐ ☐ ☐ * * ☐ * * * ☐ * ◊ * * ∆ Female employ- ment rate % 70.5 71.6 65.6 60.0 72.2 50.9 56.0 68.1 78.8 50.7 65.6 60.5 46.2 65.8 GEM 0.833 0.861 0.853 0.455 0.608 0.932 0.599 0.615 0.816 0.866 0.560 0.755 0.603 0.610 0.844 GINI 25.0 24.7 26.9 28.1 35.8 25.8 25.8 25.4 28.3 26.9 36.0 28.4 34.5 30.9 GNP 28335 32136 27956 5664 13107 39535 14060 16265 28104 31074 15342 27777 20450 11428 28590 Index gender role attitudes 3.5 3.6 3.4 2.7 3.0 3.4 2.9 2.9 3.2 3.4 2.8 3.1 3.1 2.8 3.3 Working time women in hours 36.3 35.4 38.2 40.0 40.3 32.9 40.4 42.0 37.7 35.9 41.6 30.8 43.2 43.0 26.9 Working time men in hours 42.7 42.0 43.8 45.8 43.3 42.2 44.4 47.2 45.5 50.9 45.5 44.2 45.6 49.0 42.8 Female share household income 3.4 3.5 3.4 3.3 3.5 3.3 3.4 3.4 3.4 3.2 3.4 3.0 3.6 3.1 2.8 Female Housework % 58 62 63 64 64 65 65 66 66 66 68 69 69 71 71 n 1 114 81 1224 979 1284 1080 1076 1674 622 373 895 1001 808 1023 1143 ) iN A ST e B lO l S DK F u e N SK CZ D( iS H g S P N ast) epublic (e r kraine stonia ermany nited Kingdom Countries Sweden Denmark Finland u e Norway Slovakia Czech g iceland Hungary u Slovenia Poland Netherlands Division of Housework in european Societies 63 Continuing TABle 2.01 Welfare state regime ∆ ∆ ∆ ∆ ∆ ∆ ◊ ☐ ☐ ☐ ☐ Female employ- ment rate % 52.6 60.7 60.5 70.3 57.4 51.9 56.5 48.3 61.7 45.2 24.3 GEM 0.855 0.815 0.816 0.797 0.753 0.776 0.681 0.614 0.289 GINI 33.0 29.1 28.3 33.7 32.7 34.3 34.7 38.5 34.3 43.6 ) * GNP 28662 30781 28104 29271 27047 61816 35879 22835 18428 19563 6949 est) Index gender role ast/We attitudes 3.2 3.0 3.0 3.0 3.2 3.0 3.1 3.0 2.9 2.7 2.4 3.0 ), post-socialistic (☐ Working time ermany g ern ( women in hours 33.5 33.9 30.6 30.7 35.2 34.8 32.7 38.3 38.5 42.0 46.3 36.5 Working time men in hours 43.0 44.1 43.6 45.8 42.4 45.1 44.5 45.0 42.3 47.2 52.4 44.8 Female share ermany-Microcensus ( household income 3.0 3.0 2.7 2.7 3.1 2.6 2.6 2.4 3.0 2.4 1.6 3.0 g ◊), mediterranean-south Female ∆), liberal ( Housework % 72 73 74 75 75 76 78 80 85 87 90 international Monetary Fund 71.2 n eport 2004 S $ 2004: eport 2004 r 1103 1199 1128 1200 1115 1013 1347 999 1207 1497 1192 u r ), conservative (☐ r B A D(W) CH F l irl e P gr T est) per Capita in PPP/ Human Development (W DP Human Development g NDP NDPu European Social Survey 2004, own calculations i: u ermany reece urkey N M: Countries Belgium Austria g Switzerland France luxembourg ireland Spain Portugal g T Overall est: egimes: social-democratic ( Sources: gi ge Employment rate: Eurostat 2004, Federal Statistical Office r r 64 Christian Bischof The countries thus far all remain under the Europe-wide average of 71.2%. The next group comprises the Central European countries. The final group is composed of the Southern European countries, with Turkey in the last position, having the female share of housework reaching 90%. The next step was to seek out the correlations between the share of housework completed by women and the other factors. To do this, regressions were calculated between the female share of housework and the macro indicators for the countries under consideration. The calculation of the factors brings in part only very low coefficients of determination: GNP (r²=0.01), GEM (r²=0.15). Thus the assumed correlations cannot be verified. Only for the employment rate, with r²=0.43, is there a clear correlation. Figure 2.01 shows the results for the analyzed countries. Here a clear cluster is formed by the Northern European countries of Iceland, Norway, Denmark, Sweden and Finland. Figure 2.01 Correlation Between Female Housework and Female Employment Rate ‹  ‰ ­ € €ˆ  ‚ ƒ † ‡ ­€ € Š €„ ‰‡  ‚ƒ       These countries feature a high proportion of women engaged in the workplace combined with a low share of domestic duties assumed by women.Turkey can be found far behind at the other end of the regression line. An additional cluster is formed by the Southern European countries of Greece, Portugal and Spain. In the middle, one can find a group of Central and post-socialistic countries, in which the latter form their own small cluster. Division of Housework in european Societies 65 Thus the macro indicators have proven to be not very well suited for answering the problem. The reasons for the large differences within Europe must be explained by other means. To do this, the individual values of the European Social Survey data set were aggregated using the calculated average and once again the regressions were calculated. Variables come into play in this analysis that are seen as relevant on the basis of theoretical consideration: first, the division of the household income between the man and woman; second, the temporal resources, indicated by the weekly working hours of both the man and the woman; and third, the index of gender roles. Both working hour values yield only extremely small correlations. A markedly stronger correlation is found with the index of gender roles, where the coefficient of determination reaches a value of r²=0.37. The same cluster that is found in the correlation with employment rate is also found here. A very strong correlation is found between the female share of housework and the female share of household income, reaching a coefficient of determination of 0.72. Figure 2.02 presents a scatter plot of the 22 countries. The values from the answer choices are scattered between one (the woman contributes nothing) and four (the woman contributes approximately half). The countries, with the exception of Greece and Portugal, lie quite clearly along the regression line, where once again the same sequence can be found as was found in the previous analysis. Turkey displays the highest proportion of household work performed by women with the least amount of income contribution. Again, the Southern European countries come next, followed by the Central European ones, succeeded by a group of post-socialistic countries and ending with the Northern European countries. Here one can see an even clearer correlation than that with the employment rate on the macro level. A high employment rate seemingly leads to women demanding more cooperation with the housework from their partners. Yet an egalitarian division of housework only appears when the female income is high enough, nearing that of the men. Through an average income comparable to that of men, women reach a position of economic independence that allows them more power in negotiating the division of household work. However, women can only arrive at a similar income level if they have the same career possibilities as men. As the present results show through the aggregated data from the country level, the share of housework assumed by women differs greatly, and can be divided into roughly three clusters of countries: the Scandinavian and post-socialistic countries, the Central European countries, and the Southern European countries. Turkey does not fit within any of these groups. Women perform the least amount of household work in countries where the female share of the household income comes close to 50%. 66 Christian Bischof Figure 2.02 Correlation Between Female Housework and the Female Share of Household Income Ž ‹ „ Œ  ‚ †‰‡ ­ ƒ  Š „  ‹Š ˆ ­€ †‡    ‚ƒ   ‘ The Gender Empowerment Measure Index, deemed important by the theoretical presuppositions, has proved to be insufficient in explaining the share of female housework. The clearest correlation between all of the investigated variables can be found between the GEM value and the gender role attitude index, reaching the high value of r²=0.82. Figure 2.03 shows the correlations between these variables and the 22 analyzed countries. Here similar groups can be seen, though this time the Eastern European countries are not grouped with the Central European countries, but instead with the Southern European ones. Determining the causalities of these two factors is quite difficult, and perhaps unnecessary, but it can nevertheless be seen how well the two factors, the GEM for the macro level and the gender role index for the individual level, mirror each other. Societies in which the household work is more nearly completed by both partners also simultaneously display maximum values for these indicators. Four groups have consistently been formed through analysis on the macro level and through the analysis of the aggregated individual data, respectively. The groups are the Scandinavian, or rather the Northern European countries, the Central European states, the post-socialistic nations and Southern European countries, where Turkey always assumes a special role and maintains a significant distance from all the other countries. These groups correspond to a wide-ranging classification of liberal welfare states (marked with ◊), con- Division of Housework in european Societies 67 servative-corporate (∆) and social democratic (☐) regimes [Esping-Andersen, 1990]. This classification is oriented to the relationship between state and market, which emerges for the allocation of resources and incomes and through this influences the social structure. The regimes can be briefly characterized through the following: the liberal welfare state (Great Britain, USA, Ireland, etc.) is defined by influence exercised by the markets and less by the government; the conservative-corporate type (Austria, France, Germany, etc.) is characterized by a maintenance of class and status differentiation; the social democratic regime (Scandinavian countries) is distinguished by a strong, egalitarian underlying tendency. Figure 2.03 Correlation Between GEM and Gender Role Attitudes Attitudes This classification was initially not oriented towards gender relations; in a more recent study, gender roles and the family have also been included [Esping-Andersen, 1999]. The social-democratic regime also promotes gender equality, both in the workplace through equal pay, and also in the private sector [Sainsbury, 1996]. In the liberal regimes, the underlying ideology does not allow for differences between individuals; thus also between men and women, one finds no state support concerning this topic [O’Conner, et al. 1999]. Conservative regimes, in contrast, support traditional gender roles and are typically shaped by Catholicism; the family is a core unit of society with an organization based on division of labour. The male-breadwinner model plays an important role, where the man is responsible for financial provisions 68 Christian Bischof and the woman for domestic organization [Sainsbury, 1996]. Taken together, the exercise of influence in the welfare state regimes occurs on three levels: in the sphere of the exercise of influence as an ideological framework and as the general conditions for redistribution [Daly, 1994]. For this work, it can be inferred that the welfare state creates the frame for the specific arrangement of domestic division of labour between the genders. Thus the constraints on female employment create a higher incidence of female housework. In contrast, the gender-equitable educational curriculum promotes an egalitarian understanding of male and female roles, which then presents the prerequisites for an equitable division of housework [Geist, 2005]. The typology of Esping-Andersen has been expanded by a few authors, thus the Mediterranean-Southern European (marked with ☐) welfare type has been added. These countries are not assigned to the conservative regimes here, because the welfare state of these countries first developed late and is relatively rudimentarily pronounced, as for example in Spain and Italy. An additional type can be seen in the post-socialistic states (*), which have, admittedly, grown closer to the liberal or conservative models in recent years, but are nevertheless still strongly shaped by their Communist heritage. Even in the question of gender equity, these countries are of great interest, since here female employment was desired by the state and there were large numbers of childcare institutions. Following the change of system in the late 1990s, this became a large financial burden for the populace and the state promotion of childrearing became limited [Deacon, 2000]. From the point of view of the concept of the welfare state regime, the results displayed in the diagram are clearly anchored in these typologies. It appears as if the regimes would be a good point of departure in explaining the differences between countries regarding the division of housework between genders. This becomes especially applicable if the data for West and East Germany is considered. Although both units exhibit the same present conditions of a governmental general framework, they clearly show differences. The former ideological shaping of the East German populace can still be seen as relevant to the egalitarian understanding of roles in society. It shows that the different socio-political structures are of great importance: “Male dominated hierarchy of the political economy, existing ideologies on gender, and birth cohort all affect the amount of power a woman can derive from her earnings [Coleman, 1990: 252].” The social norms imparted by society determine the behavior of men and women even more strongly than individual mindsets. The following chapter seeks to illuminate the interdependence between individual dispositions and the macro structure. Division of Housework in european Societies 69 Explanations on the Individual and National Level In the analysis thus far, the individual data has only been used in aggregated form. In the next step, this data will also be analyzed on the individual level in combination with the national context variables. To this end, a multi-level analysis is conducted, by which the micro attributes from the individual level are connected to the macro factors from the national level. Several random intercept models are calculated; in these models, the constants of the individual countries vary on different levels. In this form of multiple-level analysis, it is assumed that the connection between the female share of the housework and the explanatory variable is the same for each country. It is, however, assumed that the level of the housework shared in each country is different; the previous analysis based on the aggregated data has already confirmed this difference in level. As the first step in the analysis, a multilevel model is calculated, which simultaneously defines the variance in the division of housework on the national and individual levels. This model is called the “empty” model; here still no explanatory variables emerge in the calculation – it contains only constants. 28,617 people in 22 countries are included in this analysis. The results of the “empty” model are cited in Table 2.02 under Model 1. The variance on the individual level amounts to 467.3, thus 91.3% of the variance of differences in household work allotment is explained on this level. The variance on the national level amounts to 47; 8.7% of the variance is explained on the national level. The value of 8.7% is also referred to as the Intraclass correlation coefficient (ICC), which means that 8.7% of the unexplained variance in household work allotment can be ascribed to the people living in different countries. The intraclass correlation is the relationship of the variance between the countries with the variance within each country. An analysis on multiple levels makes sense only if a difference between the macro units is identifiable. Though the differences on the national level are relatively small, a hierarchical multi-level analysis seems nonetheless justified. In the next step, the explanatory variables on the individual level are taken into the model; the values can be found under Model 2. 19.6% of the variance on the individual level can be explained through the incorporation of these variables. The integrated variables explain a decidedly higher percentage of 65.5% on the national level. The explanatory power of the added individual variables is thus significantly higher on the national level than on the individual level. As an explanation for these figures it can be assumed that the populaces distinguish themselves on the national level so much regarding their composition and their response behavior that these large differences can be reproduced on the national 70 Christian Bischof level. Both of the last values show how large the share of the explanation of this model is for the entire variance – thus how much can be explained through the integration of additional variables for both levels. These values are designated as effect size. Through the additional variables, the variance is reduced both on the country level as wel as on the individual level. Through the individual variables, the –2 Log Likelihood value is reduced by 1921 to 76611; this value is significant at the 5% level. This model attains a higher predictive ability for the share of female housework than the model without the individual variables. In the next model, the variables from the country level are now also introduced. These are the values for the Gross Domestic Product, the GINI Index, the Gender Empowerment Measure Index and the female employment rate. The –2 Log Likelihood value is once again reduced and is likewise significant; thus this model also enables a higher prognosis of the dependent variable than the previous model. The effect size remains the same on the individual level and increases on the national level to a value of 84.9%. Through this model, a reduction of the variance from 47 of the “empty” model to 6.9, thus 84.9%, can be reached by means of the variables on the national level. As the variance on the individual level is, however, notably higher, the macro indicators have collectively only a very minor explanatory power. Thus the GINI value is the only indicator on the macro level that turns out to be significant. With a lower GINI value, as well as a lower disparity in wealth, the share of female involvement in housework also sinks, with the other independent variables held constant. Model 4 additionally allows for the different welfare state regimes – Table 2.02 gives an overview of the classification of the individual states. Once again, no additional variance on the individual level can be explained through these indicators, though on the national level, the explained variance climbs 12.5 percentage points to 97.4%. The variance on the national level only amounts to 1.2 more, also the –2 Log Likelihood value is reduced to a significant level, thus this model also better explains the problem itself. The influence of the welfare state regimes with the social democratic type as a reference category is noticeable in the following way: the conservative and Mediterranean regimes exhibit a significant effect, thus the female share of the housework increases in the conservative regime by 4.4 percentage points and by Mediterranean even more clearly with 9 percentage points. On the individual level, on the other hand, all of the variables except for the duration of the cohabitation and the Protestant and Islamic religious groups are significant. In order to enable the comparability of the variables measured on different scales, the value of each explanatory variable was subjected to a Z-transformation. The factor with the greatest influence proves to be the respondents themselves; when the respondent is female, the share of the Division of Housework in european Societies 71 female housework increases by 12 percentage points. Thus there is a large discrepancy, depending on whether the information is provided by a man or a woman. The country with the least difference in this regard is Poland, with a difference of 3.7, while the greatest difference occurs in the Czech Republic, with 14.4. The Mediterranean welfare state emerges as the second most influential factor, with an increase in the female share of 9 percentage points. In the next position appears the female share of the household income with a standardized coefficient of 0.227; when the value of the answer scale shifts a category, for example, from “less than half ” to “about half ” – thus when women start providing for more of the household income – the female share of the housework shrinks by 4 percentage points. The fourth greatest value, with a standardized coefficient of 0.175 appears as the conservative regime, with an increase in the female share of 4.3 percentage points. In fifth place emerges the Orthodox Christian faith: people who replied as belonging to this religious community display a 3.8 percentage point higher share. Education appears as an additional important value; if the respondent graduated from high school, the share sinks by 3.5 percentage points. 2.5 Summary The explanation for the strongly differing share of housework for women might thus be found primarily on the individual level. The greatest differences on the national level emerge only through individual characteristics and not through macro factors. Natural y, this stil does not mean that the respondent exhibits this behavior simply from him or herself, that is, independent from the social context to which he or she is connected on many levels. This results, consequently, in the conclusion that individual behavior cannot be explained through the context factors used here. The objective of further studies must be to be able to better analyze the connection between the micro and macro levels and to discover the factors that function as the connecting link between the two. 72 Christian Bischof TABle 2.02 Multi-level Model of the Effects on the Female Share of Household Work 11 0.407 0.070 -0.510 -0.130 0.097 0.063 0.028 0.158 0.065 -0.226 0.122 -0.080 -0.1 -0.058 0.046 0.001 stand. coefficients ** * ** ** ** * * ** ** ** ** ** * 91.973 1.685 -3.130 2.326 1.512 0.678 3.808 1.602 -3.918 0.224 -0.150 -4.209 -0.572 0.073 0.002 Model 4 -12.253 ** * ** ** ** * * ** ** ** ** ** * 79.366 1.666 -3.438 2.303 1.504 0.827 4.375 3.891 -3.940 0.224 -0.152 -4.131 -0.531 0.075 -0.002 Model 3 -12.254 ** * ** ** ** * * ** ** ** ** ** * 94.293 1.660 -3.514 2.315 1.817 0.774 4.383 3.614 -3.952 0.223 -0.154 -4.132 -0.516 0.075 -0.004 Model 2 -12.262 ** 69.816 Model 1 ef.: without)r oman Catholic astern Orthodox espondent (ref.: female) ducation (ref.: without r e islam orking time men orking time women Constant Coefficients individual level Married (ref.: no) r e matriculation) Child in household(ref.: no) Confession ( Protestant ears of cohabitate Female share household income W W index gender role attitudes Household income Age Y Division of Housework in european Societies 73 Continuing TABle 2.02 0.270 0.075 -0.122 -0.017 -0.150 0.175 0.089 0.368 21038 stand.coefficients 26 * * * 19.6 97.4 * 1.2 0.364 -0.030 -3.337 4.364 2.188 9.023 376.1 76568 Model 4 -19.124 0.000568 eport r * * 1 6.9 22 0.636 19.6 84.9 -5.946 -0.134 376.1 7661 Model 3 0.000335 Human Development * NDP 1 i: u 18 Ngi 376.1 15.7 19.6 65.5 7661 Model 2 3 European Social Survey 2004, own calculations 467.3 45.7 78522 Model 1 ICC 8.7 est); Rest: index Germany East/W ariance Human Development Report 2004; Employment rate: Eurostat 2004, Federal Statistical Office per Capita in PPP/US $ 2004: International Monetary Fund; mpowerment UNDP e M: i-index ge NP N ender elfare-regime (ref.: soc. dem.) ariance individual level ariance country level Country level Coefficients country level g gi g Female employment rate W liberal conservative post-socialism Mediterranean V V Percentage of V explained: Individual level -2 Log-Likelihood Numbers parameter ermany-Microcensus ( * Significant p<0.05; ** Significant p<0.001; Dependent variable: female housework Sources: GDP 2004; g Attitudes towards Immigrants and the Effect of Migration in European Countries 3 Jana Chaloupková | Petra Šalamounová 3.1 Introduction International migration is a hot topic for both scholars and the general public. Additionally, the issue of migration has come to the forefront in the debate over the demographics of developed countries with their ageing and declining populations. Together with the continuing processes of globalization, the intensifying social relations among various nations, states and regions of the world, and the growing mobility of information, capital, goods and people, even European countries, which have been ethnically and culturally fairly homogenous in the past, are now beginning to face an influx of migrants, who often come from very different social and cultural backgrounds. As the intensity of migratory flows can be expected to increase further in the future, even inhabitants of those European countries that have stood outside of the main migratory flows will begin to face inflows of migrants. Data from the European Social Survey (ESS)1 allow us to compare the attitudes of people in twenty European countries towards migration. The first round of this international survey was carried out in 2002 and included an extensive module dealing with attitudes towards migration and migration policies. The main emphasis was on the perception of how migration impacts host countries.2 This paper will focus only on three of the topics covered by the survey. First, we are interested in knowing how the public in individual European countries perceives the inflow of migrants, and second, how the public views the social, cultural and economic impact of migration on the host country. Furthermore, we will examine the attitudes of people in European countries towards various forms of integration of immigrants in their country. This paper will attempt to establish whether there is any connection between attitudes towards migration on the part of the inhabitants of individual European countries and the experience of the given country 1 The European Social Survey took place in a total of 22 countries. This paper does not analyze data for Israel, which is not usually considered a European country, or for Slovenia, which is not a member of OECD and thus some information about migration pertaining to this country was not available. The data file covering 20 European countries includes a total of 38,341 respondents. ESS surveys apply high research standards in the area of quantitative research, and respondents are selected strictly by a probability method. The minimum sample size is approximately 1,500 respondents aged 15 and over in each of the participating countries. For more detailed information, please visit http://www. europeansocialsurvey.org. 2 Effects of migration may be felt even in the home countries of the migrants. For example, the public in less affluent European countries may be concerned about the structure of the people leaving their home, as they will include a disproportionate number of the educated elite. However, attitudes on this issue were not the subject of this international survey. 78 Jana Chaloupková | Petra Šalamounová with immigration and the overall economic situation of that country. The difficulty of this approach lies in the fact that the number of immigrants in individual European countries is not known precisely and that, in addition to the registered immigrants, many people reside in host countries illegally. In the first part, we provide a brief overview of both the historical and the more recent developments in the area of migration in Europe, and – owing to problems of reliability and comparability of data on international migration – we also offer an account of our sources. 3.2 Migratory Developments and the Current Situation in Europe During the 20th century, Europe changed from a region of emigration to a region attracting immigrants [Martin, Widgren, 2002].3 Concerning their experiences with international migration, European countries may be divided into two groups4 [Bauer, Lofstrom, Zimmermann, 2000]. The “old” countries of immigration in Europe include primarily those with postcolonial immigration ( e.g. UK, France and the Netherlands). Another group of countries with long-term experience with immigrants includes those countries of Western Europe that experienced an economic boom in the 1960s and sought foreign guest workers from less developed countries in Southern and Southeastern Europe, Morocco, Turkey and elsewhere to meet an acute labour shortage. Labour migrants primarily chose Germany, but also France, the Netherlands or Sweden. The original intention was for foreign workers to remain only temporarily. However, although the economic recession of the 1970s limited the job opportunities for migrants, most of them decided not to return to their home countries but instead to bring their families to join them in their new country. The second group comprises the European countries having no historical experience with migratory inflows that have been transformed from sources of emigration into destinations for immigration. These “new” immigrant countries include Italy, Ireland and Spain. Beginning in the late 1970s, labour 3 For an outline of the developments in the field of international migration in Europe after World War II, see Martin, Widgren, 2002; Zimmermann, 1995; Bauer, Lofstrom, Zimmermann, 2000; see also Rabušic, Burjanek, 2003: 15. 4 From a global point of view, the traditional immigrant countries are the United States, Canada, Australia and New Zealand. Immigration played a crucial role in the formation and development of these countries, and they remain open to inflows of immigrants to this day. Attitudes towards immigrants and effect of Migration in european Countries 79 emigrants started returning to these countries, and beginning in the 1980s, foreign migrants started arriving [Bauer, Lofstrom, Zimmermann, 2000]. The early 1990s saw a considerable increase in migratory flows across Europe, bringing more refugees and asylum seekers. The fall of communism in Central and Eastern Europe resulted in significant changes [Martin, Widgren, 2002; Bauer, Lofstrom, Zimmermann, 2000; see also Rabušic, Burjanek, 2003: 15]. Since the 1990s, countries in Central Europe, which before the fall of communism were often countries where international migration originated,5 have been, thanks to their geographical location, becoming transit countries and also increasingly destination countries [Wallace, 2002]. Sources of European Data on Migration Comparing the size of the migratory flows into individual European countries is complicated not only by the multitude of methods used to determine the number of foreigners, but also by the different interpretations of who is to be considered an immigrant. Generally speaking, foreigners include people without citizenship of the country in question.6 According to the broadest definition used by the United Nations, a migrant is a person living outside of his or her country of origin for more than one year. Consequently, migrants are not only those who have permanently settled in a foreign country, but also those who have come to a foreign country for a limited period of time, e.g. to work or study.7 The boundary between temporary and permanent residence in the given host country is not always clear. The method of determining the number and structure of immigrants depends on the practices used in the given country. One of the best sources of data for pan-European comparison is the information from the Organization for Economic Cooperation and Development (OECD), obtained from national coordinators. However, even this information is based on different sources, and thus its quality is not always fully analogous. The most common source of data on the number and structure of migrants used in most Western European countries 5 International migration is not an entirely new phenomenon for the former communist countries. However, international migratory flows were strictly limited and regulated before the fall of communism [see Václavíková, 2002]. 6 This distinguishes migrants from members of national minorities who are citizens of the given country and set apart from other citizens usually only by their ethnic origin, language, distinctive culture and traditions. 7 According to John Salt [2001], migrants include those leaving their home country for seasonal work or regularly crossing the border for work. 80 Jana Chaloupková | Petra Šalamounová is the population register. In Southern and Eastern Europe and Switzerland, data from successful applications for permanent or temporary residence permits are used. Other sources of information about migrants include censuses and social surveys.8 Because legislation covering the residence of non-citizens in foreign countries ( e.g. visa requirements) differs from country to country and also according to the country of origin of migrants, official statistics may underestimate certain categories of foreigners. Furthermore, such statistics do not include foreigners who have already obtained citizenship of the given country. The conditions under which migrants may obtain citizenship of their new homeland and the length of a waiting period differ among European countries. Different rules also apply to the entitlement to citizenship for children of migrants born in the host country. Refugees, i.e. persons fulfilling the criteria of the Geneva Convention on the Status of Refugees, fall into a special category of migrants. Statistics regarding their number and structure are usually separate and prepared based on information from relevant government ministries.9 Apart from legal foreign residents, all European countries are home to illegal migrants. The total number of illegal immigrants living in Western Europe at the beginning of the 1990s was estimated to be between 3.5 and 5.5 million [Drbohlav, 2001b]. The number is likely to be higher today. For comparison, the total number of foreigners settled in Western Europe at the end of the 1990s was around 20 million [ ibid]. It is therefore clear that illegal immigration accounts for a considerable portion of total migratory flows. The share of illegal immigrants is not distributed evenly across the individual countries. For example, it is estimated that there is only a small group of illegal immigrants living in Scandinavia, while the number of illegal immigrants in Southern Europe significantly adds to the total number of immigrants. The main reason for this is that the Scandinavian countries are geographically too distant for migrants from less developed countries, and compared to Southern Europe, these countries have a more advanced system of registration and regulation of migration [Drbohlav, 2001b]. While it is apparent that numbers of illegal 8 OECD data provided in Table 3.01 have been collected using the following sources: data for Belgium, Denmark, Finland, Luxembourg, Germany, the Netherlands, Norway, Austria and Sweden are based on population registers. Data for the Czech Republic, Italy, Hungary, Poland, Portugal, Spain and Switzerland are based on statistics of government ministries pertaining to residence permits. Data for Greece and France are based on censuses, data for Ireland and the United Kingdom are based on workforce surveys. 9 Some countries such as Germany and the Netherlands also include among migrants asylum seekers not living in facilities designed for them. Norway includes all asylum seekers among migrants, regardless of where they live. Attitudes towards immigrants and effect of Migration in european Countries 81 immigrants may significantly influence attitudes towards international migration, it is not possible to carry out a more extensive international comparison due to the great difficulty of estimating the number of illegal immigrants and the lack of comparable data regarding the extent of illegal migration in individual European countries. Overview of Data Regarding Migration in the European Countries Under Consideration Today, the largest numbers of non-citizens live in countries that have a long history of experience with immigration (Table 3.01). Luxembourg is rather particular in this respect, as every third inhabitant is a foreigner. Switzerland also has a high share of immigrants, making up one-fifth of the country’s population. Germany, Austria and Belgium have a relatively large percentage of immigrants within the total population (approx. 8 or 9 %). Greece follows with 7% of its total population composed of immigrants, but is an unusual case because, during almost the entire the 20th century, it was a country of emigration. Generally speaking, there is a higher share of migrants and asylum seekers in the wealthier European countries.10 Southern Europe (except Greece) has a relatively low share of legal residents from other countries. Hungary, the Czech Republic and Poland, all post-communist countries, have the smallest share of immigrants in regards to the total population because they only opened their borders beginning in the 1990s. According to official statistics, 1 in 50 inhabitants of the Czech Republic is not a Czech citizen. The official number of migrants in Poland is less than 0.5 % of the country’s total population. Finland has a similarly low share of non-citizen population. In addition to the share of immigrants within the total population, their structure based on country (or region) of origin, education, religion and reason for emigration is also important. Data that can be used for a pan-European comparison are available based on country of origin. One problem with such a comparison, though, is that the OECD only publishes information on the countries of origin for the largest groups of migrants moving to the host country in question. Data for less common countries of origin have been grouped under the heading ‘other countries.’ Because each country has a specific structure of immigrants due to its geographical location and history ( e.g. people coming from former colonies), it was necessary for the purpose of an international comparison to categorize countries of origin into a number of groups. 10 There is a significant correlation between the share of migrants and GDP (Table 3.04). 82 Jana Chaloupková | Petra Šalamounová TABle 3.01 Share of Immigrants in Population and Structure of Immigrants According to an Area of Origin (%) did Other 22.1A 12.7 17.7 28.6A 11.2 0.1 22.2 30.7A 21.7 35.6 34.7 25.1 38.6 26.3A 29.7 9.8 11.0A 24.1 23.6 19.7A countries which in . latin * * * * * 0.4 * 2.1 * 4.7 * * * 2.2 * * * * 2002 counties America 21.5 12.1 in eu the seekers * * * * * * * from Africa 13.1 2.1 34.8 2.0 5.5 4.9 15.3 3.5 21.2 19.5 45.2 4.4 1.0 asylum of Asia * 5.6 . 14.3 26.1 5.4 11.5 6.4 14.6 27.7 13.4 16.0 13.6 * 3.3 13.6 1.7 12.9 12.8 8.4 8.2 immigrants exist. r Numbers SS did * * * 2.1 * 5.3 * * * * * * * * * * 7.7 Former u 32.9 34.9 41.8 (1999). immigrants France of Former * 24.2 35.4 13.4 1.2 0.7 * 5.6 13.1 1.9 1.8 8.2 * * 2.7 * 2.4 4.0 9.3 1.5 and immigrants according to an area of origin (total 100 %) ugoslaviaY (2002) structure ast the e * * * * * * * * * in Structure of urope 4.3 1.0 3.3 2.2 2.2 6.3 e 16.3 66.8 18.4 36.1 39.6 Kingdom nion u .edu.pl/osrodki/cmr/wpapers/pdf/056.pdf nited u .iss.uw 77.9 57.5 16.2 25.5 68.0 11.8 36.6 41.7 29.9 39.5 32.3 49.5 61.4 25.5 14.0 31.1 4.7 19.8 11.0 21.5 countries uropean Developed e Poland, the immigrants in the particular country are included in the category „other“. from Ireland, , 000 23.5 36.0 46.0 8.6 18.2 5.1 8.6 37.0 12.4 18.7 11.6 38.5 29.7 1.5 1.3 0.2 8.3 6.6 6.3 1.3 incomes Number of asylum seekers per 10 inhibitants Germany for is cell .oecd.org/document/36/0,2340,en_2825_494553_2515108_1_1_1_1,00.html; http://www 100_SAP_5_euro03.htm; http://www xception special 37.5 19.7 9.4 8.9 8.2 7.0 5.6 5.3 5.0 4.4 4.3 4.1 3.9 2.7 2.4 2.2 2.0 1.9 1.1 0.1 e. no Share of immigrants (%) A 2001 letter ep. from the r by OECD; http://www Data ermany reece .K. luxembourg Switzerland Austria g Belgium g France Sweden Denmark u Netherlands Norway ireland Spain italy Portugal Czech Finland Hungary Poland Note: labelled not belong to the most usual countries of origin of the Source: statistics.gr/eng_tables/S1 Attitudes towards immigrants and effect of Migration in european Countries 83 The first group includes migrants from EU member states (15 members states at the time the data for ESS was gathered11), EFTA member states, the United States and Canada. The second group includes migrants from former communist countries in Central and Eastern Europe. We list the migrants originating from the former Yugoslavia and the former USSR separately. Other migrants are listed by continent of origin. In general, the wealthier countries in Europe have a higher share of migrants from Western Europe and the United States than the less developed countries (see Table 3.04). The largest number of migrants from post-communist countries (including the former Yugoslavia and the former USSR) is to be found in the former Eastern bloc countries as wel as in neighboring countries ( e.g. Greece, Finland, Austria). African migrants also tend to move to nearby destinations, in particular Southern Europe and France. The largest group of migrants in Germany are Asian, in particular, Turkish. Only four of the twenty European countries examined (Luxembourg, Ireland, Belgium and Switzerland) have a predominant share of immigrants from developed countries.12 Asylum seekers represent an important category of migrants. During the past decade, the number of asylum seekers has changed significantly in the individual countries. The most marked growth in the number of asylum seekers in the past ten years has occurred in Austria and Ireland. On the other hand, Germany has experienced a sharp drop in the number of refugees due to the enactment of stricter legislation. The number of asylum seekers in Poland and Southern Europe remains low. In 2002, Austria had the highest relative number of asylum seekers, with almost 50 applications per 10,000 inhabitants, followed by Sweden, Norway and Switzerland (over 30 application per 10,000 inhabitants). The number of applications per 10,000 inhabitants in the Czech Republic in 2002 is comparable to those in Germany and France. 3.3 Attitudes towards Immigrants in European Countries Previous international comparisons have shown that the populations of the post-communist countries have more negative attitudes towards immigrants than people from Western Europe, and that the level of intolerance towards 11 New EU member states have not been included in this category because their historical development in the past 50 years is different. It is also important to note that their accession to the EU was often seen as worrisome due to the potential growth of labour migration from these new member states to Western Europe. 12 It is very likely that the number of migrants from less developed countries would significantly increase in many European countries if il egal immigrants were included in official statistics. 84 Jana Chaloupková | Petra Šalamounová minority groups in post-communist countries is also higher than in Western Europe [ e.g. Rabušic, 2000; Burjanek, 2001; Wallace, 2002]. Unlike Western Europe, countries in Southern, Central and Eastern Europe have no long-term experience with migration, and it may be assumed that they do not see themselves as countries needing to accept immigrants. However, with the growing support of political parties wanting to toughen immigration policies, the increasing number of violent attacks against immigrants, as well as growing social unrest among immigrants within the past decade in many Western European countries, the issue of immigration is also becoming an extremely sensitive one for the populations of these countries [Martin, Widgren, 2002]. These feelings may be further strengthened in Western Europe by concerns related to the inflow of labour migrants from the new EU member states as well as from other countries that could join the EU in the future. Different attitudes towards immigration across Europe are the result of a number of social, economic, political and cultural factors, e.g. the character of the political culture or the level of openness of the society.13 We may assume, however, that feelings about immigration may also be influenced by the number and structure of immigrants in the given country. Many studies have focused on the connection between attitudes towards immigration and individual social and economic characteristics of the respondents and, in particular, on their economic status, education, age or socio-psychological characteristics. Such studies agree that people with higher education see migration in a more positive light. On the contrary, less educated people with a low economic status tend to see migrants more negatively because migrants represent competition on the job market. According to other explanations, better educated people are more tolerant towards immigrants and express fewer ethnic or racial prejudices and stereotypes. In comparing attitudes towards immigrants in 12 OECD member countries, Bauer et al. did not confirm that differences in attitude in the individual countries could be explained by the different effects of individual characteristics of the respondents, such as education or economic status [Bauer, Lofstrom, Zimmermann, 2000]. However, other studies have shown that better educated people in wealthier countries are more open to migration, while the correlation be-13 For example Rabušic provides evidence that the unwillingness to accept foreigners and the symbolic social exclusion of foreigners in general are connected to the level of interpersonal trust. Rabušic assumes that if a person does not trust people in general, he or she is not very likely to tolerate “others” in the neighborhood. In countries where intolerance towards immigrants is higher, fewer people believe that most other people can be trusted. People living in developed democratic countries show higher levels of interpersonal trust [Rabušic, 2000]. Attitudes towards immigrants and effect of Migration in european Countries 85 tween education and positive attitude towards migration in countries with smaller GDP is negative [Mayda, 2004]. Similarly, Clark and Legge, who have examined attitudes towards immigrants in Germany, concluded that economic status influences attitudes towards immigrants in Eastern Germany more profoundly than in Western Germany [Clark, Legge, 1997]. Only a few studies, however, have analyzed the correlation between attitudes towards immigrants and the character of immigration in individual countries and the broader economic and social situation. One of the main reasons for this is the lack of international y comparable surveys that focus on the acceptance of immigrants and the perception of the impact of migration, coupled with the relative complexity of comparing international migration data. Openness (or Lack Thereof) towards Immigrants The ESS survey asked respondents to specify the extent to which their government should allow immigrants to come to their country. The advantage of the ESS survey is that it examines the willingness to accept immigrants from different countries of origin. The questionnaire contained separate questions regarding the willingness of people to accept people from six categories of countries of origin: 1) people of the same ethnicity/race, 2) people of other ethnicity/race, 3) people from richer countries in Europe, 4) people from poorer countries in Europe, 5) people from richer countries outside Europe, and 6) people from poorer countries outside Europe. Respondents could select from four different answers: 1) allow many to come and live here, 2) allow some, 3) allow a few, and 4) allow none.14 Although the percentage of those agreeing or disagreeing with how many immigrants should be allowed into the given country differs based on whether the immigrants come from a wealthy or poor country or an area with different culture, there is a high degree of correlation regarding the willingness to accept immigrants from various countries of origin. People willing to accept immigrants from wealthy European countries are also in favor of the immigration of people from poorer European and non-European regions. In contrast, those respondents who rejected people from poor regions also did not want immigrants from economically more developed countries. For simplification, it was possible to calculate a sum index demonstrating the total rate of openness to migratory flows (average response values for the six questions). 14 For the wording of the questions, please see attachment. 86 Jana Chaloupková | Petra Šalamounová The higher the index, the less the public was in favor of immigrants settling in their respective country.15 Table 3.02 shows the sum index values and percentages of respondents from the individual countries stating that their country should accept no or only a few immigrants from the analyzed categories. The inhabitants of Sweden and Switzerland were the most open. It is interesting to note that respondents from these two countries more often said that they would rather accept people from poorer regions than from wealthier regions. On the other side, Greeks and Hungarians showed the least willingness to allow any new inhabitants to enter their countries. There are numerous reasons for the unwillingness to accept foreigners in these two countries. Apart from both countries having become destination countries only recently, the structure of newcomers may also play a significant role. Moreover, negative attitudes towards immigrants may be fueled by the frequency of illegal migration and problems related to it ( e.g. unemployment, crime, etc.). More than half of respondents in Austria and Portugal also said that they would allow only a few or no immigrants into their country. Luxembourg, where every third inhabitant is a foreigner, has also shown a relatively low level of acceptance of new immigrants. Overall, Europeans prefer to accept immigrants of the same race (ethnicity) and also European immigrants over those coming from other continents. Perception of Social, Cultural and Economic Effects of Migration: A Benefit or Threat? It is usually assumed that the reasons for negative attitudes towards immigrants fall into two groups. The first group includes reasons of an economic nature. An inflow of migrants may cause the public to be concerned that salaries will fall and jobs will become scarcer for domestic workers. Inflows of migrants change the ethnic, racial and religious structure of the population in the destination countries, and for this reason, a growing number of immigrants may raise concerns about the destination country’s cultural identity being threatened. 15 Reliability coefficient (Crombach’s alpha) of the scale is 0.94. Values of this coefficient calculated for each of the examined countries separately are similar (ranging from 0.90 to 0.97). Attitudes towards immigrants and effect of Migration in european Countries 87 TABle 3.02 Openness of Society Towards Immigrants. Sum Index and Share of Answers (in Percentage) Allow a Few and Allow None in Categories 15.4 31.3 38.1 36.2 40.8 38.4 53.8 43.3 51.0 43.7 48.7 48.8 52.8 44.0 51.5 60.1 64.9 62.4 88.0 85.8 Poorer countries outside Europe 24.9 36.8 31.5 38.5 39.0 44.9 41.3 33.9 46.7 45.1 34.7 52.2 50.7 50.0 48.6 58.8 63.4 56.9 76.5 72.6 Richer countries outside Europe immigrants creates average of 13.3 26.9 34.7 31.7 35.8 33.3 43.5 41.5 48.8 38.1 46.0 42.6 48.8 41.6 46.8 53.6 61.1 61.0 83.6 83.5 Poorer countries in Europe 21.2 31.4 31.4 31.9 34.6 37.3 30.7 31.8 45.1 38.9 34.3 43.0 48.3 45.6 44.0 50.2 57.1 56.8 70.3 67.1 Richer countries in Europe . Sum index of willingness to accept 17.0 33.6 36.3 35.0 42.1 43.5 51.5 45.3 48.3 44.8 53.1 45.4 53.8 42.1 49.8 62.7 66.2 61.4 86.2 85.8 Different race/ethnic group 11.2 19.5 29.6 21.7 26.1 28.2 25.1 33.1 43.6 31.7 44.1 35.1 42.1 36.8 35.3 41.7 55.2 56.1 50.6 70.3 Same race/ ethnic group Acquired values on scale from 1 (allow many people) to 4 (allow none) - Sum in dex 1.92 2.18 2.23 2.24 2.29 2.29 2.30 2.31 2.35 2.40 2.42 2.46 2.46 2.46 2.50 2.52 2.59 2.74 2.88 2.91 ep. SS 2002 r e Total (100%) include answers allow some and allow many ermany .K. reece Sweden Switzerland italy ireland g Norway Denmark Poland Spain Belgium Czech France luxembourg Netherlands u Finland Austria Portugal Hungary g Note: answers on six monitored items. Source: 88 Jana Chaloupková | Petra Šalamounová The ESS measured the impact of migration on the destination countries using a battery of six questions.16 Respondents were asked to choose an answer on a scale from 0 to 10, with the lowest value denoting an attitude that migration has negative effects on the host country and the highest value denoting a belief that the effects of migration are positive: A. People who come to live here generally take jobs away from workers in [country], or generally help to create new jobs? B. Most people who come to live here work and pay taxes. They also use health and welfare services. On balance, do you think people who come here take out more than they put in or put in more than they take out? C. Would you say it is generally bad or good for [country]’s economy that people come to live here from other countries? D. And, using this card, would you say that [country]’s cultural life is general y undermined or enriched by people coming to live here from other countries? E. Is [country] made a worse or a better place to live by people coming to live here from other countries? F. Are [country]’s crime problems made worse or better by people coming to live here from other countries? First, we verified whether cultural and economic effects of migration are related. A series of factor analyses calculated separately for each of the twenty countries as well as for the entire data file have shown that all the above specified statements are fueled by a shared factor.17 It is, therefore, not possible to state that perception of economic and cultural threats represented competing dimensions of attitudes towards migration. Using these statements, a sum index of migration effects was construed for the purpose of further analysis (average value of responses to the six questions). The smaller the index, the more migration is perceived as a threat, while, the greater the index, the more the positive effects of migration are emphasized. Reliability coefficient of the scale (Crombach’s alpha) is 0.84.18 It is important to note that the average sum index values appear more in the bottom half of the 11-point scale. As a 16 This battery of questions contained a seventh question: ”When people leave their countries to come to live in [country], do you think it has a bad or good effect on those countries in the long run?” This question, however, dealt with the effects of migration on countries of origin and did not correlate with the other questions, and was therefore excluded from further analyses. 17 The factor analysis was executed using the main component method. The factor extraction criterion was a characteristic number greater than 1. As regards the analysis executed using the entire data file, the extracted factor accounted for 56 % of the total variance. 18 Index reliability was also verified for each of the countries separately. Reliability coefficients (Crombach’s alpha) range from 0.74 to 0.88. Attitudes towards immigrants and effect of Migration in european Countries 89 result, people in European countries can be said to believe that the arrival of migrants is more negative than positive for the host country or, at best, that advantages and disadvantages of migration balance each other out. People in Luxembourg and Sweden feel the least threatened by migrants as compared to the other countries under consideration. In contrast, respondents in Greece, the Czech Republic and Hungary view effects of migration most negatively (Table 3.03). A possible rise in crime is perceived as the biggest concern. People in European countries tend to believe that inflows of migrants could worsen problems related to crime. Another cause of considerable concerns, especially in Southern Europe and countries that underwent economic transformation in the 1990s, is the growth of unemployment for the domestic population. However, most people in most countries covered by our analysis view immigration as enriching, rather than damaging to the host country’s culture. Relation between Attitudes towards Migration, Migration Indicators and Economic Situation in the Host Country Is there a correlation between the share and structure of migrants in a country and the attitudes of the domestic population towards them? Are these attitudes related to the given country’s economic situation? In order to assess factors influencing attitudes towards migrants in the selected countries on the macro level, an aggregated data file was created where a single line represents each of the analyzed countries. The aggregated file contains two sum indexes construed using attitudinal items (the index of willingness to accept migrants and the index of perception of the effects of migration – see above), indicators of volume and structure of migration, as well as an indicator of economic situation in the analyzed countries (GDP per capita in terms of purchasing power parity in 2003).19 As it is expected that attitudes towards migration may be also influenced by the type of migrants entering the given country, the analysis included not only the total share of migrants in the host country’s population, but also the share of migrants coming from non-European and/ or undeveloped countries ( i.e. migrants not coming from 15 EU members, EFTA, the United States and Canada) as well as the number of asylum seekers per 10,000 inhabitants (see Table 3.01). Migrants from countries that have a similar education system, language or culture posses more easily transferable workplace skills and therefore probably also better fulfill the prerequisites for economic and cultural integration in the new country. 19 Data taken from OECD statistics. 90 Jana Chaloupková | Petra Šalamounová TABle 3.03 Perception of Social, Cultural and Economic Effects of Immigration by Country Sum Countries A B C D E F Index luxembourg 5.4 4.8 6.9 7.2 5.8 3.0 5.6 Sweden 6.1 4.7 5.5 7.1 6.2 3.4 5.5 Finland 5.1 4.1 5.3 7.3 5.3 3.4 5.1 Switzerland 4.9 4.2 5.9 6.2 5.2 2.9 4.9 Spain 4.8 4.8 5.4 5.9 4.8 3.3 4.9 Denmark 5.6 4.1 4.8 5.8 5.5 3.3 4.8 Norway 5.5 4.7 5.4 5.8 4.8 2.7 4.8 Austria 4.7 4.5 5.6 5.9 4.8 3.3 4.8 italy 4.6 5.0 5.3 5.3 4.5 3.6 4.7 France 5.0 4.4 5.1 5.3 4.6 3.5 4.6 ireland 4.2 3.7 5.0 5.6 5.3 4.2 4.6 Netherlands 5.1 4.3 4.8 6.1 4.7 2.9 4.6 Poland 3.9 4.1 4.5 6.2 5.2 3.3 4.6 germany 4.2 3.9 5.2 6.2 4.9 2.9 4.6 Portugal 3.8 5.5 4.8 5.2 3.9 3.1 4.4 Belgium 4.3 4.0 4.6 5.8 4.3 3.0 4.3 united Kingdom 4.4 3.9 4.4 5.2 4.6 3.6 4.3 Hungary 3.4 3.5 4.1 5.1 4.0 2.7 3.9 Czech republic 3.7 3.5 4.4 4.6 4.3 2.4 3.8 greece 2.8 3.2 3.7 3.6 3.4 1.9 3.1 Note: Average of answers on the scale from 0 to 10, where 0 means negative effects and 10 means positive effects. Sum index of immigration’s effects – average of answers on all items. A. People who come to live here generally take jobs away from workers in [country], or generally help to create new jobs? B. Most people who come to live here work and pay taxes. They also use health and welfare services. On balance, do you think people who come here take out more than they put in or put in more than they take out? C. Would you say it is generally bad or good for [country]’s economy that people come to live here from other countries? D. And, using this card, would you say that [country]’s cultural life is generally undermined or enriched by people coming to live here from other countries? e. is [country] made a worse or a better place to live by people coming to live here from other countries? F. Are [country]’s crime problems made worse or better by people coming to live here from other countries? Source: eSS 2002 Moreover, immigrants from more developed countries are also usually better qualified. In contrast, economic integration of immigrants accepted Attitudes towards immigrants and effect of Migration in european Countries 91 for humanitarian reasons may be more difficult and, from the point of view of the host country, also more expensive [Bauer, Lofstrom, Zimmermann, 2000]: it is for this reason that the public in countries with a high share of asylum seekers may be more concerned about the effects of migratory flows. It is likely, however, that attitudes towards immigration may be significantly influenced by a number of other factors, including the above-mentioned extent of illegal immigration or the educational and social structure of immigrants. Also important is the overall level of public awareness regarding the immigration process and the number of foreigners living in the given country. Immigration and problems related to it are often covered by the media, but it would be extremely difficult to establish the specific situation existing in the individual European countries at the time of data collection and the extent to which the situation influenced respondents. Table 3.04 shows the correlation coefficients for statistics concerning the size of migration and its structure, the economic situation of the given country and the public attitude towards migration in the analyzed countries. It is necessary to underscore in this respect that, as stated above, statistics concerning the numbers of immigrants do not fully reflect the actual number of foreigners living in any given country. Not only are data about migrants collected differently in different countries, but they also do not take into account illegal and unregistered immigrants. Furthermore, distortions arising in the comparison of the individual countries are also caused by the irregular distribution of this error, as the volume of illegal immigration is different in the various countries. It is the existence of a large community of illegal immigrants that can significantly contribute to a negative perception by the majority of the effects of immigration. Because a larger share of migrants and asylum seekers move to the more developed European countries, and because the structure of migrants in terms of their country of origin is different in Western, Southern and Central and Eastern Europe (for example, Southern European and post-communist countries have a greater number of migrants from the former Eastern bloc), we have controlled the individual countries’ gross domestic product in our correlation calculations. According to one hypothesis, the willingness of the inhabitants in the host countries to accept immigrants decreases as the presence and concentration of foreigners in the given country increases [Gang, Rivera-Batiz, Yun, 2002]. However, the correlation between the share of migrants and asylum seekers and the willingness to accept foreigners was not found to be significant. This may support the idea that countries which are more open to immigrants attract them in greater numbers. Theories explaining the causes of migration 92 Jana Chaloupková | Petra Šalamounová usually emphasize two groups of factors at work during the decision-making process: first, there are the factors driving people from their country of origin (or keeping them there), and second, there are the factors attracting people to a destination country (or factors dissuading them from a given country). Factors that make countries more attractive for migration include, above all, economic aspects (job opportunities, difference between salaries in the home country and the destination country), but also other aspects, such as the presence of social networks of fellow countrymen, liberal immigration policies, a generous welfare system and, last but not least, the overall “openness” of the destination country and the willingness of the public to accept newcomers. TABle 3.04 Correlation Analysis Share of Share of Number Sum index 1 Sum index 2 migrants migrants of asylum (willingness (perception not from seekers to accept of migration developed per 10,000 migrants) effects) countries inhabitants Share of migrants not from -0.63 developed countries Number of asylum seekers 0.38 -0.50 per 10,000 inhabitants Sum index (willingness to -0.08 0.38 -0.39 accept migrants) Sum index (perception of 0.41 -0.50 0.42 -0.67 migration effects) HDP 0.77 -0.75 0.57 -0.31 0.64 Partial correlation (controlled for gDP) Share of migrants not from -0.14 developed countries Number of asylum seekers -0.11 -0.14 per 10,000 inhabitants Sum index (willingness to 0.26 0.23 -0.27 accept migrants) Sum index (perception of -0.17 -0.04 0.09 -0.64 migration effects) Note: Pearson’s correlation coefficients, partial correlation coefficients (controlled for GDP). Sum index 1 (willingness to accept migrants): the higher is this index, the more negative attitude towards inflow of migrants. Sum index 2 (perception of migration effects): the higher is this index, the more positive is the perception of effects of migration. Source: Aggregated data file (ESS 2002, OeCD statistics). Attitudes towards immigrants and effect of Migration in european Countries 93 The openness (or lack thereof) of the public towards immigration is closely related to the question whether people in the destination country tend to view migration as beneficial for their country or rather as threatening. The stronger the conviction that immigration has a negative influence on the economy and other areas of life in the destination country, the greater the public support for setting limits on migratory flows. Figure 3.01 demonstrates this correlation. Figure 3.01 Willingness to Accept Immigrants and the Perception of the Effects of Migration Average Values of Sum Indexes Note: For description of indexes see Tables 2 and 3. A-Austria, B-Belgium, CH- Switzerland, CZ-Czech Republic, D-Germany, DK-Denmark, e-Spain, F-France, FIN-Finland, GB-Great Britain, GR-Greece, H-Hungary, I-Italy, IRL-Ireland, L-Luxemburg, Nl-Netherlands, N-Norway, PL-Poland, P-Portugal, S-Sweden. Source: eSS 2002 Our analysis has confirmed that the inflow of migrants raises more public concern about the negative effects of migration in poorer European countries than in the more developed countries. However, the correlation between the given country’s economic situation and the willingness to accept immigrants has been shown to be relatively weak. Findings also support the notion that effects of migration are perceived more negatively in countries in which the majority of immigrants do not come from the most developed countries, both in Europe or overseas. However, if we control for GDP, this correlation disappears, as a higher share of immigrants from non-developed countries is found in the economically less developed 94 Jana Chaloupková | Petra Šalamounová European countries. Results do not suggest that the public in countries with a high share of asylum seekers are any more sensitive to possible negative economic, social and cultural effects of immigration. On the contrary, it seems that unlike the public in other European countries, the public in countries that attract asylum seekers does not perceive migration as a threat. Attitudes towards the Integration of Immigrants The openness of a society is not only determined by its willingness to allow immigrants to reside in the country, but also by the rights the immigrants are granted and the extent to which the society is willing to allow foreigners (as well as minorities) to preserve their cultures. There are three models of approaches that nation-states take towards the integration of immigrants [Drbohlav, 2001a; Barša, 2003]: 1) assimilation model, 2) discrimination model (differential incorporation), and 3) multicultural model (or cultural pluralism). Allowing for a certain degree of simplification, these models may be derived from the approaches of France, Germany and the United Kingdom [Barša, 2003]. Today, however, these approaches are undergoing certain adjustments, and it is, therefore, necessary to consider them as idealized types, whose traits overlap to some extent in reality. The first approach, i.e. the assimilation model, is characteristic for France. It assumes a complete dissolving of specific group characteristics into the homogenous body of free and equal citizens. Immigrants are allowed to obtain citizenship and well as the rights and responsibilities of a citizen of the host country at the price of having to accept the cultural standards of that host country. The multicultural model, in contrast, is based on tolerating group differences, and it grants immigrants full rights equal to those of the established majority, while allowing immigrants to preserve their specific cultural identity. European countries applying policies of cultural pluralism include the United Kingdom, the Netherlands and Sweden. The discrimination model (model of differential incorporation) is typical for Germany and Austria. It is based on the assumption that the residence of immigrants in the host country is only temporary. This model distinguishes the economic involvement of immigrants in the host society from their civic and political involvement and cultural integration. While immigrants enjoy fundamental human and social rights, it is very difficult for them to obtain citizenship [Barša, 2003].20 In the 20 As stated before, the above specified principles represent ideal types. In the past decade in particular, the principles have been subjected to an extensive debate and have been, to a large extent, reevaluated. It is worth noting that the Netherlands and the United Kingdom Attitudes towards immigrants and effect of Migration in european Countries 95 ESS survey, respondents used a scale of 1 to 5, where 1 denotes complete agreement and 5 complete disagreement, to indicate the extent of their agreement with the statement that it is better for a country if all people share the same customs and traditions. This enables a comparison of European countries based on their preference of cultural homogeneity or cultural pluralism. Using the same scale, the survey extracted answers to the question whether people who have moved to the host country should enjoy the same rights as others. Table 3.05 shows the share of respondents who agreed with this statement. The strongest agreement with the statement that it is better for a country if all people share the same customs and traditions was recorded in Southern Europe and the post-communist countries included in the survey. At the same time, however, respondents in these countries were willing to grant immigrants the same rights as those enjoyed by others. Hungary was an exception, as respondents there expressed the lowest rate of agreement with granting the same rights to newcomers. In contrast, respondents in Germany, Austria, Belgium and Switzerland voiced less support for the statement that it is better for a country if all people share the same customs and traditions. Respondents from these countries were, at the same time, less willing to grant foreigners the same rights. Respondents from Scandinavian countries were the most likely to grant immigrants the same rights and to accept different customs and traditions. in particular are moving away from multiculturalism and towards individual civil integration of immigrants [Baršová, 2005]. Also, Germany made it easier for children of foreigners born on German soil to obtain German citizenship. 96 Jana Chaloupková | Petra Šalamounová TABle 3.05 Attitudes towards Integration of Immigrants; Cultural and Legal Integration Better for a country if almost People who have come to live here everyone shares the same should be given the same rights as customs and traditions everyone else Switzerland 35.4 46.8 Sweden 36.5 86.1 germany 39.2 58.7 Austria 39.9 56.6 luxembourg 41.2 76.2 Norway 42.6 81.4 Netherlands 43.6 65.2 ireland 43.6 75.0 united Kingdom 44.7 67.3 Denmark 46.3 79.4 italy 50.7 70.2 France 52.1 62.0 Belgium 54.1 56.9 Finland 54.3 71.8 Hungary 57.3 40.4 Spain 58.1 73.9 Czech republic 69.9 72.4 Portugal 71.3 79.7 Poland 71.5 70.5 greece 83.4 65.9 Note: Answers agree strongly and agree (in %) Source: eSS 2002 3.4 Conclusion Almost all European countries face inflows of immigrants, though with varying intensity. Changes related to immigration bring to the destination countries the risk of growing economic and social insecurity and the threat of potential conflicts, but also potential benefits. Public opinion differs in the individual European countries regarding the extent of incoming migration. People in Greece and Hungary are the least Attitudes towards immigrants and effect of Migration in european Countries 97 open to potential immigration. Despite viewing effects of migration with similar concern as Hungarians, Czechs expressed a greater willingness to accept immigrants and came close to the attitudes held by people in Western Europe. Inhabitants of Sweden and Switzerland voiced the most favorable attitudes towards immigration. The ESS survey examined the willingness to accept immigrants based on whether they are of the same or different race or ethnicity as the majority and whether they originate from wealthier or poorer European or non-European countries. Although the attitudes towards immigrants from different regions are closely related, and those who do not welcome newcomers from one region usually do not welcome any other immigrants, it is generally true that people in European countries are more willing to accept immigrants of the same race and from other European countries than those of different races or coming from non-European countries. The analyses did not confirm that people from countries with a higher share of immigrants are more likely to support setting limits on migratory flows. A more important factor influencing declared support for setting limits on immigration is the assessment of effects of migration on the host country. Our findings confirm that effects of immigration are viewed more negatively by people from less affluent countries in Southern Europe and also respondents from Hungary and the Czech Republic, where the majority of immigrants originated from less developed countries. Data from the European Social Survey also confirm that there are different attitudes towards the integration of immigrants in European countries. While respondents from Scandinavian countries are willing to grant immigrants the same rights as those enjoyed by the majority, and do not seem to privilege a society in which everybody shares the same customs and traditions, people from most Western European countries are open to societies composed of more than one culture but are less likely to grant immigrants the same rights. On the other hand, people from Southern Europe, Poland and the Czech Republic are more likely to agree that it is better for a country if all people share the same customs and traditions, but at the same time believe that immigrants should enjoy the same rights as the majority. Cure for Loneliness? Social Activity and Its Role in Changing European Societies 4 Anna Domaradzka 4.1 Introduction The starting point of this paper is Ulrich Beck’s thesis which posits that nowadays close family relations are disappearing, leaving individuals alone with their problems and emotional needs [Beck, 1992]. It can be stated that, as a result of this, a new pattern of social networks is emerging; these networks are less connected with family and more with various self-development groups, voluntary associations and other forms of social activity. This pattern especially concerns older adults, who are less active in the labour market but still able to lead full lives, and are often affected by the “civilization disease” of loneliness. This problem may become visible after their children leave home, which is often referred to as “empty nest syndrome“. The question is: have close family ties diminished, and if so, have they been replaced by new forms of social activity? We may also ask, do people tend to close themselves off at home and withdraw from social life, or do they try to establish new forms of activity and substitutes for the family network? We can ask whether a scarcity of family contacts is indeed visible among surveyed countries, and if there are any differences between models of family life in European regions. The second issue is the substitute for close family relations. If the family ties are weakened, do new forms of social relations emerge to replace the traditional family networks? It can be anticipated that in the case of diminishing personal relations with relatives, people will try to build new forms of social networks, centered on participation in different social activities. The present analysis is based on data from the module “Social Networks” conducted in 2001 as a part of the International Social Survey Program. Among the researched countries are those with high as well as those with low levels of social participation (participation in voluntary organizations), which should allow for a richer international comparison. While the samples from each country are too small for this kind of analysis, the comparison between two European regions (Western Europe and Central-Eastern Europe) is presented. The research question is investigated using bivariate analysis. Most of the presented results come from mean differences (deviation from the mean) analysis. T-tests were used for significance checks. The goal of the analysis is to test the hypothesis that the weakening of relations between family members results in the engagement in different forms of social networks. It seems probable that a high level of social activity correlates with the disintegration of family bonds, but it may also be that the inclination towards social activity is a form of extended family activity. 4.2 Definitions While analyzing the data, various hypotheses were taken into account. Most of them are based on widely known definitions of social processes or phenomena, which nevertheless should be defined prior to presentation of the hypotheses. Conjugal family 1 is a nuclear family of adult partners and their children in which the family relationship is principally focused inwards. Ties to extended kin are voluntary and based on emotional bonds rather than strict duties or obligations. Most important is the link between spouses and their children, while connections to other relatives are comparatively unimportant. Extended family,2 in societies dominated by the conjugal family, refers to kindred (network of relatives that extends beyond the domestic group) who do not belong to the conjugal family. Extended family often means that many generations live under the same roof. Such families can include, aside from parents and their children, cousins, aunts, uncles, grandparents, foster children etc. Empty nest syndrome 3 refers to the grief that many parents feel when their children move out of home. This condition is usual y more common among women, who are more likely to have had the role of primary caregiver. In many cases, empty nest syndrome is compounded by other difficult life events or significant changes that happen around the same time, such as retirement or menopause. By empty nest syndrome, we understand a general feeling of depression and loneliness that parents feel when one or more of their children leave home. A strong maternal or paternal bond between the parent and child can make the condition worse. Empty nest syndrome has become more prevalent in modern times, as the extended family has become less common than in previous generations, and the elderly are left to live by themselves. However, some research has shown that empty nest syndrome is not necessarily a negative experience. For instance, it is argued that children leaving home may improve the relationship between partners [Kalmijn, Poortman, 2006]. It can also be argued that the empty nest experience may have positive or negative consequences for a woman’s well-being, depending on her historical cohort membership and employment status. In one of the studies [Adelman, Antonucci, 1989], it was posited that the empty nest was likely to be a negative experience among the particular set of women who reached adulthood during the period of strong societal emphasis on women’s maternal role (known as the feminine mystique), while it would be experienced 1 See www.wikipedia.org 2 See www.wikipedia.org 3 See www.wikipedia.org Cure for loneliness? 103 positively among the earlier group who, as young adults, were encouraged to enter the labour force (during World War II). The results of this study4 show that both employment and cohort have important independent associations with a woman’s well-being at midlife, but that the experience of the empty nest is especially influenced by generational experiences. Loss of motherhood 5 Empty nest syndrome can afflict both parents, but mothers seem to be most susceptible. Many mothers may have dedicated 20 years or more of their lives to bringing up their children and see motherhood as their primary role, even if they were also working during this time. Once the last child has moved out, the mother may feel that her most important job is finished, and she may feel worthless, disoriented or unsure about the meaning of her future. However, most mothers adapt over time. Solitude is the state of being alone and secluded from other people, and it often implies having made a conscious choice to be alone. Loneliness 6 is an emotional state in which a person experiences a powerful feeling of emptiness and isolation. Loneliness is a feeling of being disconnected and alienated from other people so that it feels difficult or even impossible to have any form of meaningful human contact. Loneliness is, therefore, unwilling solitude. Personal contacts are direct, face-to-face interactions (in this case, with other members of the family). Remote contacts are contacts through phone, Internet or mail (abstracted type of communication). It is said that people often mitigate loneliness by interacting with others via the phone or the Internet. However, it is widely believed that purely remote relationships are no substitute for face-to-face relationships, and also that such relationships are less stable. 4.3 Questions and Hypotheses Several hypotheses have been formulated concerning the issue of the replacement of family ties with other forms of social networks. Hypothesis 1: Level of social activity is positively related to the strength of family ties as measured by the frequency of contact with other members of the family. 4 Analyses of covariance tested the relationships among empty-nest status, cohort membership, and employment status, and three measures of psychological well-being, adjusted for age, education, and marital status. 5 See www.betterhealth.vic.gov.au 6 See www.wikipedia.org 104 Anna Domaradzka Hypothesis 2: Level of social activity decreases if the type of contact with relatives is remote. Hypothesis 3: Parents of adult children, especially those who no longer live together, are more socially active than others (empty nest syndrome hypothesis). Hypothesis 3a: Mothers of only sons are most affected by empty nest syndrome (Scheler’s hypothesis7). Additional questions arose in response to anticipated differences based on region and gender. 1. Why are people socially active? 2. What are the characteristics of the activists? 3. Is there a difference between Western Europe and Central and Eastern Europe? 4. Is there a difference between men and women? 5. Is there a difference that depends on the child’s gender? 6. Are parents of only children different in terms of social activity? 4.4 Used and Created Measures In studying these hypotheses, based on Social Network ISSP data, several important restrictions must be recognized. One is that the quality of family life was not measured – no question concerning satisfaction from family/marital life was included. There is also no information about household composition, so it is not possible to determine the inhabitants of a respondent’s household. However, information about the household can be partly derived from other variables. Since the author is restricted by the size of this essay, instead of presenting a wide range of issues, the differences between various groups of people will be described according to selected questions and hypotheses. Thus, it is not possible to include an analysis of social activity which takes into account such variables as age, level of education, place of living, or employment, as these factors have been deemed unimportant for social engagement. 7 According to Max Scheler’s [1994] thesis, mothers of only sons are most affected by resentment (which, in this case, can also be called an empty nest syndrome); when their sons leave their households, these mothers often suffer significant emotional pressure. It can be argued that there is a difference between sons and daughters because the latter tend to stay more connected with the family and also bring in new members (husband, children). In case of a son, Scheler argues, the mother views his departure as a lost investment – after years of caring and giving, she is forced to let him go, often to another woman. This is why mothers of only sons are of special interest in this paper. Cure for loneliness? 105 Another limitation is that social activity is seldom in many of the countries under consideration,8 resulting in sub-samples that are too small to allow any quantitative analysis. Therefore, the analysis is focused on three aggregated groups of countries. The regional groups are: 1) Western European countries (Great Britain, Northern Ireland, Austria, Norway, Spain, France, Switzerland, Finland); 2) Central and Eastern European countries (Czech Republic, Hungary, Latvia, Slovenia, Poland, Russia); 3) both Western and Central and Eastern European countries (1+2).9 The sample size for all fourteen European countries is N=17963, for WE N=9942 and for CEE N=8022. Also, not every question in the database was used in each country, which made some calculations impossible. Nevertheless, to make the most of the Social Networks database, conducted analyses are focused on the data that allowed a comparison between social activity and the family situation and other variables. In some cases, the scales used in the questions selected for analysis were recoded to be more useful measures of intensity or frequency. Additionally, some new measures were created. In the Social Network 2001 database there is one question concerning social activity: “People sometimes belong to different kinds of groups or associations. The list below contains different types of groups. For each type of group, please tick a box to say whether you have participated in the activities of this group in the past 12 months.” The list of groups and associations is: A political party, club or association A trade union or professional association A church or other religious organization A sports group, hobby or leisure club A charitable organization or group A neighborhood association or group Other associations or groups For each item there are four possible answers: I have participated more than twice I have participated once or twice I belong to such a group but never participate I do not belong to such a group. To establish the measure of activism intensity, a specific number of points was attributed to each answer. “I have participated more than twice” received 8 European countries included in 2001 Social Networks database were Great Britain, Northern Ireland, Austria, Hungary, Norway, Czech Republic, Slovenia, Poland, Russia, Spain, Latvia, France, Switzerland and Finland. 9 From this point forward, “Western Europe/Western European” will be abbreviated as “WE” and “Central and Eastern Europe/European” as “CEE.” 106 Anna Domaradzka 6 points; “I have participated once or twice” was awarded 3 points; “I belong to such a group but never participate” received 1 point and “I do not belong to such a group” was recoded as 0. As a result, the original scale has been transformed into intervals that can better describe the level of engagement. Of course, we have to remember the difficulty in creating this kind of interval scale. While there is a qualitative difference between active and non-active, the difference between active and very active is rather quantitative. Thus, in some of the analysis, non-active cases are excluded. When analyzing the levels of activity in different types of organizations, it should be underlined that personal preferences may not be the only reason for higher levels of some type activity and that the structure of the voluntary sector in certain countries or regions might also be important. To determine the inhabitants of the respondent’s household, questions Q3, Q7, Q9 and Q11 were used (How often do you see or visit your brother or sister/ daughter or son/father/mother?). One of the possible answers was “He/She lives in the same household as I do.” By using al four questions, it was possible to determine that respondent is living with at least four of his/her close relatives. The variable was then coded to show with whom the respondent is living. To measure the frequency of contacts, a new scale was applied. Again, a specific number of points was attributed to each answer. It was derived from simple calculations based on one week and than multiplied to make calculations easier. In case of daily contacts, 7 points were awarded, several times a week was translated into 3.5 points, once a week into 1, once a month into 1/4, several times a year into 1/16 and less often into 1/180. The number of points was then multiplied by 180. The final point assignments are as fol ows: “Daily” = 1260; “At least several times a week” = 630; “At least once a week” = 180; “At least once a month” = 45; “Several times a year” = 11; “Less often” = 1. While not an ideal method of translating verbal data into numbers, it is very practical. In the case of frequency and type of contact, it should be remembered that the type of contact may be the only possible type in cases where the family is dispersed. Thus, both types of contact are taken into account in some of the calculations. During the analysis, various correlations were calculated. Most of the correlations are significant but very low, which is why the presented results were ultimately based on an analysis of deviations from the means. To answer the question concerning characteristics of people who are active, two-variable correlations were calculated and several regression models were applied. However, the percentage of explained variance was very low. The main reason for this was the small size of sub-sample of activists. 4.5 Results Presented below are the activity and frequency of contact indicators and also an analysis of deviations from the means for several issues: – Level of activism by type of organization and region – Type of contact by region – Social activity of parents of only children (by child’s gender) – Social activity of parents of only children no longer living with their parents (by child’s gender) – Social activity of parents of one or more adult children – Social activity vs. frequency of contact with relatives. Figure 4.01 shows the difference in levels of activism in both the WE and CEE regions. The results are shown on a scale of 1 (I belong to such a group but never participate) to 6 (I have participated more than twice in a last 12 moths). Figure 4.01 Type of Social Activity (mean) by Region   The significance of the results was calculated using a t-test for equality of means for independent samples. For each type of social activity, the mean is higher for WE countries than for CEE countries. Those differences in politi- 108 Anna Domaradzka cal and social participation are widely discussed and usually connected to a post-communist heritage and lower economic stability. In countries with developed democracies, people tend to be more engaged socially and politically than in countries where everyday economic problems remain the main focus for the majority of citizens. It is also a matter of political tradition; in most WE countries, the tradition of social dialogue and cooperation between the voluntary sector and the government is well established, while that practice is still not very popular among governments in CEE countries. It is interesting to compare the results concerning social activity with frequency of contact, as depicted in Figure 4.02. Figure 4.02 Type of Contact (mean) by Region There is a visible and significant difference between frequency of personal contact with family in WE compared to that in CEE, with respondents from CEE meeting their relatives more frequently. This may be a result of weakening family ties in the more developed Western countries while the more traditional family forms have remained stronger in CEE. The difference between CEE and WE is significant in the case of personal contact as wel as that of cumulative personal and remote contact. In the case of remote contact, differences in the means exist, but are smal er and insignificant. Cure for loneliness? 109 Social Activity – Europe Figure 4.03 shows the overall differences between the types of activism practiced by men and women (average for European countries included in 2001 Social Networks database). The results confirm the differences shown by other research concerning social participation ( e.g. World Values Survey). In short, it can be stated that men tend to participate in political or professional groups, while women are more active in religious and charitable organizations (although in cases of charitable organizations, the difference is not significant). The results are shown on the scale Figure 4.03 Type of Social Activity by Gender in Europe   of 1 (I belong to such a group but never participate) to 6 (I have participated more than twice in a last 12 moths). The significance of the results was calculated using a t-test for equality of means for independent samples. Tables 4.01 and 4.02 show deviations from the mean social activity levels for women and men in all fourteen of the European countries under consideration. The yellow cells represent a deviation greater than 0.2 and negative (<-0.2) and the green cells show a deviation greater than 0.2 and positive (>0.2). The mean was calculated respectively for men and for women. 110 Anna Domaradzka Table 4.01 shows that in cases of adult only children, there is a noticeable difference between mothers of sons and mothers of daughters. Mothers of only sons tend to be more active in political parties and charitable organizations than the average woman, but less active in trade unions or professional organizations. In contrast, mothers of only daughters are less active in political parties and charitable organizations, but more active in associations connected with the labour market. In other words, two opposite trends can be observed when comparing mothers of only sons to mothers of only daughters. If we interpret the participation in charitable organizations as a measure of frustration that is released or vented by helping others, then this result is an interesting depiction of Scheler’s thesis. The results of the empty nest syndrome can be more clearly observed when the data concerning mothers of adult sons and daughters not living together is examined. Also in this case, there is an observable difference that depends on the only child’s gender. Mothers of only sons tend to be more active in church and charitable organizations, while mothers of only daughters are more active in professional associations and less interested in charitable organizations. In the case of religious activity, which is more specific to women than men, there is a difference between women with no adult children (including child-TABle 4.01 Social Activity of Women in Europe - rade union or Political party T professional association Church or re ligious organi- zation Sports group Charitable organization Neighbourhood group Mothers of adult only child 0.07 -0.05 0.05 0.01 0.11 -0.08 Mothers of adult only sons 0.41 -0.26 0.12 -0.10 0.43 -0.10 Mothers of adult only daughters -0.40 0.22 -0.14 0.00 -0.18 -0.09 Mothers of adult only sons / not living together 0.04 -0.05 0.24 -0.25 0.32 -0.74 Mothers of adult only daughters / not living together -0.49 0.28 -0.05 -0.26 -0.44 -0.04 Women without adult children -0.12 -0.05 -0.18 -0.01 -0.15 0.02 Women with one or more adult children 0.13 0.14 0.21 0.05 0.18 0.01 Cure for loneliness? 111 less women) and mothers with at least one adult child. The latter seem more inclined to contribute to church or religious activities, which can be also connected to a religiosity that increases with age. Table 4.02 shows that when adult only children are considered, the difference between fathers of sons and fathers of daughters is not as marked as it is with the mothers. Usually fathers of only children are less active than average in religious organizations. This changes when it comes to children who do not live with their parents. It seems that the absence of adult children may increase religious engagement in the case of sons but decrease it in the case of daughters. This may be the effect of a smaller average engagement of men in religious organizations as compared to women. Compared to mothers, fathers are more active in neighborhood organizations. The result of male empty nest syndrome is also visible in cases of charitable organizations. Compared to women, however, there is no difference that depends on the only child’s gender. Fathers of only sons tend to be less active than average in political parties, but more active in trade unions or professional organizations. Conversely, fathers of only daughters are more active in political parties, but less active in religious associations. Political and professional activity seem to be more TABle 4.02 Social Activity of Men in Europe rade union or Political party T professional association Church or religious organi- zation Sports group Charitable organization Neighbourhood group Fathers of adult only child 0.09 -0.05 -0.33 -0.19 0.12 -0.02 Fathers of adult only sons 0.16 0.01 -0.29 -0.35 0.15 -0.35 Fathers of adult only daughters 0.08 -0.03 -0.40 0.02 0.11 0.35 Fathers of adult only sons / not living together -0.20 0.21 0.17 -0.14 0.37 -0.12 Fathers of adult only daughters / not living together 0.26 0.04 -0.34 0.13 0.48 0.25 Men without adult children -0.15 -0.16 -0.13 0.11 -0.15 -0.16 Men with one or more adult children 0.15 0.30 0.30 -0.17 0.19 0.22 112 Anna Domaradzka specific to men than women; the differences between men with no adult children (including childless men) and fathers of at least one adult child are also visible, but not as much as with women. The latter seem more inclined to participate in professional and religious activities. Social Activity – Western Europe In the case of WE countries, the main differences concerning social activity by gender are visible in trade unions/professional organizations, which are more frequently male spheres of activity, compared to church and religious organizations, in which more women participate. In other cases, the differences by gender are not significant. Charitable organizations are a favorite form of activism for mothers of only children, especial y of male children. In most other cases, mothers of only sons are less inclined to engage in social activity. In regards to political parties, however, there is a visible difference between mothers of only sons and mothers of only daughters. The latter are much less active in politics than is average for women in the region, while mothers of only sons are more engaged. Figure 4.04 Type of Social Activity by Gender in Western Europe   Cure for loneliness? 113 Empty nest syndrome in the case of mothers of only sons results in greater activity in charitable organizations, while activity in most of the other kinds of organizations is lower than average. Also in regards to mothers whose daughters do not live with them, the tendency to become involved in political parties is much lower than average. TABle 4.03 Social Activity of Women in Western Europe - rade union or Political party T professional association Church or re ligious organi- zation Sports group Charitable organization Neighbourhood group Mothers of adult only child -0.09 -0.24 -0.10 -0.03 0.25 -0.03 Mothers of adult only sons 0.36 -0.52 -0.21 -0.22 0.48 -0.14 Mothers of adult only daughters -0.81 0.13 -0.07 0.05 0.12 0.09 Mothers of adult only sons / not living together 0.17 -0.37 -0.49 -0.31 0.51 -0.82 Mothers of adult only daughters / not living together -0.86 0.05 -0.10 -0.18 -0.24 -0.06 Women without adult children 0.00 -0.01 -0.21 0.00 -0.17 0.07 Women with one or more adult children 0.03 0.14 0.31 0.00 0.15 -0.07 TABle 4.04 Social Activity of Men in Western Europe - rade union or Political party T professional association Church or re ligious organi- zation Sports group Charitable organization Neighbourhood group Fathers of adult only child -0.12 -0.18 -0.42 0.09 0.50 -0.07 Fathers of adult only sons -0.16 -0.11 -0.23 -0.01 0.52 -0.40 Fathers of adult only daughters -0.20 -0.17 -0.57 0.27 0.52 0.27 Fathers of adult only sons / not living together -0.16 -0.07 0.43 0.01 0.60 -0.18 Fathers of adult only daughters / not living together -0.13 0.06 -0.54 0.34 0.60 0.03 Men without adult children -0.04 -0.17 -0.14 0.11 -0.18 -0.05 Men with one or more adult children 0.09 0.35 0.31 -0.23 0.13 0.09 114 Anna Domaradzka In the case of religious associations, women without adult children tend to be less active, while women who have at least one adult child are more inclined to activity connected with religion. This may be the effect of religiosity increasing with age. In WE, fathers of only children are more inclined to be active in charitable organizations than is average for men in the region. They are, however, less inclined to be engaged in religious organizations. Only in cases of neighborhood groups is there a visible effect of the child’s gender. While fathers of sons are less active, those with daughters tend to be more active than average. Empty nest syndrome is apparent when looking at charitable and religious organizations, and in the latter group, the child’s gender is significant. Only fathers of sons are more active in the religious organizations, while the opposite is true with fathers of daughters. Also, with most of the activities, fathers of at least one adult child are more active than fathers without adult children, especially in professional and religious organizations. This may be, as it is with women, connected to age (experience and conservatism). Social Activity – Central and Eastern Europe In CEE countries, the main differences of social activity by gender are visible in the case of political parties, which is more frequently a male sphere of activity, as compared to church and religious organizations, in which more women participate. It is worth mentioning that the differences are more visible in CEE than in WE. It is also interesting to note that the level of trade union participation is, on average, the same for men and women in CEE, while there is a significant difference in WE. In both groups of countries, the differences by gender regarding sports groups, charitable organizations or neighborhood groups are not significant. In WE, charitable organizations are favored by mothers of only children, while in CEE these women are much more active in political parties. There is a visible difference between women without adult children and those with at least one adult child. Mothers of adult children are much more active in political parties than other women in the region, which further supports the empty nest hypothesis. In most of the cases, mothers of only sons are more likely to engage in social activity (unlike in the western countries). Concerning political parties, there is a difference between mothers of only sons and mothers of only daughters, but much less of one than in WE. Mothers of only sons tend to be more engaged in political parties than the average for both regions. Empty nest syndrome in the case of mothers of only sons results in much higher activity in religious organizations, especially when the son no longer Cure for loneliness? 115 lives with his mother. Compared to the average, church organizations are much more popular among mothers of only sons in CE than in WE. In the case of daughters, there is no visible difference from the average level of religious activity. This trend may also be connected with growing secularization in the western countries. Figure 4.05 Type of Social Activity by Gender in Central and Eastern Europe   When comparing average activity in charitable organizations, we can see a difference based on the only child’s gender. While mothers of sons are more active, mothers of only daughters tend to be much less active in this type of organization. This is even more the case when the daughter no longer lives with her mother. On the other hand, these mothers tend to be more active in trade unions and professional organizations than other women in the region. In cases of mothers and daughters not living together, the tendency for the mother to be involved in political parties is higher than average, as opposed to those in western countries. Though they are very involved in church organizations, mothers whose only sons do not live with them are less active in neighborhood organizations. This 116 Anna Domaradzka is similar in the western countries, but instead of church, mothers of only sons are active in charitable organizations. In CEE countries, political activity is high not only among the mothers of only children but also among the fathers. In CEE, fathers of only children are more inclined to be active in political parties and trade unions or professional associations than is average for men in the region, while less inclined to engage in charitable organizations and sports groups. This is exactly opposite in the WE countries. Only in the case of neighborhood groups is there a visible effect of the child’s gender. While fathers of sons are less active, those with daughters tend to be more active than average (same in WE). TABle 4.05 Social Activity of Women in Central and Eastern Europe rade union or Political party T professional association Church or religious organization Sports group Charitable organization Neighbourhood group Mothers of adult only child 0.68 0.16 0.40 0.19 -0.13 -0.10 Mothers of adult only sons 0.72 0.01 0.87 0.30 0.40 0.04 Mothers of adult only 0.66 0.30 -0.14 -0.02 -0.60 -0.26 daughters Mothers of adult only sons / 0.13 0.47 1.42 -0.07 0.13 -0.54 not living together Mothers of adult only 0.52 0.60 0.11 -0.37 -0.72 0.09 daughters / not living together Women without adult children -0.72 -0.14 -0.11 -0.05 -0.09 -0.09 Women with one or more 0.43 0.14 -0.05 0.03 0.22 0.20 adult children Empty nest syndrome is visible in cases of political parties and trade unions/ professional associations, and in both cases, the child’s gender makes a big difference. A father of an only daughter is much more likely to be active in political parties (comparable to the activity in church organizations of a mother of an only son), while in the case of a son, he is more likely to be active in professional organizations and trade unions. Similar to WE, in most of these cases, fathers of at least one adult child are more active than fathers without adult children. In CEE, they are more active in political parties, church and charitable organizations as well as neighborhood organizations than men without adult children. As with the women, this may be connected to age. Cure for loneliness? 117 TABle 4.06 Social Activity of Men in Central and Eastern Europe rade union or Political party T professional association Church or religious organization Sports group Charitable organization Neighbourhood group Fathers of adult only child 0.66 0.22 -0.06 -0.72 -0.57 0.13 Fathers of adult only sons 0.97 0.27 -0.32 -0.84 -0.17 -0.19 Fathers of adult only 0.76 0.28 0.15 -0.60 -1.25 0.53 daughters Fathers of adult only sons / -0.17 0.71 -0.21 -0.51 -0.02 0.05 not living together Fathers of adult only 1.25 0.03 0.35 -0.44 -0.83 0.76 daughters / not living together Men without adult children -0.72 -0.15 -0.12 0.15 -0.03 -0.46 Men with one or more 0.36 0.13 0.20 -0.07 0.34 0.57 adult children Social Activity – Summary The analyses presented above enable us to compare how parents of only children are different from the average women or men in each of the three regions. The differences between WE and CEE can be summarized as follows: 1. In WE, charitable organizations are the favored form of activism for mothers of an only child, while in CEE these mothers are much more active in political parties. 2. In CEE, fathers of only children are more inclined to be active in political parties and trade unions/professional associations than is average for men in the region, while they are less inclined to engage in charitable organizations and sports groups. This is exactly the opposite of the trend found in the WE countries. 3. Compared to the average, church organizations are much more popular among mothers of only sons in CEE than in WE. 4. In CEE, mothers of only sons are more eager to engage in social activity, while in WE the opposite is true. 5. Mothers of only sons tend to be more engaged in political parties than is average for both regions. 6. In both regions, fathers of sons are less active, whereas those with daughters tend to be more active than average. 118 Anna Domaradzka Social Activity by Type of Contact In CEE, there is a visible difference between those who have no contact with their relatives and those who have very frequent contact with their families, be it through phone/mail or personal visits. Both scale extremes (lack of contact and high frequency of contact) seem to have the same effect of lowering the tendency to participate in political parties. This is true for each type of contact, whether remote or personal. Those who contact family frequently tend to be more active in professional organizations, as well as charitable and neighborhood groups. Those who do not contact their close relatives are less active than average (for CEE countries) in cases of political parties, religious and charitable organizations as well as neighborhood groups. TABle 4.07 Social Activity by Type of Contact in Central and Eastern Europe Activity Personal contact remote contact mean lack low medium high lack low Medium High Political party 3.06 -0.42 0.65 0.09 -0.40 -0.49 0.21 0.20 -0.56 Trade union or professional association 2.43 -0.07 -0.24 -0.01 0.34 -0.08 -0.11 -0.02 0.25 Church or religious organization 3.35 -0.23 0.08 0.05 -0.05 -0.28 0.34 0.05 -0.18 Sports group 4.62 0.13 0.17 -0.05 -0.08 0.19 0.49 -0.08 -0.17 Charitable organization 3.42 -0.30 -0.03 0.00 0.32 -0.33 -0.78 0.12 0.25 Neighbourhood group 3.23 -0.56 -0.56 0.14 0.22 -0.65 -0.06 0.05 0.32 There is also a difference in activity depending on the kind of contact with family. Respondents from CEE who have little remote contact are more active in church organizations than those who have little personal contact. In the case of neighborhood groups, a rising trend can be observed. In other words, the higher contact index, the more frequent the activity. A similar trend can be observed in the case of charitable organizations and personal contact with family: with the growing frequency of contacts, activity also rises, as compared to average. The means of activity in WE are higher than in CEE for each type of organization mentioned in the questionnaire. More contact is visibly related Cure for loneliness? 119 to higher activity in the case of trade unions. Intensity of remote contact seems to be somehow connected with higher than average church/religious activity. Lack of contact with family, both personal and remote, visibly lowers the engagement in charitable organizations, as compared to the average for WE countries. This could indicate a lack of time, or perhaps a withdrawal from social life. Results from all fourteen European countries are less diverse, probably because the differences in the averages for both CEE and WE more or less balance out. Most of the results are close to average. In this case, the connection between lack of contact with family and lower activity in charitable organizations (also visible in western countries) becomes more interesting. TABle 4.08 Social Activity by Type of Contact in Western Europe Activity Personal contact remote contact mean lack low medium high lack low Medium High Political party 3.91 -0.01 -0.24 0.05 0.06 -0.24 0.18 0.02 0.01 Trade union or professional association 2.63 -0.13 -0.10 0.00 0.40 -0.13 0.05 -0.03 0.32 Church or religious organization 3.66 0.15 0.05 -0.07 0.37 0.22 0.23 -0.12 0.48 Sports group 5.05 0.03 0.04 0.03 -0.37 0.00 0.03 0.04 -0.33 Charitable organization 3.90 -0.29 0.12 0.02 -0.08 -0.26 0.01 0.01 0.18 Neighbourhood group 3.54 0.03 0.00 0.00 -0.04 -0.05 -0.04 0.02 -0.06 Social Activity by Type of Contact – Summary The analyses presented above enable us to see the differences in social activity by type and by frequency of contact with family in each of the three regions. The main results can be summarized as follows: 1. The mean of activity in every type of organization mentioned in the questionnaire is higher in WE than in CEE. 2. In WE countries, lack of contact with family, either personal or remote, visibly lowers the engagement in charitable organizations as compared to the average. 3. In CEE, those with frequent family contact tend to be more active in professional organizations, as well as in charitable and neighborhood groups. 120 Anna Domaradzka 4. In CEE, those who do not contact their close relatives are less active than average in political parties, religious and charitable organizations, as well as in neighborhood groups. 5. Respondents from CEE who have little remote contact are more active in church organizations than those who have little personal contact. 6. In WE, a higher intensity of remote contact seems related to a higher than average church or religious participation. Social Activity by Frequency of Contacts Figure 4.06 shows the mean frequency of contact by different types and by level of engagement in CEE. For CEE countries, the results are significant in two cases. In trade unions and professional associations, the mean frequency of contact with family is significantly higher for those who are more actively engaged. In the case of participation in church and religious organizations, the opposite trend can be observed. TABle 4.09 Social Activity by Type of Contact in Europe Activity Personal contact remote contact mean lack low medium high lack low Medium high Political party 3.70 -0.18 -0.03 0.06 -0.17 -0.36 0.10 0.08 -0.18 Trade union or professional association 2.56 -0.13 -0.12 0.01 0.32 -0.14 -0.03 -0.01 0.26 Church or religious organization 3.59 0.01 0.07 -0.05 0.16 0.02 0.22 -0.08 0.27 Sports group 4.95 0.01 0.09 0.02 -0.32 0.01 0.14 0.03 -0.33 Charitable organization 3.80 -0.35 0.10 0.03 -0.01 -0.33 -0.24 0.04 0.14 Neighbourhood group 3.45 -0.17 -0.08 0.04 0.03 -0.26 -0.06 0.04 0.06 Figure 4.07 shows the mean frequency of contact by different types and by level of engagement in WE. In WE countries, the results are significant in three cases. In trade unions and professional associations, the mean frequency of contact with family is significantly higher in cases of more active engagement. Also, in the case of participation in church and religious organizations, a similar trend can be observed. The opposite is true for sports groups. Figure 4.08 shows the mean frequency of contact by different types and by level of engagement in European countries. In European countries, the results are Cure for loneliness? 121 significant in three cases. In trade unions and professional associations, the mean frequency of contact with family is significantly higher in cases of more active engagement. The opposite is true for sports groups and other associations. Figure 4.06 Activity in an Organization within the Last 12 Months vs. Frequency of Contact (CEE) ­ ­ €      Figure 4.07 Activity in an Organization within the Last 12 Months vs. Frequency of Contact (WE)      122 Anna Domaradzka Figure 4.08 Activity in an Organization within the Last 12 Months vs. Frequency of Contact (Europe)      Social Activity by Frequency of Contact – Summary The analyses presented above enable us to see the differences in social activity by type and by frequency of contact with family in each of the three regions. The main results can be summarized as follows: 1. In CEE countries, two trends are observed: a. In trade unions and professional associations, the higher the mean frequency of contact, the higher the level of activity. b. Regarding participation in church and religious organizations, the higher the mean frequency of contact, the lower the level of activity. 2. In WE three trends are observed: a. In trade unions and professional associations, the higher the mean frequency of contact, the higher the activity. b. In regards to participation in church and religious organizations, the higher the mean frequency of contact, the higher the activity. c. Concerning participation in sports groups, the higher the mean frequency of contact, the lower the activity. 4.6 Summary Central and Eastern Europe 1. Concerning empty nest syndrome, political parties are for men what church organizations are for women. This is visible in cases of mothers of only sons and fathers of only daughters who no longer live with their children. The gender of the only child seems to be an important variable when analyzing the social activity of the parents. 2. The gender of the only child is important. a. Fathers of daughters are more likely to participate in neighborhood groups than the average man. b. Parents no longer living with their adult only sons are less politically active, while the opposite is true with daughters. c. Those who are most active in church and religious organizations are the mothers of only sons and the fathers of only daughters who no longer live with their children. 3. The level of political activity is lower than average for parents without adult children, while it is higher than average for parents with at least one adult child. There is no visible difference between women and men in this respect. 4. Trade unions and professional organizations are significantly more popular among men, while church and religious organizations are more preferred among women. Western Europe 1. The level of religious activity is lower than average for parents without adult children, but higher than average for parents with at least one adult child. There is no visible difference between women and men in this respect. 2. In cases of fathers whose adult only sons no longer live with them, activity in church organizations is higher, while in cases of daughters, it is lower than average. For women this is partly reversed: mothers of only sons who no longer live with them are less likely to become involved in religious organizations than the average woman. 3. Mothers of only sons are more likely to participate in charitable organizations than mothers of only daughters. 4. Political parties are significantly more popular among men, while church and religious organizations are more preferred among women. 124 Anna Domaradzka Western Europe vs. Central and Eastern Europe 1. For parents of adult only children in WE, one of the main areas of activity are charitable organizations, while in CEE, political, professional and church organizations are more popular. 2. Higher political activity among parents in CEE as well as in trade unions and professional organizations may be associated with the historical heritage of Communism and also the political instability of many of those countries. On the whole, however, CEE citizens are less active in politics than WE citizens. 3. Empty nest syndrome is less visible in WE than in CEE. 4. The frequency of personal contact with family members is lower in WE than in CEE. 5. Mothers of only sons are more active in charitable organizations and political parties both in CEE and in WE. 6. Mothers with only daughters who no longer live with them are less active in charitable organizations than average, a trend that is more visible in CEE. 7. In WE, fathers with adult only sons who no longer live with them are more active than average in church and religious organizations, while the opposite is true for fathers of only daughters who no longer live with them. However, in CEE, fathers with only sons who do not live with them are less active in church, while fathers with only daughters who do not live with them are more active. 8. Regarding neighborhood groups, fathers of adult only daughters are more active than average in CEE as well as WE. 9. Mothers with adult only sons who no longer live with them are less active than average in neighborhood groups, both in CEE and in WE. 4.7 Conclusions Most of the assumptions are partly positively confirmed. The level of social activity seems to be indeed related to the strength of family ties, as measured by frequency and type of contact with other members of the family, although there is no simple dependency. Parents of adult children, especially of those who no longer live at home, are more socially active in some of the organizations, but there are important inter-regional differences. Scheler’s hypothesis that mothers of only sons are most affected by empty nest syndrome is also illustrated. Social participation is still a rather rare phenomenon in the researched countries. The only significant result concerning people who tend to be active in Cure for loneliness? 125 various organizations is that activity in one type of organization correlates with activity in a different type of organization. Thus, when seeking potential activists, attention should be directed towards individuals who are already active in some form of organization. While likely that passive members also comprise a group of potential activists, this cannot be proved by the obtained data. By analyzing gender differences, it can be shown that men tend to participate in political or professional groups, while women are generally more active in religious and charitable organizations. This difference is stable in time and visible in other research on social participation. Data from 2001 ISSP Social Networks module illustrate that for each type of social activity, the mean is higher for WE countries than for CEE countries. Those differences in political and social participation can be explained as remnants of the communist regimes, when social participation had been mostly forbidden. Lower economic stability of CEE countries may also be a discouraging factor. In countries with developed democracies, people tend to be more engaged in social and political life than in countries where everyday economic problems are still a main area of concern for the majority of citizens. Political tradition and governmental attitude towards social participation are also important here. In WE, the tradition of social dialogue and cooperation between the voluntary sector and the government is quite developed, as compared to that in CEE. On the other hand, while social activity levels are higher in WE, the frequency of personal family contact is visibly lower than in CEE. This may be a result of the weakening of family ties in the more developed, western countries, coupled with the continued existence of more traditional family forms in the central and eastern countries. It can be argued that the results presented above support Beck’s thesis concerning changes in modern societies. Beck argues that the more developed the world is and the more possibilities of self-development it offers, the weaker traditional family networks become. In modern societies, people live alone or in nuclear families instead of in “outdated” extended families. They also lose touch with their closest family, due to increased mobility, growing individualization or the need for independence. The families are smaller and the blood bonds are weaker. Nevertheless, humans, still with the same emotional needs as before, seek contact with other people and long to feel useful. In modern societies, though feelings of alienation are growing, some new models of building networks are emerging. Participation in different types of organizations – voluntary associations, charitable organizations, religious groups etc. – is becoming an alternative to being alone. 126 Anna Domaradzka Taking this theory into account, we can treat WE countries as examples of more developed societies than those in the CEE region. Frequency of contact with family is visibly lower in WE than in CEE, while levels of activity in all types of organizations are higher. This paper also discusses the controversial issue of the empty nest syndrome. It is argued that the interpretation of the moment when children leave their parents as either positive or negative is strongly related to the society’s dominant ideology, especially concerning motherhood. It seems quite reasonable, however, to point out that this change in parents’ lives usually results in a shift of inter-family relations and may, therefore, result in some kind of empty space. While this moment is often combined with lower activity in the labour market, for many people, it is a chance to make the most of their free time and savings. For those who cannot afford it, or do not value it, the alternative may be to withdraw from social life or to become engaged in some kind of activity, be it the informal neighborhood group or professional association. Perhaps this is why empty nest syndrome seems to be less visible in WE, where people have more possibilities (financial as well as institutional) when organizing the later phases of their lives. The analyses presented above suggest that there is a difference in the type and level of activity depending on the family situation, especially for parents of only children. There is an interesting effect of child’s gender, which seems to support Scheler’s thesis. Furthermore, a similar effect is visible with fathers of only daughters. As mentioned above, mothers of only sons are more active in charitable organizations and political parties in both regions, while fathers of only daughters are more active in neighborhood groups. In other cases, the region is an important factor when analyzing the type of organization. In conclusion, the analysis of the deviations from the means shows that in cases of diminishing personal relations with family, people build new social networks based on participation in different social activities. Moreover, it can be argued that a child’s moving out of the home, reaching adulthood and becoming independent, seem to be important in the parents’ activity level in voluntary organizations. Section II Family, Gender, Life-Satisfaction Life Satisfaction, Happiness and Marital Status in Four Central European Countries: International Comparison Based on 5 the European Social Survey Dana Hamplová 5.1 Introduction Psychological, sociological, and medical studies systematically document that quality of life, life expectancy, mental and physical health, and life satisfaction and happiness are linked to marital status. Married men and women live longer and have a lower mortality rate [Joung, et al. 1997; Lillard, Waite, 1995; Kučera, 1994; Ross, et al. 1990], suffer from mental disorders less often [Brown, 2000; Earle, et al. 1998; Coombs, 1991; Ross, et al. 1990], are healthier [Waite, Gallagher, 2000; Cheung, 1997; Coombs, 1991; Ross, et al. 1990] as well as happier and more satisfied [Hamplová, 2006, 2004; Ryan, et al. 1998; Mastekaasa, 1994; Coombs, 1991]. The positive association between marital status and life satisfaction seems to be almost universal. Stack and Eshleman [1998] find that married men and women are happier in sixteen of the seventeen countries examined (the only exception in their sample being Northern Ireland). Ryan et al. [1998] confirm a positive correlation between marriage and life satisfaction in seven of the eight countries examined (the only exception was Finland). A positive correlation between marriage and life satisfaction is confirmed by Mastekaasa’s [1994] comparative study of nineteen countries. In this chapter, we ask whether married men and women are also happier in Central Europe and possibly whether the differences between various marital status categories vary across countries in the region. Four countries are included in the present comparison: the Czech Republic, Austria, Hungary and Slovenia. 5.2 Life Satisfaction and Marriage Why should married people be more satisfied than others? There are two types of explanations: selection and causality. The selection theory is based on the assumption that people are not happier and healthier because they are married, but that people who marry have certain characteristics that also predetermine that they will be happier and more satisfied [for discussion, see Stack, Eshleman, 1998; Mastekaasa, 1994, 1992; Ross, et al. 1990]. For example, temperament and personality traits such as neuroticism or extroversion1 1 For example extroverts seem to be happier even if they live alone. While personality traits such as extroversion, neuroticism, optimism and positive self-image influence how happy people feel, their impact differs across cultures. For example, the association between extroversion, self-respect, and life satisfaction is stronger in individualistic cultures [Diener, 132 Dana Hamplová influence the assessment of one’s own life [Mastekaasa, 1992; Blatný et al. 1998; Blatný, Osecká 1998]. The selection theory also assumes that these personality traits will affect the ability to establish and maintain long-term relationships. However, the selection theory has been criticized for three reasons. First, longitudinal studies measuring mental and physical condition of people before and after marriage do not clearly support the selection hypothesis. They have found either no or only a relatively weak selection effect [Brown, 2000; Cheung, 1997; Horwitz, et al. 1996; Marks, 1996]. Second, widows and widowers lose the advantages of marriage with their partner’s death and face the same risks as single people2 [Lillard, Waite, 1995]. Third, medical research indicates that single people do not die more often as a result of genetically related diseases but rather of diseases that can be explained by behavioral factors [Waite, Gallagher, 2000; Rogers, 1995; Ross, et al. 1990]. If better mental and physical health of married people cannot be attributed to the selection of certain people for marriage, how do we explain it? Supporters of the causal theory offer three different explanations. First, married life leads to a healthier lifestyle. Presence of a partner in the household helps prevent risk behavior such as smoking or excessive drinking. Partners monitor each other’s health; they encourage each other to promptly visit a doctor if necessary. If one of them becomes seriously ill, there is a higher chance of survival thanks to family care and support [Waite, Gallagher, 2000; Joung, et al. 1997; Goldman, et al. 1995; Kučera, 1994; Rogers, 1995; Ben-Shlomo, et al. 1993, in: Joung, et al. 1997]. Second, married people tend to be better off and have a higher standard of living because they manage their finances together, share expenses and, in case of unemployment or sickness, take economic responsibility for their partners wwStack, Eshleman, 1998; Joung, 1997, Rogers, 1995; Wilmoth, Koso, 2002; Holden, Kuo, 1996]. A better economic situation improves both “objective” quality of life indicators such as mortality and physical health as well as subjective life satisfaction [Miech, Shanahan, 2000; Luo, Waite, 2005]. Third, higher life satisfaction among married people is also a result of emotional support of partners, which enhances feelings of self-worth and sig-Lucas, 2000]. According to some estimates, around 80 percent of differences in how happy people feel can be attributed to genetic factors [Lykken, Tellegen, 1996, in: Averill, More, 2000]. 2 It is also true that divorced people are equally well off or worse off than single people, but in their case, one may consider the selection effect (a certain type of people could be more susceptible to divorce and mental problems) [Mastekaasa, 1994]. life Satisfaction and Marital Status in Four Central european Countries 133 nificance, and helps a person to overcome life challenges [Ryan, 1998; Ross, et al. 1990; Joung, et al. 1997]. Thus, marriage helps one to maintain and develop personal identity, seen among both men and women [Berger, Kellner, 1979]. Furthermore, married people are also more involved in social networks and are able to draw on social support more often [Ross, et al. 1990; Stack, Eshleman, 1998]. 5.3 Husbands and Wives In the 1970s, the American sociologist Jessie Bernard [1975] formulated a hypothesis that marriage is beneficial only to men while being harmful to women. According to Bernard, research findings showing married women to be happier do not mean that they are actually happy. Married women allegedly adjust their responses to the existing social norms that expect them to be happy.3 Bernard was later criticized for using few selective sources while also ignoring a large body of contradicting findings [for more information, see Marks 1996]. Despite these shortcomings, she did open an important discussion about potential differences between men and women. Today, there is a general agreement that marriage positively affects both men and women. Married men and women have a higher life expectancy, are healthier and suffer less from depression, anxiety and other mental problems [ cf. Earle, et al. 1997; Joung, et al. 1997; Lillard, Waite, 1995; Mastekaasa, 1994; Goldman, et al. 1995, in: Joung, et al. 1997; for an overview of older studies, see Ross, et al. 1990]. However, the question remains whether this positive effect of marriage is as strong for women as it is for men. Some studies show that the differences between married and single men are greater than between married and single women [Marks, 1996; Kučera, 1994; Ross, et al. 1990; Gove, et al. 1983].4 Other studies assert that marriage brings more benefits to women [Hirschl, et al. 2003; Cheung, 1997; Glenn, 1975, in: Mastekaasa, 1994]. Still other studies find no variation between men and women [Stack, Eshleman, 3 In addition, Bernard believed that the relative satisfaction of married women might also be explained by the fact that the alternatives are even worse than marriage. According to Bernard, the price that married women pay for their relative satisfaction is their mental health, as they are more likely to suffer from mental disorders and depression. 4 According to Gove et al. (1983), marriage improves the quality of life for both men and women, but men benefit more from married life. Stack and Eshleman (1998) criticize this study from a methodological standpoint (differences between men and women were not subjected to statistical testing) and assert that existing research in no way supports the hypothesis that marriage brings more benefits to men than to women (see Stack 1990). 134 Dana Hamplová 1998; Horwitz, et al. 1996; Mastekaasa, 1994; Lillard, Waite, 1995]. Based on longitudinal research, Horwitz et al. [1996] suggests that men and women benefit from marriage approximately the same amount, but in different areas of life (marriage reportedly protects men from depression and women from alcoholism). According to Stack and Eshleman [1998], however, the issue of variation between men and women is complicated by the fact that men enjoy better mental health irrespective of their marital status. Moreover, men and women often suffer from different types of mental disorders, which further clouds the comparison [Rosenfield, et al. 2000]. 5.4 Is Cohabitation as Good as Marriage? During the second half of the 20th century, most European societies underwent sweeping demographic changes. Cohabitation has become increasingly popular. The fact that it has also been gradually becoming an arena for child-bearing indicates that it acts as a substitute for marriage to some degree. This, therefore, raises the question whether cohabitation enhances quality of life in the same way as marriage. American researchers have established fairly well that, in the United States, people living with a partner are better off than single people, but worse off than married couples5 [Waite, Gallagher, 2000; Brown, 2000; Stack, Eshleman, 1998]. Stack and Eshleman [1998] compared seventeen countries and have reached the same conclusion. However, their analyses are based on pooled data from seventeen countries. It is therefore impossible to determine whether their conclusions are valid for all individual countries. The results may be driven by a few countries in which this effect is especially pronounced. 5.5 Research Question Data from the European Social Survey (ESS) provide an opportunity to study the effect of marital status on overall life satisfaction and happiness in Central Europe (Austria, Hungary, Czech Republic and Slovenia). The survey also enables an investigation into whether the effect holds even after financial situation and health have been controlled for. Three research hypotheses are tested: 5 This is attributed to the relative instability of informal unions (Brown 2000). life Satisfaction and Marital Status in Four Central european Countries 135 Hypothesis 1: Married people are happier than people without partners in all the countries examined. The advantages of married couples can be explained by their better health and better economic situation. Hypothesis 2: The effect of cohabitation is weaker than the effect of marriage in all four countries. Hypothesis 3: Men and women benefit from marriage differently. Hypothesis 4: The effect of marriage on life happiness and satisfaction varies across the selected countries. The same holds true for cohabitation. Data from ESS I (2002) has been analyzed. The respondents were asked to answer the question “Taking all things together, how happy would you say you are?” using an 11-point scale, where 0 points indicates “completely unhappy” and 10 points “completely happy.” The analysis used weights to control the population distribution in the individual countries ( i.e. dweight in ESS data). 5.6 Descriptive Statistics Figure 5.01 shows “average happiness” in the four selected European countries on a scale from 0 (minimum) to 10 (maximum). The figure clearly demonstrates that positive responses prevail. People from all of these countries feel fairly or very happy. The data is not normally distributed (right-skewed distribution). To test whether the differences are significant, the Kolmogorov-Smirnov test was used. The results indicate that Austrians are on average happier then the rest. Czechs and Slovenes do not significantly differ in the average level of happiness. Hungarians seem to be the least happy. When interpreting the results, it is necessary to take into account three important facts. First, while it may seem at first that average happiness corresponds to the economic situation in the given country, a more detailed comparison of twenty-one countries does not confirm this assumption (see Hamplová 2006). Second, the declared “happiness” does not reflect only objective mental wel - being but also cultural notions about what it means to be truly happy and whether it is acceptable for a person to admit that he or she is unhappy. Third, Austrians may be the happiest people in the region, but overal , they belong to a group of European nations showing only average happiness ( ibid). 136 Dana Hamplová Figure 5.01 Taking All Things Together, How Happy Would You Say You Are? Average 5.7 Descriptive and Control Variables The main objective of this chapter is, however, not to assess “average” happiness of Central European nations but to show association between marital status and happiness. OLS regression was applied to the research question. Control variables include elementary demographic characteristics: age and sex. Marital status is measured using three categories: married, cohabitating, and single. As a result, never married, divorced or widowed people may be included both in the cohabitation category and the single category. Our categorization is based on the presumption that the key factor is the presence of a partner in the household and the form of this presence. It is usually assumed that marital status influences life satisfaction through three mechanisms: economic situation, health, and emotional support. ESS data cannot be used to measure the quality of the relationship between partners, but it can provide information about the subjective health and the satisfaction regarding income of the respondents. Health was measured in five grades (very good, good, satisfactory, poor, very poor), income satisfaction in four grades (income is comfortable, able to make ends meet, difficult to make ends meet, very difficult to make ends meet). Controlling for these variables, life Satisfaction and Marital Status in Four Central european Countries 137 we can estimate to what degree the potential effect of marital status can be attributed to better financial situation or better health. Results Hypothesis 1 Table 5.01 shows a simple model that includes the effects of demographic characteristics (age and sex) and the three marital status categories. Three points can be made. First, men and women do not significantly differ in the levels of their happiness in any of these countries. Second, older people are less happy than younger people. Third, marital status influences happiness. On the whole, however, this model explains only a small portion of the variance. TABle 5.01 OLS Regressions. Dependent Variable: “Taking All Things Together, How Happy Would You Say You Are?” Austria Czech Rep. Hungary Slovenia coef. s.e. coef. s.e. coef. s.e. coef. s.e. Age -0.011 ** 0.004 -0.022 ** 0.004 -0.043 ** 0.004 -0.044 ** 0.004 Male -0.045 0.092 -0.079 0.125 -0.183 0.117 -0.110 0.110 Marital status Married 0.577 ** 0.108 0.950 ** 0.141 1.147 ** 0.139 0.660 ** 0.134 Cohabiting 0.309 0.165 0.084 0.272 0.627 0.225 0.221 0.250 Single (comp.) Constant 7.826 ** 0.154 7.385 ** 0.246 7.591 ** 0.191 8.453 ** 0.171 BiC -2.62 -37.66 -113.83 -79.98 Adj. r2 0.02 0.06 0.09 0.08 Number 1912 1061 1409 1258 * significant at 0.05 ** significant at 0.00 Source: eSS 2002 Table 5.02 presents a more extensive model, including a subjective assessment of health and income satisfaction. It is clear that the positive effect of marriage continues to be significant even after controlling for subjective health and income satisfaction. The first hypothesis is therefore supported only partially. Married people are happier than single people, but this effect cannot be fully explained by their better financial situation or health. 138 Dana Hamplová TABle 5.02 OLS Regressions. Dependent Variable: “Taking All Things Together, How Happy Would You Say You Are?” Austria Czech Rep. Hungary Slovenia Coef. s.e. coef. s.e. coef. s.e. coef. s.e. Age -0.003 0.003 -0.003 0.005 -0.019 ** 0.005 -0.027 ** 0.005 Male -0.085 0.088 -0.353 ** 0.119 -0.291 ** 0.109 -0.279 ** 0.101 Marital status Married 0.440 ** 0.103 0.728 ** 0.140 0.989 ** 0.128 0.673 ** 0.124 Cohabiting 0.201 0.155 -0.036 0.252 0.726 ** 0.207 0.303 0.219 Single (comp.) Subjective health Very good (comp.) good -0.605 ** 0.096 -0.760 ** 0.158 -0.475 ** 0.218 -0.515 ** 0.146 Fair -1.089 ** 0.136 -1.307 ** 0.193 -1.023 ** 0.238 -1.092 ** 0.176 Bad -1.507 ** 0.406 -1.809 ** 0.355 -1.970 ** 0.293 -1.767 ** 0.263 Very bad -1.423 1.410 -3.263 ** 0.876 -2.485 ** 0.447 -2.303 ** 0.503 Subjective Income living comfortably (comp) Coping -0.339 ** 0.094 -0.595 ** 0.174 -0.673 ** 0.189 -0.506 ** 0.105 Difficult -0.779 ** 0.157 -1.030 ** 0.202 -1.339 ** 0.207 -1.348 ** 0.193 Very difficult -1.319 ** 0.331 -1.825 ** 0.376 -2.035 ** 0.262 -1.757 ** 0.413 Constant 8.438 ** 0.165 8.451 ** 0.282 8.555 ** 0.288 9.019 ** 0.186 BiC -137.36 -157.57 -299.88 -242.74 Adj. r2 0.10 0.19 0.23 0.22 Number 1912 1061 1409 1258 * significant at 0.05 ** significant at 0.00 Source: eSS 2002 life Satisfaction and Marital Status in Four Central european Countries 139 Hypothesis 2 The effect of cohabitation on life satisfaction is much more diverse and cohabitants are a more heterogeneous group then married people. A 95% confidence interval for the marriage coefficient is positive in all four countries. It is thus safe to assume that marriage enhances life satisfaction, and married people are, generally speaking, better off than people living alone. With the exception of Hungary, however, a 95% confidence interval for cohabitation ranges from negative to positive values. In Austria, for example, the effect of marriage on overall life satisfaction ranges between 0.23 and 0.64 with a 95% confidence interval while the effect of cohabitation ranges from -0.10 to 0.50. The same conclusions apply to the Czech Republic, where the cohabitation coefficient ranges from -0.53 to 0.46 and Slovenia (from -0.12 to 0.73). Hypothesis 3 The third hypothesis focuses on the potential variations between men and women (see Table 5.03). The possibility that marriage improves quality of life differently for men and women is tested using interaction effects Marriage*Man. However, the assumption is not confirmed in any of the countries. The interaction effects do not improve the model. An interaction effect between sex and cohabitation is also tested, but does not prove significant. We therefore cannot conclude that men or women in Central Europe benefit more from marriage than the other sex. Hypothesis 4 Hypothesis 4 assumes that the effect of marriage and cohabitation on life happiness and satisfaction will vary across the selected countries. While models presented in Tables 5.01 –5.03 suggest this possibility, it is necessary to estimate models with all countries and interactions in order to test the differences. Comprehensive models for the entire region are presented in Table 5.04. The first model includes age, sex, marital status, health and economic situation, but does not allow for variation of the effects across countries. V 140 Dana Hamplová TABle 5.03 OLS Regressions. Dependent Variable: “Taking All Things Together, How Happy Would You Say You Are?” Austria Czech Rep. Hungary Slovenia Coef. s.e. coef. s.e. coef. s.e. coef. s.e. Age -0.003 0.004 -0.002 0.005 -0.020 ** 0.005 -0.028 ** 0.005 Male -0.103 0.138 -0.287 0.205 -0.479 0.172 -0.475 ** 0.158 Marital status Married 0.428 ** 0.127 0.769 ** 0.188 0.842 ** 0.168 0.519 ** 0.165 Cohabiting 0.202 0.156 -0.039 0.251 0.730 ** 0.207 0.299 0.219 Single (comp.) Subjective health Very good (comp.) good -0.604 ** 0.096 -0.755 ** 0.157 -0.484 0.218 -0.538 ** 0.145 Fair -1.088 ** 0.136 -1.303 ** 0.191 -1.029 ** 0.237 -1.103 ** 0.175 Bad -1.505 ** 0.406 -1.807 ** 0.354 -1.980 ** 0.292 -1.781 ** 0.262 Very bad -1.421 1.413 -3.259 ** 0.874 -2.500 ** 0.447 -2.322 ** 0.505 Subjective Income living comfortably (comp) Coping -0.340 ** 0.094 -0.596 0.174 -0.662 0.190 -0.510 ** 0.105 Difficult -0.781 ** 0.156 -1.030 ** 0.202 -1.332 ** 0.208 -1.345 ** 0.193 Very difficult -1.319 ** 0.331 -1.819 ** 0.377 -2.024 ** 0.262 -1.760 ** 0.415 Men* Married 0.028 0.180 -0.095 0.250 0.315 0.222 0.337 0.207 Constant 8.448 ** 0.177 8.411 ** 0.308 8.669 ** 0.301 9.170 ** 0.208 BiC -129.83 -150.76 -294.67 -238.25 Adj. r2 0.10 0.19 0.23 0.22 Number 1912 1061 1409 1258 * significant at 0.05 ** significant at 0.00 Source: eSS 2002 life Satisfaction and Marital Status in Four Central european Countries 141 TABle 5.04 OLS Regressions. Dependent Variable: “Taking All Things Together, How Happy Would You Say You Are?” Model 1 Model 2 Model 3 Model 4 Age -0.013 ** 0.002 -0.013 ** 0.002 -0.013 ** 0.002 -0.013 ** 0.002 Male -0.217 ** 0.051 -0.221 ** 0.051 -0.221 ** 0.051 -0.223 ** 0.051 Marital status Married 0.671 ** 0.061 0.798 ** 0.128 0.835 ** 0.087 0.859 ** 0.091 Cohabiting 0.327 ** 0.100 0.315 ** 0.100 0.314 ** 0.100 0.033 0.237 Single (comp.) Subjective health Very good (comp.) good -0.536 ** 0.067 -0.540 ** 0.067 -0.540 ** 0.067 -0.544 ** 0.067 Fair -1.091 ** 0.084 -1.092 ** 0.084 -1.092 ** 0.084 -1.094 ** 0.084 Bad -1.862 ** 0.144 -1.862 ** 0.144 -1.862 ** 0.144 -1.866 ** 0.144 Very bad -2.570 ** 0.290 -2.565 ** 0.288 -2.566 ** 0.288 -2.574 ** 0.289 Subjective Income living comfortably (comp) Coping -0.469 ** 0.061 -0.467 ** 0.061 -0.467 ** 0.061 -0.467 ** 0.062 Difficult -1.058 ** 0.086 -1.059 ** 0.087 -1.057 ** 0.087 -1.055 ** 0.087 Very difficult -1.695 ** 0.153 -1.687 ** 0.153 -1.687 ** 0.153 -1.694 ** 0.153 Country Czech republic (comp.) Austria 0.235 ** 0.076 0.393 ** 0.126 0.419 ** 0.108 0.457 ** 0.120 Hungary -0.355 ** 0.082 -0.380 ** 0.136 -0.339 ** 0.082 -0.377 ** 0.085 Slovenia -0.259 ** 0.082 -0.074 0.136 -0.048 0.119 -0.037 0.128 Interaction Austria* Marriage -0.240 0.154 -0.278 * 0.122 -0.336 * 0.133 Hungary* Marriage 0.064 0.168 Slovenia* Marriage -0.294 0.164 -0.331 ** 0.134 -0.362 * 0.142 Austria* Marriage 0.158 0.284 Hungary* Marriage 0.637 * 0.298 Slovenia* Marriage 0.242 0.324 Constant 8.651 ** 0.124 8.549 ** 0.150 8.524 ** 0.134 8.520 ** 0.137 BiC -1383.784 -1366.751 -1375.22 -1354.76 Adj. r2 0.23 0.23 0.23 0.23 N = 5640; * significant at 0.05 ** significant at 0.00 Source: eSS 2002 142 Dana Hamplová The second model adds the interaction effects between a country and marriage. Thus, it directly tests the possible differences in the effects of marriage. To test significance of the interaction effects, models are compared by an LR-test, since the individual coefficients cease to be reliable [Jaccard, Turrisi, Wan, 1990]. The effect of marriage on life satisfaction in the Czech Republic is the comparative category. Model 2, which includes interaction effects, seems to be clearly superior from the point of view of a likelihood ratio (LRtest), which suggests that the effect of marriage on overall life satisfaction varies among these countries. Model 3 leaves out the interaction effect Hungary*Marriage and tests the possibility that the effect of marriage in Hungary is not different from the effect of marriage in the Czech Republic. Fit statistics shows that the interaction effect is not significant. Model 4 tests whether cohabitation influences life satisfaction differently in these countries. Three interaction effects Country*Cohabitation are included. The effect of cohabitation in Czech Republic again serves as a comparative category. This model does not improve the overal model fit either by the Bayesian Information Criterion, or by the likelihood ratio. The lack of significance may, however, be explained by the high heterogeneity among cohabiting couples. 5.8 Conclusions This chapter has examined the influence of marital status on men’s and women’s overall life satisfaction in Central Europe. Extensive theoretical and empirical literature suggests that marriage enhances quality of life in many respects and, among other things, also improves mental well-being and happiness. Data from the European Social Survey confirm this notion. Married people are happier in all studied countries. This result applies to men and women alike. While higher life satisfaction of married people can be partially attributed to their subjective health status and income satisfaction, physical health and income cannot explain all the differences between married people and others. Softer factors such as mutual companionship, support, help and 6 Women’s Role in the Family: A Comparative Study Filip Raciborski 6.1 Introduction The outbreak of World War II revolutionized the structure of employment. The labour market experienced a significant shortage as large numbers of men were drafted into the military and sent to fight. The economy, propelled by huge demands from the armed forces, needed a new workforce. Women provided that workforce. Housewives abandoned their household duties and filled the gap in the market, which they managed to do very well. However, with the end of World War II, the labour market became saturated, and the easiest solution to this problem seemed to be to convince women to leave professional work and to return to the household. It was not only an issue of the labour market, but also of rebuilding the demographic potential that had been, to a large extent, destroyed during the war. However, the changes that had taken place could not be reversed so simply. Attitudes towards role division in the family had changed, as well as the very aspirations of women. Additionally, combining family life and work proved very difficult, requiring more help from men. Though these changes proceeded differently in each country, the basic model was similar. These changes are still in progress; we are far from reaching the stage of complete equality and fair division of responsibilities. It should be recognized that the degree and the pace of these changes differ from country to country. This problem is exactly what I want to concentrate on in my work. The data used here was collected in the study called Family and Changing Gender Roles III, conducted in 2002, as a part of the International Social Survey Program. In the analyses, I have included Poland, the Czech Republic, Slovakia, Hungary, Austria, as well as Spain, Ireland, France and Great Britain. Though the first set of countries corresponds to the CEE-focus of the book, I believe that these countries should be included in a comparative study of this kind anyway. In the case of the remaining countries, Ireland and Spain lend themselves especially well to a comparative study with Poland because of some essential similarities, such as fast economic growth and the important role of Catholicism. France and Great Britain provide good examples of multi-cultural countries. Furthermore, France is considered to be the state in which postmodern values are the most widespread. I am interested to see if this translates into a distinct way of perceiving women’s roles in the family. 6.2 Methodology The ISSP study Family and Changing Gender Roles III was conducted in the countries mentioned above mainly in 2002. Three countries had different timeframes: Ireland (the research started at the end of 2001 and continued into 2002), Spain (in 2003), and Austria (at the end of 2003 and into 2004). The sample sizes differed and depended on the size of the country, ranging from 1023 in Hungary to 2471 in Spain. All the analyses are made using an ISSP data set from the year 2002. This data set is weighted using the weights included in this set. In the majority of the countries, those surveyed are of age, i.e. over 18 (thus, people under 18 are excluded from the analyses (this pertains to Ireland and Austria)). All in all, after weighting the data, removing the under-age respondents and those from countries other than the chosen nine, 14218 cases remain in the data set. 6.3 Modern vs. Traditional Role Division in the Family Defining modern and traditional role divisions in the family is not any easy task. It is possible to distinguish several aspects that are the key to defining these categories. One of them is certainly the attitude towards women’s professional work. The modern approach places no limits to women’s access to the labour market. Conversely, the traditional – or conservative – approach maintains that women should only tend to the household and that men should be the “breadwinners,” supporting the family financially. Another aspect is having children. In this case, equality in the division of responsibilities meets a certain biological barrier, seeing as only women are able to bear children – a situation not likely to change. But while men are not biologically equipped to bear children, they can engage in taking care of their offspring. The conservative approach holds that women, independent of any professional work, should take care of children above all else. The modern approach assumes that the professional work of both partners should be considered and that a couple should arrive at an optimal solution for both spouses, which could mean that the man stays at home to raise the children, forgoing his other pursuits. The third aspect to consider is the division of household responsibilities between partners. The conservative stand maintains that women should do all the chores, maybe with the exception of minor household repairs. There is no stereotypical thinking connected with household responsibilities in the modern Women’s role in the Family: A Comparative Study 147 approach, though. This means that household chores are divided according to what is convenient for both partners, in such a way that the chores are an equal burden for both of them. As all of these dimensions overlap, drawing a dividing line between them becomes virtually impossible. Therefore, in my analyses I restrain from pointing to specific aspects, but instead talk about certain groups of them. 6.4 Women’s Roles in the Family The ISSP study Family and Changing Gender Roles III contains an entire set of questions concerning family life issues, such as the division of household chores, the opinion on women’s professional work, and the managing of household finances. Presenting all of this information here would be impossible, and even more so if age, sex, education and income are considered. Thus, it is necessary to create an index that will show these phenomena collectively. I focus on the first two sets of questions from the survey, which refer to opinions about women’s professional work and the division of household chores. Overall, there are 10 questions. Even though the questions seem to address very similar problems, the statistical analysis proves that, on their basis alone, it is impossible to create a reliable index that will reveal a conservative or modern approach to women’s role in the family.1 After a detailed analysis of the questions’ content, I decided to use the following four statements to create a scale: 1b. A pre-school child is likely to suffer if his or her mother works 1c. All in all, family life suffers when the woman has a full-time job 1d. A job is fine, but what most women really want is a home and children 2b. A man’s job is to earn money; a woman’s job is to look after the home and family Respondents could choose between the following answers: strongly agree, agree, neither agree nor disagree, disagree, strongly disagree, can’t choose. In my opinion, the chosen questions represent all three of the aspects mentioned above. Principal Components Analysis reveals one main component, which proves that the questions are consistent. The reliability analysis (Alpha Cronbach) of the scale gives the result of 0.75, which is not very high. Nevertheless, considering that the scale is composed of only four questions, this result is acceptable. The index was created by computing the mean on the basis of answers that 1 The Principal Components Analysis shows that it is not acceptable to treat the questions as one dimension, but rather as three or four independent ones. The reliability analysis (Alpha Cronbach ) of the scale, based on these questions, corroborates these findings. 148 Filip raciborski were given to the questions described above, with values ranging from 1 (an extreme conservative attitude) to 5 (an extremely modern attitude). Among the studied countries, the most modern approach to women’s roles in the family can be found in Ireland and Great Britain, with the mean index equaling 3.3. Poland falls in the middle (2.8), between Austria and Spain, while Hungary has the lowest results with less than 2.5 points. To make interpretation easier, one can imagine that a one-point difference in the index means moving one category higher or lower in all the questions, for instance a change from strongly agree to agree. The point 3.0 being the middle of the scale is of great importance for interpretation purposes. This is also the estimated result for Spain. Higher results are noted for France, Great Britain and Ireland. The remaining countries are lower on the scale. Figure 6.01 Traditional (1) vs. Modern Approach Towards Women’s Roles in the Family Great Britain Ireland Poland France Austria Czech Republic Spain Hungary Slovakia As expected, women’s opinions regarding the division of household responsibilities are more modern than those of men. These differences are not the same in all the countries, though, and in some cases they are not even always statistically relevant (specifically in the Czech Republic and Hungary). The statistically relevant differences (at the level of p < 0.00l) are especially notable in France (0.34), Austria (0.24), and Great Britain (0.21). 6.5 Factors Differentiating the Attitudes Towards Women’s Roles in the Family Though sex is an important factor that influences the respondents’ attitudes towards women’s roles in the family, it is only one of many variables. This work concentrates on the basic variables of education, professional work and political party affiliation. Figure 6.02 The Influence of Age on Women’s Roles in the Family Great Britain Ireland Poland France Austria Czech Republic Spain Hungary Slovakia Note: The mean value of the index of the traditional approach (1) vs. the modern approach (5) for particular groups The relationship between age and a modern attitude towards the family seems an obvious one. We expect older people to have more conservative opinions than younger ones. The question remains, what are the extent and the differentiation of this phenomenon in the studied countries. To address this, I used Pearson’s correlation, measuring the relationship between the respondents’ positions on the scale and the respondents’ ages in years. The results confirm the hypothesis that the acceptance of the conservative model of the family increases with age. All the computed correlations prove to be statistically relevant at the level of p < 0.001. The strongest linear relationship is noted in France (r = -0.44) and is closely followed by Spain (r = -0.41). In 150 Filip raciborski Ireland and Great Britain there is also a quite strong relationship, with the correlations at the level of r = -0.37 and r = -0.35. The weakest relationship is found in Slovakia (r = -0.17). In the Czech Republic and in Hungary, the results are similar, at r = -0.18 and -0.19 respectively.2 Figure 6.03 The Influence of Formal Qualifications on the Attitudes Towards Women’s Roles in the Family Great Britain Ireland Poland France Austria Czech Republic Spain Hungary Slovakia No formal Lowest formal Above lowest Higher Above higher University qualification qualification qualification secondary secondary level, degree completed below full uni completed Note: The mean value of the index of the traditional approach (1) vs. the modern approach (5) for particular groups. groups smaller than 20 people were excluded from the analysis The Figures presented below partly explain the cause of such a low correlation in these three countries. In the first 3 or 4 age groups, we observe little diversity. It is only first with the respondents over 45 or 55 (depending on the country) that the results clearly begin to differ from the results of the younger respondents. There should be a relationship between education and acceptance of the modern family model, in which as education level increases, so does the inclination towards the modern model. Indeed, in the majority of countries this relation does exist. The analysis of Pearson’s correlation between the number of years 2 Result for remaining countries: Austria (r= -.28), Poland (r= -.31) Women’s role in the Family: A Comparative Study 151 of education3 and the index shows a strong linear relationship in Spain (r = 0.38, p < 0.001) and in France (r = 0.37, p < 0.001). The weakest relationship (but stil statistical y significant at the level of p < 0.001) is noted in Great Britain (r = 0.18), the Czech Republic (r = 0.19) and Slovakia (r = 0.19).4 In the case of Great Britain, the Figure showing the index of acceptance of the modern family model versus education level clearly depicts this relationship, though it is a bit distorted by the group of very wel educated people. In the Czech Republic, the situation is not that clear, since the most modern attitude is present among both the least and the most educated respondents. There is little diversity within the middle categories. It is quite possible that distortions in the Figure are partly the result of different education systems, with their varying structures of schools and universities. It is not possible to establish the correlations for Austria, due to the lack of appropriate data. Nonetheless, the information depicted in the Figure suggests the existence of a very strong relationship between education and acceptance of the modern family model. The analysis of the situation connected with professional work brings a lot of insightful findings. The most progressive group is undoubtedly students, both male and female, with women usually scoring higher on the scale of the modern approach towards family. Presumably, this result is not only connected with the level of education but also with age. This is further supported by retired respondents being the most conservative group. The comparison between full-time and part-time workers and people who maintain households is interesting. Respondents working professionally score relatively high on the scale of the modern approach towards the division of responsibilities in the family, while those fundamentally occupied with domestic duties are more conservative. It is worth posing the question about the reasons for this. Is it an effect of acquiring certain attitudes while performing specific duties (in line with the Marxist thesis that “existence precedes consciousness”)? Or is it more basic, namely, that people with specific psychological predisposi-tions choose this particular mode of self-fulfillment? In the first case, external factors are responsible, while in the second, internal ones are at work. I shall leave this question open. 3 In the analysis, the variable used is recoded in such a way that the maximum number is 18. The last category also includes all answers indicating an even longer time of study. It is done in order to increase the level of standardization, since in some countries values over 17 are not taken into consideration at all. Austria is excluded from the analysis, as this variable is not accessible there. 4 Result for remaining countries: Hungary (r= .23), Ireland (r= .24), Poland (r= .35) 152 Filip raciborski Figure 6.04 The Influence of Professional Work (Chosen Categories) on the Attitudes Towards Women’s Roles in the Family Great Britain Ireland Poland France Austria Czech Republic Spain Hungary Slovakia Female Male Employed-full time Employed-part time Unemployed Studt, school, vocat. trng Retired Housewife, home duties Note: The mean value of the index of the traditional approach (1) vs. the modern approach (5) for particular groups. groups smaller than 20 people were excluded from the analysis The analysis of the index of acceptance of the modern division of responsibilities in the family and political party affiliation brings very diverse results. Some of these results are predictable for a few of the studied countries, as is the case for France, Spain, Great Britain, and quite significantly – Ireland. In this group of countries, we observe an almost linear relationship between political affiliation and the level of acceptance of the modern family model. With the shifting of political inclination from the extreme left to the extreme right, the acceptance of the modern role division in the family increases. In other countries, it is not possible to trace such a relationship. This situation may stem from a different understanding of the left-right division in countries belonging to the former Communist bloc. Women’s role in the Family: A Comparative Study 153 Figure 6.05 The Influence of Party Affiliation on the Attitudes Towards Women’s Role in the Family Great Britain Ireland Poland France Czech Republic Spain Hungary Slovakia Note: The mean value of the index of the traditional approach (1) vs. the modern approach (5) for particular groups. groups smaller than 20 people were excluded from the analysis. The results for Austria are nor presented because of the lack of essential data. 6.6 Household Chores Since the ISSP is an international survey, it is possible to compare responses from different countries. I find the series of questions concerning the division of household chores among spouses especially interesting. The questionnaire contains questions about which member of the household does the following chores: laundry, small repairs around the house, looking after sick family members, shopping for groceries, household cleaning, and preparing meals. Given that these questions concern role division in the family, only married and cohabitating couples were addressed. Respondents were to determine who in their household performs certain chores. They could also choose a third person in addition to themselves or their partner. They could declare that they do housework together or divide responsibilities equally. Respondents who indicated themselves or their partner could also define whether 154 Filip raciborski they perform a certain chore always or usually. This difference, being very difficult to define, is not reflected in the Figures or analyses. In the fol owing analysis, responses by women are mainly used. This approach has been chosen because in most families nowadays, women are general y responsible for the housework, independent of their professional work. This is confirmed by the results of research on time budget with regard to sex (Siemieńska R., 2006). Hence, women’s responses seem to be more trustworthy. Figure 6.06 Household Chores. Household Member Responsible for Specific Chores: Laundry Great Britain Ireland Poland 100% France Austria Czech Republic Spain Hungary Slovakia 100% Usually always man About equal or both together Done by a third person Usually always woman Female Male The results of the research show that, among the chores in question, laundry is the one most frequently performed by women, irrespective of country. In all analyzed countries, 81% to 94% of women in permanent relationships declare that they perform this duty. Men’s opinions fully corroborate this observation. The highest percentage of women who do laundry is found in France (94%), Slovakia (94%) and the Czech Republic (92%). Interestingly, in France there is also the biggest discrepancy between the opinions of men Women’s role in the Family: A Comparative Study 155 and women. According to women, 94% of them do laundry, while according to men it is 75% of the women who do the laundry. In other countries, this discrepancy is also observable, but does not exceed 10 percentage points. Among the other chores addressed in the study, cleaning and preparing meals are the ones in which women are engaged the most. On the basis of women’s responses, we can conclude that in most of the countries, women are responsible for preparing meals in about 80% of the relationships. However, there are countries in which the situation looks a bit more optimistic, for example, in Great Britain (62%) and in France (69%). This does not mean that men have taken over this responsibility, though. Rather, we observe cooperation, i.e. sharing the duties with a partner. In all countries under scrutiny, women are mainly responsible for cleaning. The percentage of relationships in which women perform this duty (usually or always) oscillates between 64% (UK) and 78% (the Czech Republic), according to the women’s responses. Surprisingly, in Poland, which is quite a conservative country, only 66% of women in permanent relationships are always or usually responsible for cleaning. Furthermore, on average, in 30% of cases this duty is shared by the partners. Cooperation is also common when shopping for groceries and looking after sick family members. In the case of shopping for groceries, the percentage of couples sharing this duty oscillates between 48% in Hungary, according to women, and 23% in Ireland. Ireland is clearly distinct from the rest of countries in question, since shopping for groceries poses a common burden for women (71% of women declaring that they are responsible for it). In other countries, this percentage is usually lower by approximately 20 percentage points. Shopping together for groceries is very popular in Great Britain and France (found in 44% of permanent relationships, according to women’s responses). Taking care of a sick family member together is especially frequent in Hungary and Poland (49% and 48% respectively, according to women’s responses). It is most rare in Slovakia (31%), the Czech Republic (33%), Austria (33%), and Spain (34%). In these countries, women predominantly perform this duty (about 62–65%). Also in Great Britain, the percentage of women declaring that they always or usually take care of sick family members is very high, at 62%. This is especially interesting in light of the previously presented data that shows Great Britain as one of the leading countries with respect to a modern outlook on role division in the family. The housework in which men are most engaged is minor household repairs. In the examined countries, the percentage of men who usually or always do this chore fluctuates between 60–80%, according to women’s responses. The highest percentages are noted in Poland (79%) and the lowest in Ireland 156 Filip raciborski (61%) and France (61%). Irish and Spanish women must exhibit the greatest resourcefulness, since they are mainly burdened with making minor household repairs (17% and 14% respectively, by women’s declarations). Figure 6.07 Household Chores. Household Member Responsible for Certain Chores: Shopping for Groceries Great Britain Ireland Poland 100% France Austria Czech Republic Spain Hungary Slovakia 100% Usually always man About equal or both together Done by a third person Usually always woman Female Male In the ISSP study, only 6 basic household chores are taken into consideration, which does not embrace the whole scope of the issue. Nevertheless, these activities are performed in every household, and thus it is possible to present the division of responsibilities in the family on the basis of this survey. In general, women perform the household chores in the majority of cases, regardless of country. Even though we definitely observe variations from country to country, the big picture does not change. Concerning almost every chore (excluding minor household repairs), men are only sporadically the sole responsible members. The percentage of such instances oscillates between 2% to 10%. However, with certain chores (mainly shopping for groceries and looking after sick family members), while men might not assume sole Women’s role in the Family: A Comparative Study 157 responsibility, they do contribute significantly. It is worthwhile to pose a question about the future development of this situation. Presumably, the percentage of women taking part in household chores will fall. However, it is still unknown whether this change will go in the direction of men taking over some of the responsibilities or towards sharing housework on a daily basis. Figure 6.08 Household Chores. Household Member Responsible for Certain Chores: Small Repairs Around the House Great Britain Ireland Poland France Austria Czech Republic Spain Hungary Slovakia 100% Usually or always woman About equal or both together Done by a third person Usually or always man Female Male The statistics show yet another very interesting phenomenon that could be very important for this issue: the discrepancy between men’s and women’s responses. Theoretically, with this sample size, the declared participation of men and women with a given chore should be the same regardless of respondents’ sex. Yet this is not the case. For instance, in France, 94% of women in permanent relationships declare that they always or usually do laundry. This percentage falls to 75% in the men’s responses. What is the nature of this phenomenon, then? This inconsistency is especially prevalent in Ireland, Spain and France, and is the least in Hungary and Poland. Hence, it is pos- 158 Filip raciborski sible to hypothesize that it is connected to changes in the way women’s roles are perceived in the common consciousness, which is reflected in particular behaviors. Significant differences in responses probably indicate apparent changes in progress, which, however, do not completely translate into real changes of attitude. It means that some people (in this case, men) adjust their responses to social expectations. Conversely, we should expect that in the countries where these differences are small, a lower participation of men in household chores does not meet with people’s disapproval. 6.7 Time Budget – Household Chores The mere fact that a household member performs a certain chore does not determine the level of involvement in running house, rather the total time spent on chores must be considered. In the ISSP survey, the question about the number of hours spent on household work was directed to the respondent, but also included the respondent’s spouse or partner. It turns out that women’s and men’s responses on the whole match each other, and the differences, if any, are small. The respondents in Ireland devote the most time to household chores. Women in permanent relationships declare that, together with their partners, they spend over 48 hours maintaining the household, 80% of which the women do themselves. According to the men, this time amounts to 38 hours, 75% of which is performed by the women. Thus, these declarations are quite similar. French people spend the least amount of time on housework: according to women, the total is 16 hours (19 hours according to men), which is less than half of that spent in Ireland. It is also worth taking a closer look at Spain. This country comes second with regard to the total time spent on housework – 36 hours according to the women. In this case, women are responsible for more than 83% of the work, which is also the highest percentage. At the other end of the scale is Poland, where the average total time according to the women – 34 hours – is only a little less than in Spain. However, the women’s share in these responsibilities is only a bit more than 62%. The statistics presented above clearly show that it is still difficult to talk about an equal division of housework. The disparity between the time spent on housework among men and women is significant. In the majority of cases, the women’s share of the total time exceeds 70%. Therefore, it is justified to say that women have essentially another part-time job at home (or even more than this), regardless of their professional work. Women’s role in the Family: A Comparative Study 159 Figure 6.09 The Average Time Spent per Week on Household Work Great Britain Ireland Poland France Austria Czech Republic Spain Hungary Slovakia Note: not including childcare and leisure time activities 6.8 Household Budget In the ISSP study, respondents in permanent relationships were also asked about who manages the household budget. In the majority of cases, women’s and men’s answers are consistent. Differences can be found, however, between particular countries. Financial independence is strongest in Austria, where about 15% of men and over 20% of women declare that they keep their money separate. The smal est number of such families is observed in Spain and Hungary (merely 1–2%). Overall, pooling money is the dominant form of managing household finances – every member of the household takes as much from the budget as they need. This strategy is most frequently observed in Poland (69% of men’s responses and 73% of women’s responses), Hungary (64%, 69%), France (67%, 62%) and in Spain (64%, 63%). Pooling money is the rarest in Austria, but even there over half of the couples practice it. 160 Filip raciborski Among the countries included in the survey, special attention should be given to Spain, Hungary and Slovakia, where women are often relatively responsible for household budget. In the case of Spain, women claim that they manage the shared budget in 24% of couples, whereas only 7% of men manage the shared budget. In contrast, the percentages in France are 2% and 3% (according to women’s declarations). Figure 6.10 Household Budget Great Britain Ireland Poland France Austria Czech Republic Spain Hungary Slovakia 100% Each keep own money separate We pool some money We pool all the money Spouse, partner manages money I manage all the money Female Male 6.9 Equal Division of Responsibilities – Assessment The time devoted to maintaining the household can be treated as a certain objective indicator of a household member’s participation in chores. It is worth remembering, however, that objective measures do not always reflect the whole truth. In some cases it might be useful to use one’s own subjective assessment of the situation in order to understand that which remains inex- Women’s role in the Family: A Comparative Study 161 plicable by objective measures. The understanding of fair division of chores differs a lot from country to country. It is possible that in those countries where the majority of chores are done by women, there is also a more common belief that this is a fair division. It becomes a matter of the cultural perception of men’s and women’s roles in the family as well as the awareness of alternatives to the established standard. Figure 6.11 The Assessment of Equality Regarding the Division of Household Chores among Couples Great Britain Ireland Poland France Austria Czech Republic Spain Hungary Slovakia On the whole, in each country women claim only rarely (not exceeding 5%) and sporadically that they do a bit less or much less than they should, if the division of chores were fair. The situation is similar for men claiming that they do a bit more or much more than they should, if the division of chores were fair (the percentage is rarely higher than 10%). The most interesting findings are seen in the analysis of women’s answers asserting that their share in household chores is greater than it should be if the responsibilities were fairly divided. The most frequent answers of this kind 162 Filip raciborski are noted in Great Britain (65%), Austria (63%), and in Ireland and France (62% each). This is not surprising in the case of Ireland where, as mentioned above, women in permanent relationships work on average for nearly 35 hours a week in the household. In France, however, this situation is peculiar, as women there work only an average of 16 hours a week in the household. The explanation of this peculiarity may be that, for female respondents, the ratio of their time spent on work to that of their partners is more important than the total number of hours spent. This is confirmed to an extent by the data, since in Ireland, France and Austria, the women’s share of the overall time spent on running house oscillates around 80%. This thesis is, however, weakened by the case of Spain, where the higher share of women in chores (over 80%) is combined with a sense of injustice (regarding the division of chores) that reaches only a medium level. In Hungary, the sense of justice connected with the division of chores is the most common – 61% of women and 57% of men in permanent relationships claim that they do their fair share in the household. This is followed by Poland (47% and 55% respectively). Ireland is also worth mentioning because of the utter inconsistency between women’s and men’s opinions. According to only 36% of women in permanent relationships, there is a fair division of chores, while the division is fair for 55% of men. 6.10 Satisfaction with Family Life and Professional Work In the conducted analysis, women’s satisfaction level with family life is examined in relation to professional work. It turns out that whether a woman works full-time, part-time or only inside the home has little influence on her level of satisfaction with family life. Students prove to be the most satisfied with their family life, but this may be the result of their youth and that some of them have not yet settled down. The most negative influence on the level of satisfaction with family life is unemployment, but this relationship is only observed in some countries. Therefore, it is possible to hypothesize that women’s professional work has very little influence on the level of satisfaction with family life. In other words, these two spheres (family and work) are to a great degree independent of each other. 6.11 Changes in the Period between 1994 and 2002 Variability of people’s attitudes over time is a key issue in this research. The ISSP’s objective is not only to survey the chosen topics in particular countries, but also to repeat this study every few years. The Family and Changing Gender Roles Module has been conducted three times already – in 1988, 1994 and 2002. Unfortunately, the majority of countries that are the focus of this paper were not included in the 1998 survey; hence, I have been forced to concentrate on the data collected in 1994 and 2002. All of the originally chosen countries are then compared, except for France and Slovakia, as the data for these countries from the year 1994 is missing. Figure 6.12 The Change of the Mean Value of the Index of the Traditional Approach Great Britain Ireland Poland Austria Czech Republic Spain Hungary Note: (1) vs. the modern approach (5) in the period of 1994–2002. France and Slovakia were excluded from the analysis due to lack of data for the year 1994. The comparison of the situation in 1994 and in 2002 leads to interesting findings. In the group of seven countries included in the analysis, the differences 164 Filip raciborski observed in the two of them, Great Britain and the Czech Republic are not statistically relevant. The results for Austria prove inconsistent when respondents’ sex is taken into consideration. While the change in men’s opinions is not statistically relevant, it is for women (p < 0.01 on the basis of t-test). In the remaining countries, the results for both women and men are clearly different in the years 1994 and 2002 (p < 0.01). In all of the countries in which the differences are statistically relevant, the level of acceptance of the modern family model increases from 1994 to 2002. Though the greatest change takes place in Hungary (by 0.4 points in the case of men and by 0.35 points in the case of women), it still remains lowest on the scale of acceptance. Significant progress is also noted in Ireland and Poland, and in the case of women – also for Austria (circa 0.35 points). Thus, Ireland, taking the lead from Great Britain, becomes the country with the highest acceptance of the modern family model of all countries under consideration. 6.12 Conclusions Instead of concluding this work as is customary, I have decided to present a multiple regression model. On the basis of this model, it is possible to establish which component has the greatest influence on the attitudes towards the modern family model, which will allow me to answer some of the questions posed above. In the model, except the variables discussed so far, religiousness has also been included (how many times a year does a person participate in different religious events). Moreover, the variable describing professional work is recoded in such a way as to merely inform if a person works (full-time or part-time) or not. I also tested other, more complex models, but adding new variables did not improve predictive ability and often resulted in a drastic decrease of cases, because of a significant lack of responses. The presented model explains a respondent’s position on the index of acceptance of the modern family type. This regression model is relatively simple, using only the information about age, sex, education, religiousness and stable employment. The effectiveness of the model turns out to vary a lot from country to country. In France the model explains over 30% of variance. The results obtained in Ireland (20%) and in Spain (21%) also seem promising. However, in the case of the Czech Republic and Slovakia the predictive ability is very limited (5% of the explained variance). To compare the power of influence of particular variables in the model, we analyzed the standardized Beta coefficients. The data collected in France shows that the biggest influence on the acceptance of the modern family Women’s role in the Family: A Comparative Study 165 model is education. The level of higher education and the level of acceptance increase together. The second most influential variable is employment – full-time or part-time workers have a more modern stance on women’s roles in the family. Interestingly, in the case of France, sex proves to be the least influential variable, though still statistically relevant. TABle 6.01 The Comparison of Beta Linear Regression Coefficients in Particular Countries Great Britain Hungary Ireland Czech Republic Beta p Beta p Beta p Beta p Years of study (max. 18) 0,06 0,019 0,17 0,000 0,09 0,005 0,15 0,000 Age -0,24 0,000 -0,11 0,003 -0,29 0,000 -0,08 0,039 Sex 0,13 0,000 0,00 0,943 0,13 0,000 0,07 0,023 does respondent have professional work (full- time, part-time) 0,11 0,000 0,10 0,006 0,11 0,000 0,07 0,084 religiousness -0,06 0,018 -0,04 0,244 -0,10 0,001 -0,04 0,190 Years of study (max. 18) 0,14 0,08 0,20 0,05 Poland Spain Slovakia France Beta p Beta p Beta p Beta p Years of study (max. 18) 0,28 0,000 0,22 0,000 0,16 0,000 0,27 0,000 Age -0,16 0,000 -0,18 0,000 -0,04 0,236 -0,14 0,000 Sex 0,11 0,000 0,17 0,000 0,08 0,014 0,10 0,000 does respondent have professional work (full- time, part-time) 0,07 0,032 0,12 0,000 0,07 0,071 0,21 0,000 religiousness -0,06 0,053 -0,07 0,003 -0,03 0,369 -0,18 0,000 Adjusted r Square 0,17 0,21 0,05 0,30 Note: Austria was excluded from the Analysis Due to lack of Data Ireland and Spain exhibit a completely different structure of influence. In Ireland, age is the biggest influence. The older the respondent, the lower the acceptance of the modern family type. Importantly, in Ireland education proves to have little influence on family attitudes. In Spain, education has the biggest influence, but is followed closely by age. In both countries, sex plays a fairly important role in the acceptance of the modern family model. 166 Filip raciborski Thanks to the analysis conducted for these three countries, we know that there is no one pattern of influence. In some countries, the attitudes towards women’s roles in the family depend more on age, while in others, education or other factors are more important. The presented analyses are to show the diversity and complexity of the phenomenon. In the contemporary world, the traditional family model is gradually being replaced by a more modern one. The pace of this change is not the same, though. Some of the countries have achieved a relatively high level of acceptance of the modern family type and the rate of change has already slowed down ( e.g. Great Britain). Other countries have recently made enormous leaps (like Ireland), joining the leaders in this field. We also observe countries that, despite a low acceptance level (the Czech Republic), have not undergone visible changes. It seems that societal changes – increasing level of education, rural-urban migration, technological progress, etc. – result in a greater acceptance of the new roles for women in the family. Nevertheless, some obstacles in the form of cultural barriers exist, which will surely slow down this process. In the present article, I assumed that changes would be found in the increase of equality and women’s engagement in professional work. Nonetheless, the changes in some countries are likely to go into the opposite direction. A surplus on the labour market, connected with the development of technology, has long been recognized [Rifkin, 1995]. Thus, it may be that instead of expanding the labour market, we should do the reverse and limit it. The questions of who will survive on this market and whether sex will play a role still remain. In the coming years we are bound to find out the answer to this exceptionally intriguing question. Appendix Figure A1 Household Chores. Declarations Who Does a Certain Chore: Looking after Sick Family Members Great Britain Ireland Poland France Austria Czech Republic Spain Hungary Slovakia 100% Usually or always man About equal or both together Done by a third person Usually or always woman Female Male 168 Filip raciborski Figure A2 Household Chores. Declarations Who Does a Certain Chore: Household Cleaning Great Britain Ireland Poland France Austria Czech Republic Spain Hungary Slovakia 100% Usually or always man About equal or both together Done by a third person Usually or always woman Female Male Women’s role in the Family: A Comparative Study 169 Figure A3 Household Chores. Declarations Who Does a Certain Chore: Preparing Meals Great Britain Ireland Poland France Austria Czech Republic Spain Hungary Slovakia 100% Usually or always man About equal or both together Done by a third person Usually or always woman Female Male Children – To Have or Not to Have. A Question in the Context of Cross– 7 National Differences Marcin W. Zieliński 7.1 Introduction For a couple of decades, a decreasing fertility rate has been observed in Western European countries. After the transition from a socialistic to a free-market economy, new EU-members have also become affected by decreasing birth rates, as the reproduction rates have also gone down in these countries. As shown in Figure 7.01, the reproduction rates in all selected countries are below the safety level, which is usually assumed to be 1.2, but also even below the simple replacement level (which is usually assumed to be equal 1). Figure 7.01 Net Reproduction Rates in Selected European Countries Norway Sweden Poland Germany Slovakia Czech Republic Slovenia Austria Hungary Source: Council of Europe, 2004] The reasons for population non–replacement trends are as follows: decreasing birth rates and late maternity decisions and increasing life expectancy rates, partially caused by the improvement of life conditions and the advances made in health care [European Commission 2005; Grantl, et al. 2004]. Some authors studying cross-cultural differences, for example, Hirschman [1994], also appeal to cultural differences in values. 174 Marcin W. Zieliński The impact of simple, generational non-replacement is relatively easy to define and explain. Every society, regardless of its structural complexity, needs new members its for existence and future security. According to specialists, decreasing birth rates and the resulting change in demographics will affect the future of these societies. One frequently discussed problem concerns the social support system, especially in the form of the social security system. The increasing number of elderly and the decreasing number of younger, employed people affects both salaries as well as old-age pensions. An interesting study on the economic effects of aging1 was presented by Nishiyama [2002]. There are two main ways of addressing the problem of successive generational non-replacement trends. The first is to incentivize people to have children, such as providing tax reductions, paying for people to have babies or supporting families with children in other ways. The second is to stimulate immigration [Grant, et al. 2004]. For the uniting of Europe, the problem is important in at least two dimensions. First, some population problems in individual countries may become EU problems but on a larger scale. Second, each integrating EU country must decide what kind of model to choose for securing the replacement of its own population. Supporting families and expanding the social support system can cause more economic problems in the future, especially if these policies fail. An increasing number of immigrants creates new challenges for most countries and can cause new problems with immigration issues in the context of generational replacement [ cf. Espenshade, 1987]. Most studies concentrate on the causes and effects of decreasing birth rates as demographic and economic phenomena. Though economic and demographic changes are unquestionable, the decision to have or not to have a baby can be also viewed as a sociological phenomenon [Hirschman, 1994]. This article will investigate the influence of certain economic and social factors on values concerning children rather than the decision to have a baby, treating values as a component or mediator in such a decision. 1 The context of fertility was also considered. 7.2 Is Having Children (Still?) the Most Important Concern over Time? The question of the importance of having children seems to be trivial. It is especially so from the population’s point of view since it is essential for the existence of society. An individual’s importance is not so obvious. The importance of something, however, generally cannot be resolved neatly into a simple positive or negative statement and should rather be recognized as existing somewhere along a continuum. It is not enough simply to say that something is or is not important. It is more interesting to investigate the extent of importance and the factors that determine the level of consensus on the issue of importance. The data used in this article is derived from the International Social Survey Programme (ISSP),2 which is an annual survey covering a variety of topics that rotate each year. Some of them, like Family and Changing Gender Roles III, are repeated periodically. The most recent ISSP concerning family matters was conducted in 2002. Eight countries: Germany (D), Austria (A), Hungary (H), Norway (N), Sweden (S), Czech Republic (CZ), Slovenia (SLO) and Poland (PL) were chosen for this analysis. The German data only comes from the West German population. Slovakian data is available only for 1994, and therefore a comparative analysis could not be undertaken. Two statements concerning the perceived position and importance of children in the family have been examined. If one defines importance along a continuum rather than merely according to extremes, not only a simple agreement with a statement should be measured but instead an entire spectrum of opinions. Being interested in measuring the importance of such a sensitive subject like children, it is easier to use an extreme located at the point where one would be more interested. In this case, we assume that a positive value is related to them. As mentioned above, two questions concerning children are analyzed: A: Watching children grow up is life’s greatest joy. B: People who have never had children lead empty lives. Both of the statements were measured on 5-point scales ranging from Strongly agree (1), Agree (2), Neither agree nor disagree (3), Disagree (4), to Strongly disagree (5) and Can’t choose (8). 2 The data utilized in this article were documented and made available by the Zentralarchiv für Empirische Sozialforschung, Köln. The data for the ’ISSP’ were collected by independent institutions in each country (see: principal investigators in the study-description-schemes for each participating country). Neither the original data collectors nor the Zentralarchiv bear any responsibility for the analyses or conclusions presented here. 176 Marcin W. Zieliński Figure 7.02 Differences in Means by Year and Country: Watching Children Grow up is Life’s Greatest Joy (A) Norway Sweden Poland Germany Czech Republic Slovenia Austria Hungary Comparing answers across years, for the first extremum – an agreement with a positive statement (A) – a declining trend is only noticed in Hungary and Austria. In the case of the negative statement (B), acceptance increased from 1994 only in the Czech Republic. When comparing Figures 23 and 3, one can conclude that the Austrians and Hungarians tend to change their views to the less extreme position, whereas the Czechs tend to stratify and take more extreme positions in comparison to 8 years ago. Other societies tend to express more consistent views – higher acceptance of the positive value and less support of the negative. As shown in Figures 2 and 3,4 the changes between 1994 and 2002 are significant in almost every selected country. In comparing trends, three different patterns can be identified. The first one can be called pro-children orientation; here the perception that joy increases with having children accompanies an increase in the perception that life without children is senseless. The second is termed free-choice orientation, where the perception that watching children grow up is the greatest joy increases while the agreement with the statement “people 3 All calculations presented in this paper were done using R (R Development Core Team 2006) 4 white indicates that the observed difference from 1994 to 2002 was insignificant at the level p<=0.05 (computed using t-test for independent samples) Children – To Have or Not to Have 177 who never had children lead empty lives” decreases. Finally, in the third trend, children start to become an independent factor and are not necessarily perceived as one of the most important values. We call this the independent orientation, and in it we find a decreasing agreement with the statement that children bring the greatest joy coupled with a decreasing belief that childless people lead empty lives. Figure 7.03 Differences in Means by Year and Country: People Who Have Never Had Children Lead Empty Lives (B) Norway Sweden Poland Germany Czech Republic Slovenia Austria Hungary Only one of the selected countries, the Czech Republic, was found to conform to a pro-children orientation; Poland and Norway can be categorized as having a free-choice orientation, while Austria and Hungary exhibit an independent orientation. The other countries do not fit into these classifications because the changes in the tested values were not significant. 7.3 Measuring the Extremes of Extremes … Similar results can be obtained by analyzing the extreme opinions for each of the statements that measure extreme views.5 As shown in Tables 1 and 5 that is by looking at the answers of “Strongly agree”. 178 Marcin W. Zieliński 2,6 the Austrians and Hungarians express less extreme opinions concerning the positive and negative values of both statements. In contrast, the Czechs and Slovenians express more extreme opinions. The Swedes express greater agreement with positive values of having children and approved less radically answers linked to societal expectations. The differences in extreme opinions in other countries are insignificant. TABle 7.01 Percent of Strongly Agree (1) to a Positive Sentence Watching Children Growing up is Life’s Greatest Joy (A) 1994 2002 Country N % N % Diff. Sig. D 862 40.1 368 42.3 +2.2 - A 672 71.8 1009 54.7 -17.1 ** Hu 1190 80.0 745 73.6 -6.4 ** N 647 33.1 577 41.0 +7.9 ** S 615 51.7 563 56.1 +4.4 * CZ 544 54.7 736 60.5 +5.8 ** Si 447 45.0 562 51.9 +6.9 ** Pl 535 35.1 500 41.9 +6.8 ** TABle 7.02 Percent of Strongly Agree (1) to a Negative Sentence People Who Have Never Had Children Lead Empty Lives (B) 1994 2002 Country N % N % Diff. Sig. D 361 18.0 109 13.4 -4.6 ** A 267 29.0 349 20.8 -8.2 ** Hu 874 59.5 474 47.4 -12.1 ** N 71 3.8 47 3.5 -0.3 - S 66 5.9 29 3.0 -2.9 ** CZ 208 20.9 309 25.1 +4.2 * Si 207 21.7 247 23.3 +1.6 - Pl 135 9.7 102 9.4 -0.3 - 6 where ** means p<=0.01 and * means p<=0.05 Children – To Have or Not to Have 179 There is a very interesting finding concerning the comparison of differences measured in this way. The changes that affected the countries tend to exhibit extreme positive values for having children contrary to the acceptance of extreme negative views of societal expectations, which remained the same or became less extreme in most countries, except for the Czech Republic. 7.4 Combining the Measurements and their Social Context The two attitudes (A & B) were combined using the principal component analysis (PCA) to form a construct of the importance of having children. Higher values indicate agreement with the statements that watching children grow up is a greatest joy and that people who have never had children lead empty lives, while lower values correlate with the disagreement to these ideas.7 Attitudes and values concerning children depend on social context, which should help explain the sources of variation and the modeling differences. If so, the effects of social contexts should be modeled contextually. The social context will be defined as factors that directly or indirectly influence measured attitudes and that depend on the individual only to a minor extent. Specifically, dependence is defined by being attributable to the individual. As mentioned above, the main axis of division among countries is the age of the democracy. Old democracies can be divided into the Scandinavian countries (Norway and Sweden) and the non-Scandinavian countries (Germany and Austria). Hungary, the Czech Republic, Poland and Slovenia can be classified as “new democracies”. To begin, a simple model indicating the relationship of selected social– demographic characteristics to the construct was created; this is presented in Figure 7.04.8 The shape indicates the strength and direction of the relationship. Additional y, violet indicates a positive relationship and blue a negative one, regardless of its significance, which is presented in Table 7.06.9 For older or less educated people who have more children, children are more important than for the younger or more educated people who have fewer or no children; these are the strongest measured relationships. Also for married people in every country children are more important than for those who are not married. This is the second strongest and non-debatable observation. 7 To make interpretation easier, scales were first reversed before combining. 8 The relevant values were rounded to the first digit and presented in Figure 7.03. 9 Individual characteristics selected for analysis are presented in detail in Appendix A. 180 Marcin W. Zieliński TABle 7.03 Values of Strength and Direction of Relationship of Selected Characteristics and the Construct Rounded to First Two Digits (Pearsons) Sig. 0.38 0.00 0.00 0.98 0.00 0.08 0.01 0.49 0.00 0.12 0.01 0.00 0.50 0.00 0.23 PL r -0.03 0.18 0.13 0.00 -0.17 -0.05 -0.08 -0.02 0.26 0.05 -0.08 0.09 -0.02 0.10 -0.04 Sig. 0.09 0.00 0.00 0.40 0.00 0.18 0.01 0.24 0.00 0.46 0.00 0.00 0.59 0.07 0.02 SI 1 r -0.05 0.22 0.16 0.04 -0.23 -0.04 -0.08 -0.05 0.20 -0.02 -0.14 0.1 -0.02 0.06 -0.07 Sig. 0.00 0.00 0.00 0.34 0.00 0.50 0.08 0.02 0.00 0.16 0.03 0.03 0.29 0.70 0.08 CZ r -0.14 0.16 0.15 0.04 -0.12 -0.02 -0.05 -0.08 0.21 0.04 -0.06 0.06 0.03 -0.01 -0.05 1 Sig. 0.38 0.00 0.00 0.17 0.00 0.05 0.42 0.00 NA 0.05 0.1 0.50 0.14 0.35 0.19 S r 0.03 0.15 0.17 0.07 -0.20 -0.07 -0.03 -0.12 NA 0.07 -0.05 -0.02 0.05 -0.03 -0.04 Sig. 0.18 0.00 0.00 0.59 0.00 0.00 0.00 0.00 NA 0.83 0.00 0.39 0.23 0.33 0.00 N 1 1 1 1 r 0.04 0.1 0.12 0.02 -0.22 -0.10 -0.1 -0.15 NA -0.01 -0.1 0.02 0.03 -0.03 -0.1 Sig. 0.00 0.00 0.00 0.90 0.00 0.09 0.00 0.00 0.00 0.12 0.26 0.88 0.75 0.30 0.01 HU r -0.17 0.15 0.14 0.01 -0.17 -0.05 -0.09 -0.13 0.26 0.05 -0.04 0.00 -0.01 -0.03 -0.08 Sig. 0.17 0.00 0.00 0.10 NA 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.52 0.00 0.00 A r NA -0.03 0.17 0.16 0.06 -0.23 -0.14 -0.24 0.37 0.14 -0.20 0.14 -0.02 0.19 -0.07 Sig. 0.00 0.00 0.00 0.01 0.00 NA 0.00 0.00 NA 0.34 0.01 0.28 0.96 0.00 0.46 D 1 r -0.1 0.18 0.13 NA NA -0.16 -0.24 -0.20 -0.18 0.03 -0.10 0.04 -0.00 0.13 -0.03 .cap. r.sex r.age r.marital r.partner r.education r.urban.rural r.working inc.per eur chil.tot.numb chil.pre-sence. in.hh r.non-believer r.catholic r.protestant r.church.attend r.top-bottom Children – To Have or Not to Have 181 Respondents’ gender differentiates attitudes towards children to a great extent in Hungary, where – similar to other countries – being female increases the importance of children. Interestingly, having a partner, regardless of it being a wife/husband or someone else, increases the perception that children are important in Austria, while it decreases this belief in Germany. Thus, Austrians with partners attribute more importance to children, while Germans with partners attribute less importance to them. The difference between these two countries may be explained through the context of religion, as Austria is homogenously Catholic. This aspect needs to be examined in further analysis. Religion-oriented attitudes and behaviors change the perception in different ways. Being a non-believer or attending church less frequently are negatively related to the attitude towards children, while being a Catholic encourages a positive attitude, in contrast to being a Protestant, which seems to be unrelated. Figure 7.04 Strength and Direction of the Relationship of Selected Characteristics and the Construct of Children Importance        General similarities and dissimilarities between countries are shown in Figure 7.05. Correlation coefficients forming a straight line indicate that there is no difference between the two countries. Of course, in reality the probability that such situation would occur is very low. One can assume that such a line in the case of two countries shows more 182 Marcin W. Zieliński the relationship between socioeconomic indicators and the construct of the importance of children. At first glance, the strongest similarities between countries are observed in Norway and Sweden, followed by Hungary and Czech Republic, where the hypothetical line seems to be closest to straight. The first pair comprises Scandinavian countries, and the second pair was part of the former Austro-Hungarian Empire. Also Poland and Slovenia appear quite similar. Other similarities and dissimilarities are less obvious. Figure 7.05 Correlation Similarity – Dissimilarity Matrix Across Countries As shown in Table 7.04,10 the Germans are most similar to the Poles, the Austrians to the Germans and the Poles, the Hungarians to the Czechs, the Norwegians to the Swedes, the Czechs to the Hungarians, the Slovenians to the Poles and the Poles to the Slovenians. One can conclude that, with 10 Table 7.04 presents the distances computed using the Euclidean method. Lower values indicate a greater similarity and shorter distance between two countries in contrast to higher values, which indicate a greater distance. For more detailed information about measuring distances, see Lucas, 2006 Children – To Have or Not to Have 183 some exceptions, there are at least three categories of countries: Scandinavian countries (Norway and Sweden), Eastern European countries (Hungary the, Czech Republic, Slovenia and Poland) and two other countries – Germany and Austria, which are similar only to some extent. Similarities in the patterns of distances are shown in Table 7.05. This measure was adopted from a response-item theory originally designated for the informal testing of pairwise associations between polytomous items.11 Values above the diagonal present country-by-country correlations of the computed distances from Table 7.04. Higher values indicate a greater similarity between two selected countries in patterns of distances to other countries; lower values indicate less similarity. P-values (below the diagonal) indicate the significance of pattern similarities. TABle 7.04 Table of Distances Between Countries D A HU N S CZ SI PL D - 0.36 0.32 0.35 0.41 0.36 0.33 0.30 A 0.36 - 0.45 0.39 0.44 0.46 0.42 0.38 Hu 0.32 0.45 - 0.26 0.25 0.14 0.27 0.25 N 0.35 0.39 0.26 - 0.18 0.27 0.24 0.26 S 0.41 0.44 0.25 0.18 - 0.23 0.27 0.25 CZ 0.36 0.46 0.14 0.27 0.23 - 0.22 0.21 SlO 0.33 0.42 0.27 0.24 0.27 0.22 - 0.16 Pl 0.30 0.38 0.25 0.26 0.25 0.21 0.16 - Significant similarities in distances can be found between Slovenia and Poland, Hungary and the Czech Republic, the Czech Republic and Austria and Hungary and Austria. These results support the hypothesis that the countries of the former Austro-Hungarian Empire have strong relationships with each other. Comparing Tables 4 and 5, one could come to the conclusion of inconsistency regarding Austria because it appears to be the country most distant from the Czech Republic and Hungary while simultaneously having the same pattern of similarities. Nevertheless, the inconsistency is more ostensible than real. The pattern of similarities is significantly the same though going in the opposite direction; this means that countries more similar to Austria in their perception of importance of children are at the same time less similar to the Czech Republic and Hungary. 11 More detailed information can be found in Rizopoulos, 2006. 184 Marcin W. Zieliński TABle 7.05 Correlations Among Patterns of Distances Between Countries D A HU N S CZ SI PL D - 0.00 -0.16 -0.37 -0.56 -0.32 -0.22 -0.14 A 1.00 - -0.76 -0.51 -0.65 -0.77 -0.59 -0.52 Hu 0.70 0.03 - 0.25 0.42 0.83 0.29 0.30 N 0.37 0.20 0.55 - 0.70 0.24 0.32 0.20 S 0.15 0.08 0.30 0.05 - 0.48 0.31 0.29 CZ 0.44 0.02 0.01 0.56 0.22 - 0.50 0.50 SlO 0.60 0.12 0.48 0.45 0.45 0.21 - 0.73 Pl 0.75 0.19 0.47 0.64 0.49 0.21 0.04 - 7.5 Methods of Predicting the Construct As was mentioned in Section 7.4, data that is collected contextually should also be analyzed contextually. Observed relationships between the construct and selected characteristics indicate that in some countries they vary non-randomly and sometimes in opposite directions. Some countries are more similar than others. There are several different methods used in analyzing cross-sectional data. Mostly meta-analysis is used to determine the differences in predictions in cross-sectional data. This is done when the original data are unavailable [Hox, 2002]. It is far more convenient when the original data is available, as is the case here. To model the data contextually, one needs to take into consideration that data from each country were collected separately and independently. Thus, even if the data set combines data from every country, we still have data from independent samples. Predicting independent variables12 is usually done by using a multiple regression analysis, which allows the measuring of “pure” effects. There are two ways of addressing independent factors. One can use a multiple regression analysis independently, doing it separately for each country and then comparing the results using meta-analysis, or one can use a multilevel hierarchical regression; this handles the problems very well. The multilevel regression assumes the existence of at least two levels of analysis. In the analysis of cross-sectional data (including cross-national), especially when the 12 in this case a construct of importance original data is available, Level 1 is created through the units of observations (respondents) and Level 2 is formed by the countries in which the data was collected. The access to the raw data allows for the use of a standard multilevel regression [de Leeuw and Hox, 2003]. There are also other methods dealing with the problem of independence that are usually found in meta-analytical approaches. For example, Apparala, Reifman, and Munsch [2003] use ordinary regressions and correlations for measuring Level 2 variability. The first step is the regression analysis for each individual country using predictors collected at Level 1. Second, the relationships between the computed intercepts and the predictors collected at Level 2 are investigated, using simple correlations. More sophisticated methods are needed if there is no access to the raw data. The study of Lorant et al. [2003] can be used as an example in which relations between different studies from different countries were examined. More examples can be found in Lumley [2006] or Schwarzer [2005]. The main advantage of using a multilevel regression is the possibility of decomposing the variance between the two levels of observation and of including (as predictors) data collected at these different levels – in this case at the respondent (Level 1) and country (Level 2) levels. Using the traditional methods of aggregating or disaggregating data is insufficient and can lead to serious problems in estimating, thus manifesting biased results, e.g. see: P. D. Bliese and Hanges [2004]. 7.6 Modeling Differences Between Countries A multilevel regression was used for modeling the data.13 The dependent variable was an index of the importance of children constructed from the two statements by multiplying them together. To simplify interpretation, the computed index was then multiplied by -1, so that higher values of the index indicate more importance and lower values designate less importance. First, one has to decompose the variance that exists at the two levels of analysis, that is between-group (τ m 0) and within-group (σ2 m 0) variability. Estimated values are presented in Table 7.06. As suggested by Hox [2002] or P. Bliese [2006a], the Intraclass Correlation Coefficient (ICC14) was computed using Formula 1 to determine the amount of variability in the perception of the importance of children attributable to the national level. Its value (ICC=0.10) indicates that 10% of the total variance 13 Pinheiro, Bates, DebRoy, and Sarkar [2006] 14 More detailed information can be found in P. Bliese [2006b]. 186 Marcin W. Zieliński is attributable to the variability between countries, which is significant at p<=0.001.15 It means that there is a significant variation between countries in the perception of importance of children. As shown in Table 7.06, σ2 within-country variance is equal to 11.229 and group means (countries) are reliable at 0.993.16 As noticed in P. Bliese [2006a], ICC2 also takes into account group sizes, which, in the case of this study, differ greatly (from 936 cases in West Germany to 1475 in Norway). TABle 7.06 Between-Group and Within-Goup Variance Estimates Variance StdDev τ m 0 1.292 1.137 σ2 m 0 11.229 3.351 As shown in Table 7.03, most of the selected socioeconomic characteristics are strongly related to the construct of importance. Three groups can be created from them. The first one describes certain demographic information of the respondent (such as: sex, age, living in the rural or urban area, marital status, working status and income17); the second describes the respondent’s experience with children (such as: total number of children and presence of children in the household); and the third group describes religious tendencies (being a nonbeliever, Catholic or Protestant, as well as church attendance). τ m 0 ICC = ––––––– (1) τ m 0 + σ2 m 0 Only some of the variables were selected for the further analysis. The decision to keep or drop certain variables was based on the strength of the previously measured relationship and the availability of the data in all countries. It was also decided that only one of the religious items would be chosen because of the high religious homogeneity in most of the countries. Austria was excluded from the analysis because of the decision to keep education as a predictor. There are two measurements of the respondent’s level of 15 To calculate significance, ANOVA test was performed 16 This measurement is also known as ICC2, see: P. Bliese [2006b] 17 Income per capita in euros. For Norway, Sweden, Hungary, Slovenia, Czech Republic and Poland, the income per capita was recalculated using the course published on July 1, 2002 by the European Central Bank http://www.ecb.int/. The top 5 % were excluded individual y for each country to remove potential outliers. Norwegian data concern the income before taxes contrary to other countries where the data concern income after taxes. Children – To Have or Not to Have 187 education in ISSP data. The first one is the highest level of education obtained and the second is the amount of education received as measured in years; these two measurements are not equivalent. The latter is not available for Austria. Education, as shown in Table 7.03, seems to be a very strong predictor of the subjective perception of the importance of children, so the decision to exclude Austria rather than education, even if controversial, was made. As suggested by Hox [2002], the model without any predictors was computed first, i.e. the intercept-only model, to estimate the amount of between-group and within-group variance existing in the data. Second, further variables were included in the model, describing data collected at Level 1 (the respondent level). The decision about the order of inclusion was based on three theoretical dimensions: socioeconomic factors, religion and experience with children. Results are presented in Table 7.07. The analysis of the values of coefficients makes it clear that even after controlling for other demographic factors almost every selected feature significantly explains the perceived importance of children in the family. Being female, older, married, less educated and coming from a lower income household is conducive to attaching more importance to children. Next, the information collected at Level 2 was introduced to the model. Three parameters were chosen: life expectancy at birth as an indicator of well–being of the society, unemployment rate, and fertility rate as reported by the Central Intelligence Agency (CIA), 2006. The decision to include the latter parameter was based on the observation that the fewer children a society has, the more important they become to that society, and conversely – the more children, the less important they are. All three country-level indicators significantly differentiate the perception of the importance of children. Lower life expectancy at birth, lower fertility rate and lower unemployment rate foster the perception that children are more important. Separate regression lines, obtained by using an ordinary regression analysis with predictors for both levels, are presented in Figure 7.06. Table 7.07 also contains three measures of fit conventionally used in the comparisons of significant differences across models (AIC,18 BIC19 and logLik20). The smaller the value of AIC or BIC, the better the fit to the data. Conversely, the smaller the value of logLik, the worse the fit. All models were sequentially tested according to the improvement of fit to the data. The inclusion of every tested variable significantly improved prediction. 18 Akaike’s Information Criterion 19 Bayesian Information Criterion 20 A value of the log-likelihood function. 188 Marcin W. Zieliński TABle 7.07 Results Multilevel Analysis of Children Importance p 0.00 0.02 0.00 0.00 0.00 0.08 0.00 0.00 0.00 0.01 0.00 0.00 .eS. 1.76 0.09 0.00 0.09 0.02 0.10 0.00 0.12 0.09 0.03 0.44 0.01 0.00 3.16 Model 12 al.V 0.00 0.51 0.00 9.96 16.28 -0.21 0.02 0.68 -0.16 -0.18 -0.46 -0.16 -3.99 -0.19 28457 28550 −14215 p 0.00 0.01 0.00 0.00 0.00 0.09 0.00 0.00 0.00 .eS. 0.44 0.09 0.00 0.09 0.02 0.10 0.00 0.13 0.09 0.92 3.16 Model 8 al.V 0.85 9.96 -3.80 -0.22 0.02 0.68 -0.16 -0.17 0.00 -0.42 0.51 28483 28556 −14230 p 0.00 0.01 0.00 0.00 0.00 0.17 0.00 0.00 . eS. 0.42 0.09 0.00 0.09 0.02 0.10 0.00 0.13 0.87 3.17 Model 7 al.V 0.76 -3.37 -0.24 0.01 0.81 -0.16 -0.14 0.00 -0.43 10.02 28512 28578 −14246 p 0.00 0.00 0.00 0.00 0.00 0.17 0.00 . eS. 0.41 0.09 0.00 0.09 0.01 0.10 0.00 0.84 3.16 Model 6 al.V 0.71 9.99 -3.48 -0.26 0.01 0.81 -0.16 -0.14 0.00 28849 28908 −14415 p 0.00 .eS. 0.43 1.14 3.35 Model 0 al.V 1.29 -4.80 11.23 38474 38495 −19234 .cap.eur .rate lik ixed Part mx iC iC F Level 1 (intercept) r.sex r.age r.marital r.education r.working inc.per r.non-believer chil.presence.in.hh Level 2 life.expectancy fertility unemp.% Random Part τ mx 2 σ Model Fit A B log Children – To Have or Not to Have 189 The most frequently used measure for the explained variance in regression models is R 2. It is not recommendable, however, in cases of multilevel models. The reasons are quite well described in literature [see e.g. Hox, 2002; or Bliese, 2006a] and are based on at least three main premises. First of all, there are different levels of analysis and different groups in which the explained variance may differ. Second, there are several measures known as R 2, calculated using different formulas. Finally, there is a dependency on model specification. There is a large family of hierarchical models, and the specification of each of them depends on the research questions asked in the modeling processes (in this analysis one of the simplest was applied). Obtained R 2 measures can differ and are incomparable. Figure 7.06 Separate Regressions for Each Country for the Model 12 Norway Sweden Poland Germany Czech Republic Slovenia Hungary Nevertheless, one of the measures of R 2 in multilevel models was computed. It was proposed by Bryk and Raudenbush [1992] and an implementation example is shown by Hox [2002]. It is based on a simple comparison of the 190 Marcin W. Zieliński variance at each level with a variance obtained at those levels in a random-intercept (that is model zero) model, using Formula 2 for Level 1. σ2 m x represents the within-group variance in model x and σ2 m 0 represents the variance in the intercept-only model. Formula 3 is used for computing R 2 expressing the part of variance explained at Level 2. t mx represents the between-group variance in model x and t m 0 represents the variance in the intercept-only model. Applying Formulas 2 and 3, the final demographic model ( Model 8) accounts for 11% of the variance among respondents. As shown in Table 7.07, adding variables that describe religiosity and experience with children does not change very much in R 2, which after rounding to the first two digits still remains 0.11. Adding variables measured at Level 2, such as life expectancy, fertility rate and unemployment rate, accounts for close to 100% of the variability between countries. One should be careful in interpreting the results in this way. The total number of countries in this study was 7, which is a very small sample. To obtain more accurate and reliable results, the sample at Level 2 should be much larger. As noticed by some authors,21 with such small samples, the results can be unrealistic, especially for explaining the variability at Level 2. 7.7 Results and Discussion The analysis was conducted in three different stages. First, the changes σ2 m 0 – σ2 m x R 2 L = ––––––––– (2) 1 σ2 m 0 over time in the levels of agreement with the two statements that form the construct were measured using ISSP data. Three different patterns were observed. People from the Czech Republic changed their views towards a more pro-children orientation, in Poland and Norway, opinions tended to τ m 0 – τ m x R2 L = ––––––––– (3) 2 τ m 0 21 See, for example, Apparala, Reifman, and Munsch [2003], Hox [2002]. Children – To Have or Not to Have 191 be more independent of societal expectations, while the Austrians and Hungarians expressed less extreme views. The second stage was the measuring of relationships between selected socioeconomic factors and the construct. Most of the selected characteristics turned out to be related to the construct. Countries differed mostly in the strength of the relationship but not in the direction. The final stage of the analysis was the application of a multilevel regression to investigate the “controlled” relationship to the construct and to explain the differences between the countries. Variables used to explain differences between countries, such as life expectancy at birth, fertility rate and unemployment rate were useful, accounting for almost all variance on the national level. Nevertheless, the small number of countries does not allow for a general reflection concerning the role of such country-level characteristics as life expectancy rate, fertility rate or unemployment rate. Results suggest that there is a relationship between these factors and the perceived importance of children, but this problem needs to be investigated on a much larger scale. Unfortunately, it was impossible to include information about the number of children of each respondent. The relationship between this variable and the construct was one of the strongest, but data was unavailable for three countries. Due to lack of information for the total number of years of education in Austria, that country had to be excluded from analysis. This was a shame, especially considering that some relationships between countries of the former Austro–Hungarian Empire were noticed. Only differences in means were examined. However, as shown in Figure 7.06, large differences between slopes were noticed. Further analysis, in addition to increasing the number of countries examined, should also investigate these differences. 7.8 Variables Used in Analysis r.sex: Respondent’s gender (0 = female,1 = male) r.age: Respondent’s age r.marital: Respondent’s marital status (0 = unmarried, 1 = married) r.partner: Respondent in steady, long-term relationship (0 = no, 1 = yes) r.education: Respondent’s education (in years) r.urban.rural: Respondent’s place of living (0 = rural, 1 = urban) r.working: Respondent’s employment status (0 = not-working, 1 = working) inc.per.cap: Respondent’s income per capita in euros chil.tot.number: Respondent’s total number of children 192 Marcin W. Zieliński chil.presence.in.hh: Respondent living with children in the same household (0 = no, 1 = yes) r.non-believer: Respondent believes in God (0 = no, 1 = yes) r.cathol: Respondent catholic (0 = no, 1 = yes) r.protest: Respondent protestant (0 = no, 1 = yes) r.church.attend: Respondent’s church attendance (1 = at least once a month, 0 = less frequently) r.top-bottom: Respondent’s Top–Bottom self-placement (1 = lowest, 10 = Highest) Section III Citizenship, Institutions, Political Participation 8 An Outline for a New Concept of Citizenship in a Trans-National Context Mitja Hafner-Fink | Samo Uhan 8.1 Introduction In a recent article, Cris Shore [Shore, 2004] poses some questions that are relevant for social scientists: (1) is it possible to create a citizenship meaningful beyond the level of the nation-state and, if so, (2) what does it mean to be a global citizen; further (3) why is it relevant to study global citizenship at all? It is not easy to find answers to the first two questions. With regard to the third it is obvious that there are several reasons, three of which will be outlined here. First, while modern citizenship is theoretically and empirically connected to the nation-state, one of the key theses of globalization is that economic and political power transcends the borders of the nation-state. Thus, globalization challenges one of the key assumptions of citizenship, namely, that society is equal to the territory of the nation-state. In this way globalization constitutes a normative democratic problem on the one hand, and an inadequate account of empirical political modes of participation on the other. Another reason, more specifically associated with adopting a bottom-up perspective on global citizenship, is that institutionalized citizens’ rights, duties etc. are not necessarily a good indicator of the actual existence of rights in practice [Heater, 1990]. When studying global citizenship from a bottom-up perspective it becomes possible to see to what extent people feel, identify and participate as global citizens. The third point refers to a more or less practical dimension; an analysis of global citizenship indicates some practical (political) controversies of global citizenship. In this way it supports the development of tools to manage these controversies. It is suggested that Western Europe provides the most encouraging site for the development of new forms of political community. In particular, the institution of European Citizenship is seen as an important step forward in giving rights to make claims on the state on the basis of an external concept of obligations [Kveinen, 2000]. What it offers is a new paradigm based on the decoupling of rights and identity – the two main elements of traditional citizenship – and on allegiance to civic and political norms rather then ethno-cultural ties. This separation of the legal from the cultural and territorial dimensions of citizenship is seen as crucial for developing civil society beyond the physical boundaries of nation-states [Shore, 2004]. However, this (European) form of supranational citizenship raises some other questions of a sociological kind: what is the character of this supranational citizenry and where (else) can the elements of a new, conditionally described, ‘postmodern’ citizenship be found. 198 Mitja Hafner-Fink | Samo uhan With the concept of ‘post-modern’ citizenship we refer to a group of phenomena that different authors qualify using different labels such as: transnational citizenship [ e.g. Castles, 2005], multi-level citizenship [ e.g. Painter, 2002], and global citizenship [ e.g. Falk, 1994; Muetzefeldt and Smith, 2002]. A common point of all these understandings is that they elaborate the concept of citizenship in the context of the globalization process. They extend and refine the traditional concept of citizenship in different directions, yet primarily the following common elements appear: – citizenship is no longer predominantly determined by nation-state and territory (encompassing concepts such as global governance, global civil society, trans-national democracy etc.); – multiple identities and individualism become prevalent; and – new forms of (political) participation (as a politico-cultural aspect of citizenship) occur. The objective of our paper is to identify empirical evidence that supports the idea of the existence of this (so-called) ‘post-modern’ citizenship. Following this aim, the paper is divided into three parts. In the first part, the traditional concept of citizenship is briefly presented through a discussion of the ‘conflict’ between the liberal and republican models of citizenship. We indicate a possibility for resolving this ‘conflict’ by introducing a new concept of ‘postmodern’ citizenship. In the second part, we present the results of analyses of survey data from the International Social Survey Programme (ISSP) module 2004 ‘Citizenship’ [see: ZA, 2006] and from the European Social Survey – Round 1 [ESS, 2002] [see Jowell, et al. 2003]. In the third part, we return to theoretical grounds: here, a possible direction for modeling a new (‘postmodern’) paradigm of citizenship is proposed. It is expected that within this framework a possible correspondence between theoretical expectations and empirical evidence might be found. 8.2 Classical Concept of Citizenship Starting with Turner’s [Turner, 2000] definition we can describe a classical concept of citizenship as an ensemble of rights and obligations that determines an individual’s access to social and economic resources. In historical terms, citizenship creates a juridical identity that determines an individual’s status within the political community. In fact, citizenship is itself one of the most important resources that a society ascribes to an individual as a legal personality. An Outline for a New Concept of Citizenship in a Trans-National Context 199 Citizenship within this framework can be seen as: (a) an inclusionary principle for the distribution and allocation of entitlements; and (b) an exclusionary basis for building solidarity and maintaining identity [Turner, 2000:23]. Its modern meaning is defined as follows by the Encyclopaedia Britannica: ‘ Citizenship is a relationship between an individual and a state in which an individual owes allegiance to that state and is in turn entitled to its protection’. According to Frey [Frey, 2003:95], three aspects of this definition have to be noted: 1. The actors involved are the citizens and the state. Citizenship is a unique and monopolistic relationship between the individuals and a particular nation. 2. The citizens have both rights and obligations. The rights refer to: – the civil sphere, that is, citizens are protected against the state by the rule of law and protected when they are abroad, and they may take up residence within the borders of their state; – the political sphere, that is, citizens have the right to vote and to hold public office; and – the social sphere, that is, citizens are protected against economic hardship within the welfare state. 3. The relationship between an individual and the state goes well beyond an exchange of taxes for public services. Citizens are expected to be public spirited and to exhibit civic virtue. The listed elements comprise the core of the modern concept of citizenship. This is grounded in membership in the community, the invocation of the demos and ensuring solidarity, all of which bear on the processes of social integration and the formation of a collective identity. The equal rights of citizens are based on membership in the political community (they are particular) and only to a lesser extent on the universal legal status of the individual ( e.g. universal human rights). Lying at the heart of the modern (classical republican) concept of citizenship is political participation, which is linked to certain prerequisites or competencies that assure the equality of the chosen while excluding an important part of the population that are not citizens. In this sense, national citizenship is constructed around institutionalized racism because it excludes outsiders from access to entitlements typically on the basis of a racial or national identity [Turner, 2000]. The inherent exclusivity of the republican model of citizenship is precisely its greatest drawback in the view of many scholars. This model can only operate openly, non-exclusively and fairly when the collective identity of the demos is treated as a purely political category rather than an ethnic or cultural one. But this happens only rarely. 200 Mitja Hafner-Fink | Samo uhan Over the past decade variants of the so-called liberal cosmopolitan citizenship model, which highlights the individual’s universal legal status, have appeared as a response to the deficiencies of the traditional republican model. Within this concept the citizen is not a political actor but a legal person [Cohen, 1999:249], who abides by the laws and may expect legal protection from the state. The core of this legal reconceptualization of citizenship is the individual’s universal legal status and individual rights. The legal dimension of citizenship is universalistic and potentially inclusive. It is not tied to any particular collective identity, allegiance or demos nor is it restricted in space, and it can ‘coexist’ with various kinds of status of the individual. The flaw of this model is that the dominance of the legal dimension of citizenship could lead to processes of depolitization and desolidarization. As such, it weakens political participation and social solidarity. In this sense, the legal and democratic dimensions collide. The former is universal and inclusive, as well as apolitical and individualistic, while the latter is internally egalitarian while exclusive and particularistic. The main deficiency of the liberal citizenship model in the view of many authors is its abstract individualism, which does not recognize particularistic identity (political and cultural), contexts and traditions, and further does not satisfactorily apply the key principle it advocates – by replacing political with universal rights based on the most abstract identity, humanity. Cohen [Cohen, 1999] critically appraises both approaches and ‘charges’ them both with analytical vagueness because they both start from the same assumptions, even if they evaluate them differently; the different dimensions of citizenship are combined into a single ensemble of rights protected by the nation-state. In Cohen’s view this approach leads to a number of wrong alternatives that cannot cope with the decomposition of the paradigmatic concept of citizenship that emerged from the democratic revolutions of the 18th century. 8.3 A Post-Modern Puzzle – Supranational or Global Citizenship Although the modern paradigm of citizenship was never normatively satisfactory, it did try to resolve the tensions between democracy, justice and identity, if only it had been institutionalized in the right way. Today, globalization has undermined its core presuppositions. The exclusive territoriality and sovereignty inherent in the nation-state model are being transformed due to the emergence of transnational economic practices, supranational legal regimes and post-national political bodies [Cohen, 1999]. An Outline for a New Concept of Citizenship in a Trans-National Context 201 The model of full sovereignty and self-sufficiency has no empirical approval. Even the model of the demos as a homogeneous collective subject is no longer accepted. On the other hand, we are witnessing the more or less successful modeling of the so-called supranational European citizenship. According to its official narratives, the European Union seeks to develop a stronger sense of European identity and citizenship above the level of the nation-state, while simultaneously contributing to the flowering of local, regional and national cultures and identities below it. We can find a theoretical background for this in Weiler’s concept of EU citizenship. What he proposes is a model of citizenship based not on emotional attachments to territory or any notion of pan-European cultural affinities (such as a shared history or language), but on a commitment to the rights and duties of a civic society covering discrete areas of public life [Shore, 2004]. Weiler’s model has triggered strong reactions. While we have no intention to evaluate EU integration, it is hard to ignore the arguments against supranational citizenship or EU identity. According to Shore [Shore, 2004: 35–36] they are: 1. The EU supranational model of citizenship is too abstract and rationalistic in its inspiration. The model itself rests on a convenient dualism between ‘civic-political’ and ‘ethno-cultural’ ties. The key question here is if this theoretical distinction can be maintained at an empirical level. From the theoretical perspective, the assumption that the baser impulses of nationalism can be sublimated into purer and more rational constitutional patriotism needs to be questioned. 2. The idea that rights and identity can be decoupled without problems to create a new post-national citizenship where law stands above the logic of culture has no empirical evidence. 3. EU citizenship based on rational civic norms and devoid of ‘emotional attachment’ or any sense of ‘shared history’ would be a peculiarly elitist, sterile and soulless form of citizenship. As Painter [Painter, 2002:99] stressed, formal EU citizenship is supplementary to the citizenship of a member state; to be a citizen of the EU one must be a citizen of a member state. 8.4 What Can Survey Data Tell us? Our main (empirical) research question asks whether we can find the elements and factors of a ‘post-modern’ concept of (global) citizenship in the situation where nation-states are still the main source of the dominant model of citizenship. In this situation, we can search for ‘tracks’ of elements of the new concept of citizenship at least on the following two levels (or as a combination of these two levels): (a) on a contextual level which, from the analytical point of view, means different comparisons of nation-states or elements and factors which are observed at the level of the nation-state (including aggregate data from opinion surveys); and (b) on the level of the attributes of individuals ( e.g. attitudes, competence, behavior, socialization, status). Parallel to understanding citizenship as the relationship between an individual and the state, we are observing, separately and in various combinations, dimensions of the model of possible global citizenship by means of indicators on the individual level (citizens) and on the country level (nation-states). We are focusing on the level of individuals and, for this purpose, we use survey data from two international surveys: a) the Citizenship Module of the International Social Survey Programme [ISSP, 2004], which was applied in 39 countries in the period from 2003 to 2005 [see ZA, 2006]; b) the European Social Survey – Round 1 (2002–2003), which was conducted in 22 countries (21 European countries and Israel) (see Norwegian Social Science Data Services).1 Using these data we are able to observe different indicators of several dimensions of global citizenship. We are particularly interested in the following five dimensions: 1. attitudes towards transnational or global civil society and governance; 2. “normative” perceptions about citizens’ duties – what does good citizen look like? 3. tolerance and understanding of minorities’ rights; 4. citizens’ empowerment – their “political competence”; 5. citizens’ involvement and political participation. We follow the idea that we can trace seeds of the concept of ‘post-modern’ citizenship at the level of citizens’ social and political consciousness (attitudes, perceptions, evaluations) and their political behavior. Yet we simultaneously also pay attention to the macro level or to contextual or systemic ele-1 Norwegian Social Science Data Services is the official archive and distributor of the ESS data. An Outline for a New Concept of Citizenship in a Trans-National Context 203 ments of a possible new concept of citizenship. In that respect, we compare and classify countries (nation-states) on the basis of individual (aggregated) characteristics. We expect that differences and similarities between countries regarding the presence (or extent) of different individual dimensions of the post-modern concept of citizenship correspond to differences and similarities on the contextual level. For example, in countries where several systemic elements (or conditions) of a possible new concept of citizenship are present, we can also expect that their citizens express this fact through their (political) attitudes and behavior. a) The Global Situation (Data from the ISSP 2004) Data from the ISSP 2004 (Citizenship module) offer us a possibility to observe the presence of a post-modern concept of citizenship on the global (worldwide) comparative level – countries from all continents are included in the survey. Using these data we are able to observe the following indicators (see the Appendix 1 for details) of the five dimensions of global citizenship: – support for transnational civil society and for global governance (attitudes on the relationship between national governance and the United Nations and preferences for the inclusion of citizens’ organizations in international organizations); – a ‘normative’ perception of a responsible citizen (citizens should show the following civic virtues: supervision of the government performance, political and social participation, political empathy and tolerance, solidarity, ecological consciousness); – understanding of rights as a universal, non-exclusive category (tolerance, respect for minorities’ rights); – subjective competence (self-evaluation of the ability to follow and understand politics); and – the individual as an active citizen (selected indicators of political participation measuring intentional and self-determined activities: communication with the mass media, use of the Internet for political communications, boycotts of products). For developing a typology or classification of countries on the basis of aggregated individual data we use the hierarchical cluster method – the process of the ‘stepwise’ joining of units (countries) on the basis of the similarities (and differences) between them. We compare countries regarding the above mentioned five dimensions of the hypothetical ‘post-modern’ concept of citizenship. We ascribe to countries (as units of classification) aggregated individual data ( e.g. the proportion of respondents who support a certain statement, an average value of answers on a measurement scale). The analysis of 204 Mitja Hafner-Fink | Samo uhan the results (dendogram, structure of clusters) supports the decision to group the countries into six meaningful clusters. Here are the key characteristics of these six ‘typical’ clusters, which can be joined in two big groups of countries (Figure 8.01). (A) The first group of three ‘typical’ clusters connects countries with a clear presence of the dimensions of a post-modern concept of citizenship. Most of these countries are developed democratic countries. Cluster 1 – economically developed countries with a longer tradition of consolidated democracy (Australia, Austria, Canada, Denmark, Finland, the Netherlands, New Zealand, Norway, Sweden, Switzerland). All countries in this cluster are classified as free countries scored with the highest rating according to the classification of Freedom House. In this group of countries we can see the large support of respondents for the role of the United Nations. This is also the group with the highest proportion of active citizens, while the proportion of citizens that sees themselves as competent is also well above the average. It is interesting that in the countries of this group we cannot speak of a high level of support for the normative concept of a responsible citizen. We can also conclude that the countries in this cluster do not confront any serious problems of systemic corruption: here we find the highest proportions of respondents who do not see corruption as a problem in their countries. Cluster 2 – new and old European democracies and Israel (Belgium-Flanders, France, Germany – East and West, Great Britain, Slovenia, Spain, Israel). With the exception of Israel, al countries from this group have also received the highest Freedom House rating. The most visible characteristic of the group is the large support for the role of the United Nations, although it is not as great as in the first cluster. Considering al other dimensions, countries in this group are at the average level. Like in the first group, support for the normative concept of a responsible citizen is also below average in this group. Cluster 3 – the United States and Ireland. Split support for the role of the United Nations is characteristic of these two countries: there is high support for UN intervention in the case of serious human rights violations occurring in a country, but simultaneously fewer respondents are in favor of a stronger United Nations. The citizens of both countries broadly (well above-average) support the idea that citizens’ organizations should be directly involved in the decision-making process in international organizations. Compared to the other clusters, the proportion of respondents who see themselves as politically competent is the highest in this cluster. There is also wide support for the normative concept of a responsible citizen in both countries of the cluster. An Outline for a New Concept of Citizenship in a Trans-National Context 205 Figure 8.01 Classification of Countries Regarding the Presence of Elements of‘Post-Modern’ Citizenship      ­   € ‚ ƒ„  † ‡€€ ‡ ‡ ˆ‚ ‚ ‰ Š  € ‹ Œ€ € iSSP 2004. results of hierarchical cluster analysis - dendogram using Ward method (B) The second group consists of three other clusters of countries where we are unable to speak of the clear and unambiguous presence of elements of the post-modern concept of citizenship. Cluster 4 – new democracies (from the third wave of democratization) with recent experience of authoritarian rule (Portugal, Bulgaria, Hungary, Latvia, Poland, Russia, Slovakia, Chile, Taiwan). According to Freedom House ratings, all countries in the cluster are classified as free (democratic) countries, although not with the highest rating. There are two exceptions: the highest rating for Portugal and the classification of Russia as a ‘partly free country’. Considering the dimensions of the postmodern concept of citizenship, the most important characteristics of this 206 Mitja Hafner-Fink | Samo uhan cluster are the following: a very low proportion of participation and a low proportion of citizens that see themselves as politically competent. According to respondents’ perceptions (a very small proportion of respondents do not see corruption as a problem), we can say that countries in this cluster have problems with systemic corruption. Cluster 5 – a group of four very different countries (Czech Republic, Cyprus, Japan, South Korea), although three of them (Cyprus, South Korea, Czech Republic) have experience with being divided. The main characteristics of this cluster are the low (below-average) presence of all dimensions of post-modern citizenship. There is only one exception: a somewhat higher proportion of support for a stronger United Nations. Cluster 6 – three Latin American countries and one Asian country with specific economic experience and with different experiences of a democratic deficit (Mexico, Uruguay, Venezuela, Philippines). The most outstanding characteristic of this cluster is the very high proportion of respondents who support the idea that citizens’ organizations should be directly involved in the decision-making process in international organizations. But, at the same time, there is a low proportion of those who support a strong United Nations. We can also see a high level of support for minority rights and for a normative concept of a responsible citizen. On the other side, there is a low level of citizens’ participation and a low proportion of respondents who evaluate themselves as competent to participate in or follow political events. According to the perceptions of respondents, countries in this group face a large degree of systemic corruption. b) European Situation (Data from the ESS 2002) Data from the ESS 2002 also offer us a limited possibility to observe the presence of some elements of the post-modern concept of citizenship on a comparative European level – we evaluate 21 European countries (Israel was excluded from the analysis). Using these data we are able to observe the fol owing indicators (see Appendix 2 for details) of the five dimensions of global citizenship: – support for international (global) governance (trust in the United Nations; support for the international level as being the level where several policies should mainly be decided); – a (normative) perception of the good citizen as an active citizen (importance of being active in politics and in voluntary organizations in order to be a good citizen); – tolerance towards immigrants and cultural diversity (support for immigrants’ rights, understanding cultural diversity as something positive); An Outline for a New Concept of Citizenship in a Trans-National Context 207 – subjective competence (self-evaluation of the ability to understand politics and to form opinions about political issues); – individual (protest) political actions (boycotting products, buying products for political, ethical or environmental reasons, signing petitions). Using the same procedure as for the ISSP data (hierarchical clustering method of aggregated individual data on the country level) we have, on the basis of these five dimensions of global (“post-modern”) citizenship, obtained a classification of European countries. Results suggest that the classification into three groups of countries is the simplest and most meaningful solution (see Figure 8.02). Figure 8.02 Classification of Countries Regarding the Presence of Elements of ‘Post-Modern’ Citizenship     €‚    eSS 2002. results of hierarchical cluster analysis – dendrogram using Ward method Group 1 is the biggest (12 countries) and consists mostly of old democracies from Western and Central Europe. There are only a few exceptions: Spain, Hungary and the Czech Republic (Figure 8.02). The group is very heterogeneous and can be split into several subgroups. But in general the group is outstanding regarding citizens’ greater openness for religious (cultural) diversity. 208 Mitja Hafner-Fink | Samo uhan It probably reflects the actual situation of cultural diversity and the historical experience. Group 2 consists of four Scandinavian countries (Denmark, Finland, Norway and Sweden). It seems to be the group with the most widespread elements of post-modern citizenship. In this group of countries we can see the greatest trust of respondents in the United Nations. This is also the group with the highest proportion of active citizens. On the other side, in the countries of this group we cannot speak of a particularly high level of support for the normative concept of an active citizen. Group 3 comprises South European countries (Greece, Italy and Portugal) and two post-socialist countries (Poland, Slovenia). We can say that all these countries are new democracies from the third wave of democratization. All countries from this group are also very homogeneous from an ethnic and religious point of view. We can assume that this characteristic is reflected by the fact that citizens of countries from this group express lower tolerance towards immigrants and towards cultural diversity. There is also a low level of individualized (protest) political participation, while on the other hand we can see widespread support for the normative concept of an active citizen. c) Discussion The results may be discussed in various ways. One of them is offered by Stephan Castles, who speaks of hierarchical citizenship [Castles, 2005]. Castles has developed a system of classification that consists of five groups of countries, ranging from those in which elements of citizenship are ful y realized to those where we cannot find any elements of citizenship and where, if we use Castles’ words, we can only speak of ‘non-citizens’ [Castles, 2005:691]. Like in Castles’ classification system we can, on the basis of our analysis, identify two large groups of countries on the global level (the ISSP data): a) The first group consists of highly developed countries with long democratic traditions and fully established democratic institutions. We can find here just two exceptions – Slovenia and Spain, both of which were democratized in the third wave of democratization. This group corresponds with the first two of Castles’ groups (the US and highly developed countries from the EU and Canada, Australia etc.). b) The second group consists of two subgroups of less developed countries (Japan is the only exception): new democracies (from the third wave of democratization in Europe) and other countries with a democratic deficit from Latin America and Asia. The group strongly corresponds to Castles’ third group: ‘transitional and intermediate countries’ with ‘lower standards of rights and legal protection’ [Castles, 2005: 691]. An Outline for a New Concept of Citizenship in a Trans-National Context 209 Considering the results of our analysis we can pose the following question: in which part of the world can the elements and conditions for the ‘postmodern’ concept of citizenship be found? There is no uniform answer to this. Although elements of the concept are present in both groups, it is obvious that we can only find institutional conditions (democratic institutions, legal order, rights) and political-cultural elements (competent citizens, political participation) in the first group. In the second group, we can only speak of some citizens’ support for global governance or a global civil society as a reaction to the democratic deficit in their countries. Figure 8.03 European Countries – Relation Between Support of Citizens for Immigrants’ Rights and Trust in United Nations  ­ ‚ƒ­ ‡ˆ     ‡†     „ ‡ € ‡‰ † † (Mean Values; eSS 2002) In general, the analysis of European data [ESS, 2002] reveals the same pattern: in highly developed countries with longer democratic traditions (particularly in Scandinavian countries), we can find more elements of postmodern citizenship. We can see a very clear pattern of positive correlation between trust in the United Nations (a kind of global governance) and support of equal rights for immigrants. We can observe the highest level of both in the Scandinavian countries (Figure 8.03). 210 Mitja Hafner-Fink | Samo uhan But we also find some interesting results that bring under question the consistency of our model of post-modern citizenship. Namely, in countries with high support for normative concept of active citizen, we find very low individualized (protest) political activities of citizens – this is typical for countries from the third “European” group (Greece, Italy, Poland, Portugal, Slovenia) (Figure 8.04). When we observe the analyzed European countries, we are not able to find a clear pattern. On the basis of the presence of these two dimensions, we can classify the countries into three groups: a) one with a very low level of participation (less developed, new democracies from the South and East); b) one with a moderate level of participation (old, Western democracies); c) one with a high level of participation (highly developed Scandinavian countries and Switzerland). Within each of these groups we are able to demonstrate the following pattern: negative correlation between the extent of individual (protest) political activities and the extent of (normative) perception of the good citizen as an active citizen (Figure 8.04). 8.5 Does this Mean we Need a New Paradigm? Theses that the nation-state is completely anachronous are unfounded. The nation-state remains sovereign in many respects; its role is still important in assuring the legal, social and economic security of citizens, and it remains an indisputable authority in ensuring its territorial integrity. The impacts of globalization that diminish the sovereignty of nation-states are primarily evident in areas related to ecology, the economy and military security. The demands of individuals and minorities for the assurance of special rights (whether ethnic or other minorities) are more strongly articulated, and the obligations of states regarding issues that stretch across national borders (such as ecology) are more pronounced. In this respect the claim that full rights can only be accorded to individuals who are permanent residents and assimilated into the dominant cultural model is losing ground (the challenges of multiculturalism and the rights of non-citizens who do not intend or want to become citizens). These challenges necessitate the construction of a theoretical framework that treats the individual’s legal status as a principle of membership in an abstract state or political system, separating it from the classical citizenship status. But here it does not suffice to decouple the rights of citizens and the recognized rights of minorities; instead the model of the three components has to be extended or replaced with a synthesis that would mean a higher political level. An Outline for a New Concept of Citizenship in a Trans-National Context 211 Figure 8.04 European Countries – Relation Between Extent of Individualized (Protest) Political Activities (Boycotting, Deliberately Buying Certain Product, Signing Petition) of Citizens and Support for a Norm of Active Citizen „„ ­    „† ­  „„ ‚ƒ  „  €  „ „ „ (Mean Values on Constructed Indexes; eSS 2002) On the practical level this means enabling individuals to enjoy a relatively broad spectrum of citizenship rights, even when they are not formally citizens of the states they live in. It also means that the citizens of a particular country rely on a supranational legal and political body for the protection of their rights. Both options exclude the full sovereignty of the nation-state over its residents and, at the same time, emphasize the significance of civil society in contemporary democracies. Demands to broaden special rights or to create new forms of protecting rights, irrespective of formal citizenship, are related to the strengthening of civil society. As the civil sphere of society expands, conditions are created for the ‘socialization’ of economic and technological processes that unfold on the supranational level and simultaneously for a framework that enables the individual to influence the political processes. The creation of an international civil society thus lends political weight to the universality of human rights. The serious deficiencies of international political and civil society bodies must not be overlooked here, namely their lack of legitimacy or democratic 212 Mitja Hafner-Fink | Samo uhan deficiency (is it possible to have a democracy without a demos). This is evident in the abstractness of the discourse on political rights and the absence of subjects with ‘clout’ who can assure the exercising of rights. The emergence of international non-governmental organizations partly addresses this deficiency and primarily lays down a normative framework for the implementation of the principles of a global political culture. The function of the post-modern concept of citizenship in Cohen’s view [Cohen, 1999] transcends both the liberal and the republican models of citizenship insofar as the weakness of the liberal model lies in its underestimation of democratic participation and its failure to shape a symbolic identity, thus suffering from a deficit of political legitimacy, while the most important weakness of the republican model is its exclusivity. Hence, the primary role of the post-modern model is to strengthen the universality of human rights, their international institutionalization and articulation at all levels in order to strengthen and protect the rights of individuals and groups. Of course, the foregoing does not imply accepting the notion of the perfection of civil society. Civil organizations can also be unfair, non-egalitarian and exclusive, as many examples show. For precisely this reason, the greatest challenge facing civil society organizations is probably the implementation of democratic principles in their operations. Here the issue of identity arises again as a motivating force. The question then is: how does one determine the place of democratic participation in the model of the post-modern paradigm of citizenship, which asserts political rights and defines them at various levels of the abstract political body, which relinquishes the fantasy of absolute authority, power and identity as such. The answer seems simple – if citizenship in the post-modern paradigm is to have any meaning, it has to include democratic participation with the experience of power and not just influence (that is the possibility of co-formulating legislation and policies). Here it is important that the rights of non-citizens are treated as universal rights [Cohen, 1999:264] and not as special rights. It would be naive to expect the post-modern paradigm to completely defuse the tensions between the various components of the concept of citizenship. But these tensions can be reduced by relinquishing the exclusive notion of territorial sovereignty, by boosting the plurality of instances protecting the rights of the individual and by precisely defining the key rights international institutions assure the individual as a member of civil society and as a member of transnational organizations. The post-modern paradigm also requires a thorough consideration of the dimension of identity. The question of national identity in the post-modern An Outline for a New Concept of Citizenship in a Trans-National Context 213 paradigm is open for democratic redefinition, but the entire concept of identity has to be reshaped and made more responsive to the existence of different kinds of identity. Various levels of allegiance have to be determined and the loci of identity – local, regional, national and global – have to be defined. The post-modern paradigm also requires the differentiation of the forms of political participation and a determination of the intersections of rights, duties and allegiances in the global context. All of this has institutional implications. The implementation of a post-modern paradigm of citizenship is not seen as creating some super-federation or some world government and, similarly, not as the adoption of a cosmopolitan legal order. Nonetheless, it will probably be incumbent on all governments to strengthen their receptors to the inputs of civil society. We share the view of Joe Painter [Painter, 2002], who sees the post-modern paradigm of citizenship as a kind of synthesis of classical conceptions of citizenship, where the main characteristic of this synthesis should be the idea of multi-level citizenship: multi-level political communities (civil societies), the co-existence of different levels of democratic governance (from local to global), multiple (situational) identities, and post-modern (post materialistic) values (environmentalism, tolerance and heterogeneity). APPENDIX 1 Indicators used for dimensions of the postmodern conception of citizenship from the ISSP 2004 – Citizenship a) Support for Transnational Civil Society and for Global Governance Now we would like to ask your opinion about international issues. V49 Thinking about the United Nations, which comes closest to your view? 1 –The United Nations has too much power 2 – The United Nations has about the right amount of power 3 – The United Nations has too little power 4 – Don’t know what the United Nations is 8 – Can’t Choose V50 Which of these two statements comes closer to your view? 1 – In international organizations, decisions should be left to national government representatives 2 – In international organizations, citizens’ organizations should be involved directly in the decision-making process 8 – Can’t choose V51 Which of these two statements comes closer to your view? 1 – If a country seriously violates human rights, the United Nations should intervene 2 – Even if human rights are seriously violated, the country’s sovereignty must be respected, and the United Nations should not intervene 3 – Don’t know what the United Nations is 8 – Can’t choose An Outline for a New Concept of Citizenship in a Trans-National Context 215 b) A “Normative” Perception of a Responsible Citizen There are different opinions as to what it takes to be a good citizen. As far as you are concerned personally on a scale of 1 to 7, where 1 is not at all important and 7 is very important, how important is it: Not at all Very Can’t important important choose V7. To keep watch on the actions of 1 2 3 4 5 6 7 8 government V8. To be active in social or political 1 2 3 4 5 6 7 8 associations V9. To try to understand the reasoning 1 2 3 4 5 6 7 8 of people with other opinions V10. To choose products for political, ethical or environmental reasons, even 1 2 3 4 5 6 7 8 if they cost a bit more. V12. To help people in the rest of the 1 2 3 4 5 6 7 8 world who are worse off than yourself c) Understanding of Rights as a Universal Non-Exclusive Category There are different opinions about people’s rights in a democracy. On a scale of 1 to 7, where 1 is not at all important and 7 is very important, how important is it: Not at all Very Can’t mportant important Choose V31 That government authorities respect and protect the rights of 1 2 3 4 5 6 7 8 minorities V32 That government authorities treat everybody equally regardless of their 1 2 3 4 5 6 7 8 position in society 216 Mitja Hafner-Fink | Samo uhan d) Subjective Competence To what extent do you agree or disagree with the following statements? Strongly Agree Neither Disagree Strongly Can’t Agree Agree Disagree Choose nor Dis- agree V38. i feel i have a pretty good understanding of the 1 2 3 4 5 8 important political issues facing (COuNTrY). V39. i think most people in (COuNTrY) are better informed 1 2 3 4 5 8 about politics and government than i am e) An Individual as an Active Citizen Here are some different forms of political and social action that people can take. Please indicate, for each one, – whether you have done any of these things in the past year, – whether you have done it in the more distant past, – whether you have not done it but might do it – or have not done it and would never, under any circumstances, do it. Have Have Have not Have Can’t done it in done it done it not done choose the past in the but might it and year more do it would distant never past do it V18. Boycotted, or deliberately bought, certain products for political, 1 2 3 4 8 ethical or environmental reasons V23. Contacted or appeared in the 1 2 3 4 8 media to express your views V24. Joined an internet political forum 1 2 3 4 8 or discussion group APPENDIX 2 Indicators for dimensions of the postmodern conception of citizenship from the ESS 2002 a) Support for International (Global) Governance CARD 11: Using this card, please tell me on a score of 0-10 how much you personally trust each of the institutions I read out. 0 means you do not trust an institution at all, and 10 means you have complete trust. No trust Complete (Don’t at all trust know) B12 … the united 00 01 02 03 04 05 06 07 08 09 10 88 Nations? CARD 15: Policies can be decided at different levels. Using this card, at which level do you think the following policies should mainly be decided? READ OUT AND CODE ONE ON EACH LINE international european National regional Don’t level level level or local know level B 35 ... protecting the 1 2 3 4 8 environment B 36 ... fighting against 1 2 3 4 8 organised crime B 37 ... agriculture 1 2 3 4 8 B 38 ... defence 1 2 3 4 8 B 39 ... social welfare 1 2 3 4 8 B 40 ... aid to developing 1 2 3 4 8 countries B 41 ... immigration and 1 2 3 4 8 refugees B 42 ... interest rates 1 2 3 4 8 218 Mitja Hafner-Fink | Samo uhan b) A (Normative) Perception of the Good Citizen as an Active Citizen CARD 46 To be a good citizen, how important would you say it is for a person to… extremely extremely (Don’t unimportant important know) e26 …be active in voluntary 00 01 02 03 04 05 06 07 08 09 10 88 organisations? e27… be active in 00 01 02 03 04 05 06 07 08 09 10 88 politics? c) Tolerance towards Immigrants and Cultural Diversity CARD 26 Using this card, please say how much you agree or disagree with each of the following statements. Firstly… READ OUT Agree Agree Neither Disagree Disagree Don’t (don’t agree nor strongly know strongly) disagree D22 People who have come to live here should be given 1 2 3 4 5 8 the same rights as everyone else D28 CARD 30 And, using this card, would you say that [country]’s cultural life is generally undermined or enriched by people coming to live here from other countries? Cultural life Cultural life Don’t undermined enriched know 00 01 02 03 04 05 06 07 08 09 10 88 An Outline for a New Concept of Citizenship in a Trans-National Context 219 CARD 38 Using this card, please tell me how much you agree or disagree with each of these statements. Firstly… READ OUT Agree Agree Neither Disagree Disagree Don’t (don’t agree nor strongly know strongly) disagree D41 it is better for a country if there 1 2 3 4 5 8 are a variety of different religions d) Subjective Competence B2 CARD 6 How often does politics seem so complicated that you can’t really understand what is going on? Please use this card Never 1 Seldom 2 Occasionally 3 regularly 4 Frequently 5 (Don’t know) 8 B4 CARD 8 How difficult or easy do you find it to make your mind up about political issues? Please use this card Very difficult 1 Difficult 2 Neither difficult nor easy 3 easy 4 Very easy 5 (Don’t know) 8 220 Mitja Hafner-Fink | Samo uhan e) Individual (Protest) Political Actions ASK ALL There are different ways of trying to improve things in [country] or help prevent things from going wrong. During the last 12 months, have you done any of the following? Firstly … READ OUT Yes No (Don’t know) B19 Signed a petition 1 2 8 B21 Boycotted certain products 1 2 8 B22 Deliberately bought certain products for 1 2 8 political, ethical or environmental reasons Trust in Institutions of the System – European Comparisons – and Slovenia 9 Niko Toš | Vlado Miheljak | Slavko Kurdija 9.1 Introduction and Definition of the Problem Definition. In their classic study on political (civic) culture, Almond and Verba [1963] emphasize the necessity of trust in order for democracy to function properly. Trust, therefore, becomes a central concept in the study of political culture, on which Inglehart also reports exhaustively [1997]. A survey of people’s attitudes towards politics, and more specifically, of people’s attitudes towards the political system, is based on the premise that the level of its legitimacy differs from one cultural and historical system to another. This can be seen in the different levels of trust in particular political and institutional systems. Whereas institutions in developed democracies are believed to enjoy higher levels of trust, it is assumed that trust in new democracies is on the rise. Needless to say, this is but a simple indication of the different levels of trust in particular systems.1 In recent decades, researchers have found several examples of initiatory and transitional circumstances in the expression of trust in the institutions of democratic systems. Lipset and Schneider [1987], for example, found a significant confidence gap with regard to the most important social and political institution in the USA. Although citizens are aware of the importance and indispensability of the institutions, they are increasingly critical of and indifferent to them. Similar findings were made by researchers in Germany, especially with regard to attitudes towards political parties. Research conducted in recent decades into trust in institutions has resulted in the identification of several important paradigms. These include the reciprocal relationship between trust in individuals and trust in institutions, the influence of conflicting structures on trust, the influence of preferential adhesion to political parties on trust, the influence of experience on trust, and the influence of the accumulated experience of the elderly on higher levels of trust – or a higher level of conformity, etc. Our analyses of trust in institutions are supported in Easton [1975], who describes political trust as the feeling that one’s own interests are taken into consideration by the political process. When one trusts another individual or institution, they express a positive feeling (emotional component) and attribute a positive value (evaluative 1 We must not overlook the fact that a highly developed democratic system also presumes, even postulates, the expression of distrust. Distrust is an expression of discord between prevailing individual and collective values, between institutions as the proponents of collective values and the representatives of institutional authority. If the centers of authority are not intimately linked to the values of the system, or if they have no clear value orientation, we can expect social instability, increased uncertainty, and a decline in social activity. A positive response to such a situation would be to initiate reinstitutionalization. 224 Niko Toš | Vlado Miheljak | Slavko Kurdija component) towards that person or institution, and the evaluation is normally based on moral norms and values criteria. In addition to emotional, affective and evaluative components, trust also comprises cognitive elements, which are more difficult to determine. Trust is always linked to some “object”, an individual (a friend or group of friends, such as colleagues), an institution (parliament, the Catholic Church, etc.), or a system (a social state, market economy) as a collective subject. Trust is based on qualities and characteristics that are seen as positive, and the person expressing the trust is not always necessarily aware of these qualities. It is important for the purpose of evaluating the trust in institutions to distinguish between specific and diffuse trust, and to determine how strongly trust is associated with individual subjects and under which conditions it emerges. Whether trust is specific or diffuse depends on our comprehension of the concept in each context. If the individual expressing the trust expects a specific result or service, trust can be defined as specific. If, however, the subject of the trust is a system and not an individual or specific organization, then the trust is diffuse relative to the subject. We must not assign absolute qualities to trust, because these are often derived from its diffuse nature. Trust is not always entirely independent of specific results. Resistance to disappointment is, above all, related to the diffuseness of expectations. Anyone expecting specific results will be disappointed if those results do not materialize and will react by withdrawing their trust. The probability of trust being generalized or transferred is influenced by the specific subject of the trust. Trust in an individual (an official, for example), does not necessarily translate into trust in an institution. If, however, one trusts in democracy (a diffuse object), they cannot feel a lack of trust in all democratic institutions. Generally, when trust is surveyed, several gaps emerge in the results if the survey only takes the subject of the trust and not the expectations into account, assessment criteria and assessed values are not. The minimum prerequisite for trusting an institution is that the institution is perceptible to the individual. A direct relationship based on personal contact is not necessary; credible mediators will do. It is much easier for us to develop trust in institutions to which we assign positive characteristics. Our values may determine how quickly our trust develops. Societal values and ideologies thus significantly affect our trust in institutions. Individuals comprehend the ambitions of institutions in different ways. In typical working democracies, political rivals and opponents are free to follow their aspirations and to play by the rules in a free contest. However, even in democracies, trust in institutions is not evenly distributed. Whenever disputes and rifts between groups of people endure, it becomes very difficult for Trust in institutions of the System – european Comparisons – and Slovenia 225 one of the disputing sides to be trusted by the other. In the case of Slovenia, it would appear that these enduring divisions are related to its recent history, the Partisan and Home Guard movements, national reconciliation, Communism and anti-Communism, the left and right political wings. Needless to say, it may also be that cultural, religious and social divisions have become increasingly important. These differences are ideological in nature: for example, the libertarian-conservative relationship, or (in an atmosphere of increasingly assertive, neoliberal economic principles) the divide between workers and trade unions on the one hand and managers and capitalists on the other. When trust in an institution is reasserted repeatedly, it may become firmly established and independent of specific results. When trust grows over an extended period (a lifetime, for example), a positive relationship may emerge between age and trust in institutions. This is a socialized assumption: trust in institutions stems from childhood. It is worth noting that certain levels of trust develop between people during primary socialization. With growing experience and the emergence of social perspectives during secondary socialization, this trust broadens to include basic groups, eventual y institutions, and the system as a whole. The expression of trust also depends on short-term circumstances; trust can be shaken by negative results, distinct changes in institutions, etc. This is particularly true when the analysis not only encompasses diffuse, stable trust which is resistant to disappointment and which evolves over a long period of time, but also the expression of (dis)trust as a response to a specific, short-term event. 9.2 Surveys to Date – Secondary Sources Survey of Trust in Western, Central and Eastern Europe in the 1990s2 In the 1990s, the most important empirical analysis of trust in institutions was undertaken [Listhaug, Wiberg, 1995]. The research was based on extensive empirical data collected in 14 European countries over the course of two observation periods.3 The analysis covered ten institutions, including the Church, armed forces, trade unions, police, parliament, major corporations, 2 In reviewing past surveys of trust in institutions, we relied on the most recently published study in Slovenian by Rus/Toš [2005] and its supplement [Toš, 2007]. The study exhaustively describes the empirical sources used in this part of the paper. The first part of the review comprises international comparative surveys from the 1990s, and the second part comprises surveys conducted in Slovenia from the 1990s up to 2006. 3 The European Value System Study Group, 1981, 1990. 226 Niko Toš | Vlado Miheljak | Slavko Kurdija etc. Richard Rose’s analysis [Rose, 1984], which classifies six institutions as government (the armed forces, education system, judicial system, police, parliament, and social services), and four as non-government (the Church, the press, trade unions, and major corporations), is also important in understanding the issue. Results show that there is a greater willingness to trust state institutions than private ones. If, however, the analysis includes the results of additional surveys, one can see that trust in government institutions is more or less stable or is declining slightly, while trust in private institutions is slightly increasing. For example, trust in government institutions is decreasing in Norway, Belgium, Sweden and the United Kingdom, while it is growing in Denmark; the most discernible increase in trust in private institutions can be seen in Italy, the Netherlands and Spain. It must be noted, however, that in most countries during the 1990s, one sees a greater inclination to trust government institutions than private institutions.4 Within the group of government institutions, parliament has enjoyed a relatively high standing in all observed countries, particularly in Norway, Finland, Germany, Ireland, the Netherlands and Spain. Intermediate comparisons show that the high levels of trust enjoyed by parliament between 1982 and 1990 are visibly declining. This is particularly pronounced in Norway, applies partly to Iceland, Germany, Ireland, France and Spain, and is standard for the 14 European countries under observation. Some of the countries display the opposite trend, namely, growing levels of trust in parliament, as is the case in Belgium, the Netherlands, the United Kingdom and Italy. Also apparent are the different levels of trust in different countries. Despite the negative trend, the levels of trust in parliament in Norway are still relatively high. The same applies to Finland, the Netherlands, Germany and, to a slightly lesser degree, Italy as well. State institutions enjoying the highest levels of trust include the education system, the armed forces and the police. The trend is also positive in these cases. This prevailing positive trend is characteristic for the education system, which enjoys high levels of trust. Only Belgium, the Netherlands, Italy and especially the United Kingdom show a departure from this trend. The police enjoyed a relatively high level of trust over the course of several observation periods in all countries; however, a negative trend was observed in individual countries (Sweden, Belgium, the United Kingdom). The Church occupies a 4 Basanez, M., Inglehart, R., Moreno, A.: Human Values and Beliefs, A Cross Cultural Source-book, World Value Survey 1990–1993; International Values Survey, SJM 92, SJM 95 in SJM 99. See Toš, N., ur., Vrednote v prehodu I–III, Dokumenti SJM, Ljubljana 1997–2004. Trust in institutions of the System – european Comparisons – and Slovenia 227 special position among civil institutions. The highest levels of trust in the Church – far higher than in any other institution – were expressed in Ireland and Iceland, albeit a negative trend has also been observed in both countries. In all other countries, a more or less negative trend was observed, especially in Belgium, Germany, Ireland, the Netherlands and Spain. This is an indication of the progression and the extent of secularization in Western Europe during the observation period. If we look at the average expression of trust in the Western European countries under observation, we notice that the highest levels of trust are seen in the police, followed by the education system, judicial system, armed forces, the Church, large corporations, parliament, trade unions, etc. The lowest levels of trust are observed in the press, followed by the trade unions. Parliament ranked sixth out of ten observed institutions. A Comparison of Studies Carried out in Slovenia (WVS, 1992)5 A comparison of studies carried out in Slovenia (WVS, 1992) and in Western European countries during the same period (1992–1994) shows a number of differences. In Slovenia, the level of trust in parliament is only half of the European average and far lower than in the observed countries. Slovenia already departs from the European average by having low levels of trust in the police, judicial system, trade unions, the Church and corporations. Trust in the education system is on par with the European average, while trust in the armed forces is slightly below the European average and trust in the mass media is significantly below the European average. It is apparent that Slovenia departs from the European average when it comes to trust in institutions. However, it is also clear that in Slovenia, the highest level of trust is enjoyed by the education system, and that the police and the armed forces also enjoy high levels of trust. The results of studies (1995 – 1999) seem to confirm this first impression. The final study shows that trust in parliament is at an exceptionally low level – far lower than initially; the same applies to public services. There is a strong affinity between trust and the education system and the mass media. Here, too, it becomes evident that Slovenia is a young democracy in which state institutions with a tradition (the police or armed forces, for example) enjoy higher levels of trust than the legislative branch (a politically plural parliament representing the democratic nature of the system). Is this a conflict between the “undemocratic” nature of the state and its citizens? Slovenia is, undoubtedly, still undergoing the legitimation of its democratic system. 5 World Value Survey, WVS Group, Ronald Inglehart et al. 228 Niko Toš | Vlado Miheljak | Slavko Kurdija Results of Surveys on Trust in Institutions in Central and Eastern European Countries6 Data collected in 1995 for ten Central and Eastern European countries indicates7 that trust in institutions in these countries is at a much lower level than in Western European countries. This is particularly true for political or state institutions. In addition, we also notice differences in levels of trust between individual Central and Eastern European countries. Of the observed countries, trust in the government is highest in Slovenia and the Czech Republic; trust in parliament is highest in Croatia and Slovenia; trust in the Church and clergy is highest in Romania and Ukraine; there is a high level of trust in banks in Slovenia, Slovakia, Bulgaria and the Czech Republic, and a very high level in Ukraine and Belarus. Of the observed institutions, the Church enjoys the highest level of trust in most of the countries, especially in Romania, Ukraine, Belarus and Poland, while the lowest levels of trust are found in Slovenia and the Czech Republic. The president of the republic, in most cases elected directly, enjoys a relatively high level of trust, except in Slovakia, and this is significantly higher than for parliament. Social movements and the mass media generally enjoy a high level of trust – much higher than, for example, parliament or the government. This leaves the impression that the legitimacy of government institutions is subject to their cooperation with non-governmental institutions. A similar survey conducted on a group of Eastern European countries [Plasser, Ulram, 1996] confirms that there are differences in levels of trust in political institutions in the observed countries. For example, the president of the republic enjoys the highest level of trust in Hungary, the Czech Republic and especially in Slovenia. This is probably related to the presidents’ personalities as well as to their limited constitutional power, which elevates them above daily political bickering by making them the representative of the system, moral arbitrator and proponent of internal balance and sensible compromise. This is in sharp contrast to the constitutional and political situation in Russia and Poland, where the president, with broad powers, direct influence on executive power and the right to veto, plays the role of a “presidentially charismatic” defender of the constitution [Offe, 1994]. In such systems, presidents do not enjoy a high level of trust, except in Croatia, where the president was highly 6 Based on studies conducted under the auspices of the New Democracies barometer [Rose, 1996]. 7 Rose, R., Haerpfer, Ch.: New Democracies, Barometer IV: A 10-Nation Survey, Studies in Public Policy, University of Strathclyde, Glasgow, 1966 Trust in institutions of the System – european Comparisons – and Slovenia 229 trusted in 1995, when the Homeland War ended. Other sources show that lack of trust in political institutions is particularly the case in Russia, where, with the exception of the armed forces – an object of national identification – all other institutions are subject to a high level of distrust and suspicion. This is similar in Poland, where the armed forces enjoy a high level of trust, in contrast to the political elite (parliament, government). Even traditional institutions, such as the Polish Church, have lost a significant share of their support and are subject to distrust at a similar level as parliament and the government. Similar changes in the patterns of trust can be observed in the Czech Republic, the difference being that the government enjoys a high level of trust, in contrast to the Church and parliament. By comparing the results of surveys conducted in Western, Central and Eastern Europe, we may conclude that, following the social changes of the mid-1990s, the levels of trust in Central and Eastern Europe are significantly lower than in Western European countries with long-established democratic traditions. The results also show that trust in civilian institutions is lower than in Western Europe, however, this departure is not as significant. As mentioned earlier, trust in the Church in the two predominantly Orthodox countries (Romania and Ukraine) far exceeds the European average, while trust in two predominantly (Slovenia) or partially (Czech Republic) Catholic countries is typically below the European average. This illustrates the peculiarity of the secularization process, which has not proceeded at the same rate or under similar conditions in all Eastern Bloc or post-socialist countries. Secularization in Eastern Europe was mainly coercive and often ostensible, while the primary religiousness of the people remained deeply rooted; in contrast, secularization in Slovenia and the Czech Republic was less influenced by political diction and – at least in Slovenia – was part of the general current that characterized the Western European social and cultural environment. In other words, they were part of the processes of modernization (and urbanization) and the prevailing liberalism. An analysis of the scale and power of trust in political institutions (parliament, government, the president) in Central and Eastern European countries gives us an expected ordering of countries: at the top of the list is Slovenia, followed by Croatia and the Czech Republic, while Ukraine and Belarus are at the bottom of the list. In the two highest ranked countries, the “pre-democratic” political practice, which evolved under the Yugoslav system of self-governing socialism, and the openness and western orientation of these two former federal entities of the defunct Yugoslavia should not be overlooked. Analyses from the mid-1990s have indicated that in Slovenia (and the Czech Republic), the government enjoys a very high level of trust and the president of the 230 Niko Toš | Vlado Miheljak | Slavko Kurdija republic even more so. The central bank also enjoys an unusually high level of trust, thus legitimizing the system’s economic dimension, as do women’s movements (also in the Czech Republic and Slovakia), which legitimizes the system’s human rights dimension. Survey of (Dis)trust in Institutions in Slovenia (1991 – 2006)8 After independence (1991), the sociological community in Slovenia participated in several international comparative projects (EVS, WVS, ISSP, CSES, ESS and others). Ever since, trust in institutions has been intensively surveyed, both conceptually as well as in terms of the institutions themselves and the frequency of the surveys (see table). This chapter focuses on the results of the most recent survey (2006), and also illustrates trends in the expression of trust. The SPOS 20069 survey applied a four-level confidence scale10 and was conducted on 20 institutions, including eight institutions representing the state (political) sector and others representing the civil society, the institutional and the individual environment. The resulting ranking shows that of all surveyed institutions, the highest average level of trust is enjoyed by family, while political parties have by far the lowest average. Our observations, however, should not be limited to the rank of the institutions; rather, we should attempt to show the relationship between the ranks of individual institutions and their broader significance and affiliation. We can do this by clustering the institutions into groups or categories according to their characteristics. The first is the group of strictly governmental institutions (parliament, government, the prime minister, the president of the republic), the disciplined and security services (police, armed forces) and the judiciary (courts of law). Next are the institutions related to the welfare state: commercial and financial institutions (banks, companies, the Slovenian Tolar) and civil institutions (schools, mass media, the Church and clergy). Political parties may also be considered among the state institutions (a precondition for their functioning), especially in relation to legislative and executive authority. Finally, there is the independent group, which includes international organizations (UN, NATO, EU). The ranking of these groups by trust level places the institutions with which people are intimate, as well as civil, welfare state and international institutions at the top of the scale, 8 Faculty of Social Sciences, Centre for Public Research under Prof. Niko Toš. 9 Slovenian Public Opinion Survey 2006/2, Role of Government, CJMMK, representative sample of adult population in Slovenia, N = 1516, October-December 2006. 10 Scale: 1 – do not trust at all / 4 – trust entirely. Trust in institutions of the System – european Comparisons – and Slovenia 231 while state institutions (political parties, parliament, the prime minister, government) rank at the bottom, as do the Church and clergy as the only civil institution so low on the scale. It is evident that, in addition to trust in family which ranks highest, people look for and find social stability mainly within the context of civil society, and that they feel limited and restricted by, even excluded from, the political and religious context. The institutions can also be ranked according to different criteria, namely, the relationship between trust and distrust.11 Here (Figure 9.01), the institutions are still ordered as before, however, the perception in the minds of people of their position in the “treasury of trust” is even more acute.12 Figure 9.01 Trust in Institutions    ­ € ‚ ƒ € „ƒ € ƒ †ƒ ‚ƒ‡ˆ ‰ƒ ­ Š ‚‰  „  ˆˆƒ‹ ‚ ‚‚ Aggregate of answers: 1-none + 2-slight and 3-very strong + 4-complete Source: Faculty of Social Sciences – Public Opinion and Mass Communication Research Centre, Slovenian Public Opinion Survey 2006/2, N=1003 In addition to family, which enjoys a high level of trust, the group in which trust characteristically outweighs distrust includes schools, the Slovenian Tolar, banks, the European Union and commercial institutions, including companies; the UN also falls into this group. Next is the group in which trust does not outweigh distrust as heavily. This group includes trade unions, armed forces, the judiciary, the president of the republic, media and the UN. The last group includes institutions in which the level of distrust far outweighs the 11 Trust: complete (4), very strong (3); Distrust: slight (2), none (1) 12 The Figure only shows the values for trust and distrust. The difference up to 100% represents the undecided. 232 Niko Toš | Vlado Miheljak | Slavko Kurdija level of trust.13 These are institutions that guarantee the functioning of the state and its democratic character. Determining Factors and Expression of Trust in Individual Institutions and Groups of Institutions We shall illustrate the structural characteristics of the expression of trust in institutions. We shall be more thorough in the discussion of state institutions and less so in the discussion of civil institutions, particularly where there are small differences in the levels of trust. Analyses of trust in family and political parties will not be discussed, as there are only slight differences in the prevailing expression of trust and distrust (4.9% for family and 7.6% for political parties), leaving little to analyze. a) National Assembly, Government and Prime Minister. The national assembly ranks the lowest of all institutions in terms of trust. The number of respondents expressing trust in the national assembly increases or decreases depending on age and to some extent on education, however, the most pronounced fluctuations are seen in connection with party preference, satisfaction with democracy and trust in the prime minister. Respondents who prefer coalitions evaluate state institutions positively when they have the feeling of participation. The other two lowest placed state institutions are the government and the prime minister. Typically, the level of distrust expressed in the government heavily outweighs the level of trust. Among the considered groups, the highest level of trust in government is expressed by supporters of the coalition parties. An unusually high level of trust in the government is expressed by respondents who also show a high level of trust in the Church, worshippers who regularly attend weekly religious ceremonies, respondents who are satisfied with the level of democracy or self-declared supporters of the political right. The highest level of distrust in the government is expressed by supporters of opposition parties, left wing supporters, individuals from lower social strata, respondents who do not attend religious ceremonies, the youngest respondents in the survey and secondary school and university students. A casual analysis shows that the harsh attitude of people towards government in general is primarily influenced by their attitude towards the principal politicians in government. 13 The percentage of distrust is greater than that of trust by a factor of two or more. At the bottom of this group, with the lowest trust and highest distrust by far, are the prime minister (1:2.1), the government (1:2.2), the Church and clergy (1:3.1), national assembly (1:4.6), political parties (1:12)d. Trust in institutions of the System – european Comparisons – and Slovenia 233 Similar rankings and correlations can be found regarding the support for the prime minister. It is interesting to note that trust in the prime minister changes dramatically from one age group to another: the youngest respondents express a lower level of trust than the oldest respondents. The trust expressed by the youngest respondents is far below average and a full three-quarters do not trust the executive branch. Among the oldest respondents, the level of trust in government is far above average while the level of distrust far below average, albeit prevalent. Analyses based on education show smaller fluctuations, however, those with a lower level of education expressed a lower than average level of distrust and those with a higher level of education expressed a higher than average level of distrust. Since levels of trust and distrust among students and pensioners are the same for the government and prime minister alike, the executive’s authority in pursuing development and reforms on behalf of the younger segments of the population is called into question. A comparison of the level of trust expressed in the government and prime minister does not show any typical deviation. As expected, the prime minister – as the personification of an institution – does not enjoy a high level of trust. Trust in the prime minister is also strongly influenced by ideological beliefs. The highest level of trust is expressed by supporters of the coalition parties, respondents who express trust in the Church or attend weekly religious ceremonies, respondents who are satisfied with the level of democracy, self-declared supporters of the political right, and those with negative views on socialism. Respondents from the youngest age group, students, supporters of opposition parties, and respondents who do not trust the Church or are not satisfied with the level of democracy have a high level of distrust. There is a conspicuous level of distrust towards the executive branch, which forms a fundamental part of the attitude towards the authorities held by the youngest, most highly educated, liberal, opposition supporting, and secular groups of voters. This again highlights the exclusion of a significant part of the electorate from the political process. b) President of the Republic. The president of the republic is the state institution that enjoys the highest level of support. The president is trusted by a good third of the respondents and is distrusted by two-thirds; generally, the president is ranked above the prime minister, government and national assembly. The aforementioned correlation between age and education is reaffirmed in this case. The youngest respondents express much higher levels of distrust than trust in the president of the republic, departing from the average in both cases. The oldest respondents express higher levels of trust and lower levels of distrust than average. Differences based on education are indistinct and 234 Niko Toš | Vlado Miheljak | Slavko Kurdija extrinsic. Supporters of the opposition parties are more trusting of the president of the republic, while supporters of the coalition parties, self-declared supporters of the political right, and respondents with negative views about socialism or who trust the Church are conspicuously distrusting. c) Church and Clergy. The Church and clergy are the lowest ranked of all civil institutions. Hardly a quarter of all respondents express trust in the Church. Analyses show that the expression of trust and distrust are unevenly distributed and confirm the logical and anticipated correlations based on age, education, employment, physical location – and, in particular, ideological belief. Analyses by to age group show a gradual and distinct shift towards trust in the case of the oldest respondents, and towards distrust in the case of the youngest respondents. A similar picture emerges regarding level of education. There is a high level of distrust among those with post-secondary and tertiary education. The level of distrust falls below the average among those with primary school education. Generally, the level of trust in the Church is above average among elderly respondents and those with lower levels of education, as well as the unemployed, people living under difficult material conditions, and the rural population. One can see a correlation between trust in the Church and clergy and (ideological) proximity to the ruling authority. Conversely, there is a very strong distrust of the Church and clergy among the youngest respondents, students, respondents with post-secondary and tertiary education, the urban population, self-declared left-wing supporters and supporters of opposition parties. In order to better understand the current low level of support for the Church and clergy in Slovenia, attitudes of believers, nonbelievers and those who attend religious ceremonies must be determined. One-half of the believers, who account for slightly more than half of the respondents, do not express trust in the Church, and a good two-thirds express distrust in the Church. The level of trust in the Church and clergy drops drastically among the religiously undetermined respondents, while only individual nonbelievers express trust in the Church and clergy. Practically all respondents who consider themselves nonreligious express distrust in the Church, as well as four-fifths of the religiously undetermined, and one-half of the respondents who consider themselves religious. A similar but more vague picture is drawn by analyses on the frequency of attendance of religious services: two-thirds of those who regularly attend Sunday Mass express trust and only one-quarter express distrust. This trend is reversed when we look at respondents who attend religious services less regularly: three-fifths do not trust the Church and two-fifths do. The percentage of those who trust the Church has fallen and is lowest among those who do not attend any religious services. As a rule, the Trust in institutions of the System – european Comparisons – and Slovenia 235 functioning of civil institutions is based on a high level of general trust for the more universally oriented institutions, or on a high level of segmental trust for institutions operating in narrower social segments. The Roman Catholic Church undoubtedly belongs to the former. Nominally, it focuses its activities on believers; however, its larger mission greatly exceeds this, as it pervades all of society in its pastoral role and as the defender of certain values. And from this perspective, the legitimate role of the Roman Catholic Church in Slovenian society is highly controversial, as it enjoys the majority support of only a small segment of those who regularly attend religious services.14 d) The Media. Among the observed institutions, television, newspapers and radio were ranked in the middle in terms of trust. The media held a relatively high and stable position in Slovenia during the late 1990s. The same applies to the SPOS 20062 survey, in which the media was ranked ninth among the observed institutions. It emerges that the expression of trust and distrust is influenced by education (a higher level of distrust of the media is expressed by respondents with secondary school or higher education). Among groups based on employment, only students and the youngest respondents show a markedly higher level of distrust than trust. Greater fluctuation in the expression of trust can be observed among supporters of political parties. Of all the variables included in the survey, two ideological variables exert the strongest influence on the level of trust or distrust: (1) a significantly larger percentage of the respondents who trust the Church also trust the media; (2) trust in the prime minister influences respondents’ opinions regarding the media. Three-quarters of the respondents who do not trust the prime minister also do not trust the media, while those who trust the prime minister express above-average trust for the media. e) Trade Unions. Trade unions fall under the category of civil institutions that ranked relatively low in Slovenia up to the late 1990s, but have moved up in the rankings to the top of the middle third in this decade. Intermediate comparisons indicate a high position and stability during the 2000–2006 period. In 2006, trade unions ranked far above most state institutions, the armed forces, NATO, the judiciary, the media and the police. Analyses hint at the circumstances that influence the expression of trust in the unions. It appears that factors influencing the expression of trust are life experience, political socialization under the previous social system, and employment status. As expected, employed persons and pensioners more often express their trust in trade unions than the unemployed and students. One can also see that very difficult material conditions influence the expression of distrust 14 This group represents approximately one-sixth of all respondents. 236 Niko Toš | Vlado Miheljak | Slavko Kurdija in the unions. Ideological belief also influences trust in trade unions, as illustrated by the findings that a significantly higher percentage of opposition party supporters express trust in the unions, while the opposite applies to supporters of the coalition parties. Another illustration of the influence of ideological belief is the fact that respondents with positive views about socialism and self-declared supporters of the political left also express a higher level of trust in trade unions. f) Judiciary and Police. The placement of the judiciary in terms of trust is somewhere in the middle; slightly less than three-fifths of the respondents express distrust and slightly more than a third express trust in the judiciary. Younger and better-educated respondents, pensioners and supporters of the political right, etc, express a higher than average level of trust. The lowest level of trust in the judiciary is expressed by the oldest group of respondents, respondents with the highest level of education, supporters of the political left, and those of the lowest social stratum. Actual surveys of trust place the police at the top of the middle third of the scale, immediately after the trade unions. The percentage of respondents who express trust remains characteristically lower than the percentage of those who express distrust. Analyses show uniformity in the responses; there are no differences between the age, education, or other categories. g) The European Union and the United Nations. The European Union is ranked in the first third, among the institutions that enjoy an above-average level of trust. Barely one-third of the respondents express distrust. Analyses show that the expressions of trust and distrust towards the European Union is rather monotonous and level. Respondents with lower levels of education, those who are dissatisfied with the level of democracy, those with negative views about capitalism, and respondents who are dissatisfied with their domestic material situation express above-average levels of distrust. Far higher levels of trust than distrust are expressed, especially among respondents with higher education, students, supporters of government coalition parties, self-declared supporters of the political right, respondents who are satisfied with the level of democracy, respondents with positive views about capitalism and a negative view about socialism, respondents who regularly attend religious ceremonies, etc. h) Slovenian Tolar and Banks. Of the civil institutions in the economic sector, the highest level of trust was expressed in the Slovenian Tolar, followed by banks and the economy (commercial companies). All three institutions are ranked very high up in the upper third of the scale, and, on average, the respondents characteristically express a higher level of trust than distrust. This Trust in institutions of the System – european Comparisons – and Slovenia 237 particularly applies to the Slovenian Tolar. In late 2006,15 almost two-thirds of the respondents expressed trust in the Tolar and only one-third expressed distrust. The oldest respondents, respondents with higher education, materially well-off respondents, supporters of the political left and right, respondents who are satisfied or very satisfied with their material situation, etc. expressed the highest level of trust in the Tolar. A correlation also appears between ideological and political beliefs. The Tolar enjoys the highest level of trust from supporters of the coalition parties, respondents who trust the prime minister and the Church, and respondents who are satisfied with the level of democracy. Banks also rank very highly among the observed institutions. Three-fifths of the respondents express trust in banks, while slightly less than two-fifths do not trust banks. More distrusting than average are respondents living under poorer material conditions or who are not satisfied with their material conditions, etc. Supporters of the political left express above average levels of trust in banks. i) Schools, the Education System. Schools rank second, immediately after family, with four-fifths of the respondents expressing support. Throughout the observation period (1991–2006), schools ranked very highly in terms of trust, which corresponds to the high value attributed to education. Analyses show that there are no fundamental differences between the observed groups in their expression of trust, however, respondents with higher education and supporters of the political right expressed above-average levels of trust. There are higher levels of dissatisfaction than average among respondents with primary school education, the unemployed, respondents with negative views about socialism, respondents who classify themselves as belonging to the lowest social stratum, and respondents who are dissatisfied with their material conditions. Trust in Institutions – Intermediate Comparisons 1991–2006 Table 9.01 (see annex) displays the results of all surveys of trust in institutions taken during the 1991–2006 period.16 (a) The extreme right column of the table displays the average of all measured percentages of trust, which provides a view of the final, overall placement of the observed institutions; (b) the bottom of the table displays the annual average of trust in all (14 selected) 15 Slovenia adopted the Euro in January 2007. 16 For each survey, the sum of the responses “trust very strongly”, “trust considerably” (3 + 4) are given for each institution individually. 238 Niko Toš | Vlado Miheljak | Slavko Kurdija institutions, which indicates the changes in the expression of trust towards all observed institutions during that period. In (a), we have already determined that at the very top of the scale, with the highest total average percentage of trust is family, while at the very bottom, with the lowest percentage by far, lie political parties. In addition to family, schools, educational institutions, the Slovenian Tolar and the Central Bank rank very highly. These are followed by the group of institutions that enjoys an average level of trust of between one-third and one-half. At the top of this group are the president of the republic, the United Nations, banks, the European Union, the Slovenian army, and neighbors, while at the bottom of the group lie NATO, the police, the prime minister, companies, God and the media. The lower group, which enjoys an average level of trust (below 33%), includes the judiciary, government, trade unions, the Church and clergy, the national assembly and political parties. In a sense, this classification is misleading because it blurs the sharp increase and decline in the expression of trust in individual institutions during the observation period. In (b), the sharp increase and decline are illustrated in the column at the edge of the table, which gives the annual average for all observed institutions and individually for the 14 institutions under observation during the entire period. Figure 9.02 illustrates the dynamics and changes in the averages during the 1991–2006 period: it shows higher levels of trust in 1991 (37.4%), the gradual decline to 1993 (28.5%), and the eventual increase, peaking in 2000 (50.7%) and reaffirmed in 2001 (46.1%). The figures for 2002 (41.1%) already indicate the characteristic decline in overall trust; figures for 2003 (45.3%) and 2005 (44.2%)17 again show an increase in the average level of trust to the 2001 level; the 2005–2006 (44.3%) period again shows a decline in the average level of trust in institutions. From the perspective of the development of a democratic institutional system, two distinct periods emerge. The first starts in 1991 with a relatively high percentage (37.4%), falls two years later (1993: 28.5%) and climbs to its initial level of trust in the late 1990s (1998: 37.5%, 1999: 39.5%). The second period starts in 2000 when overall trust in institutions reaches its highest point (50.7%) before falling to its lowest point in 2006 (42.3%), which is still higher than during the first period. Figure 9.02 illustrates the trends in trust in civil and state institutions. With the exception of the starting period (1991–1993), civil institutions are ranked far above state institutions; the final average level of trust in state institutions is almost half of that in civil institutions. 17 In the 2004 general election, the government was replaced after three consecutive mandates 1992–2004. Trust in institutions of the System – european Comparisons – and Slovenia 239 Figure 9.02 Trust in Institutions: Averages for Civil Society and State Institutions, and Overall Average for the 1991–2006 Period (Sum of the responses: “trust completely” + “trust considerably”) 60 50 40 30 20 Civil society institutions Total 10 State institutions 1991/2 1993/1 1995/1 1997/1 1999/4 2001/2 2003/3+4 2006/2 1992/3 1994/4 1996/1 1998/2 2000/1 2002/1 2005/1+2 Source: Public Opinion and Mass Communication Research Centre, Slovenian Public Opinion Survey, 1991–2006 Comparisons between the First, Final Surveys and the Average of All Surveys. Figure 9.03 illustrates the level of overall trust in the observed institutions: the first column shows the level trust from the first survey for each institution, followed by the average level of trust for all the surveys, and finally the third column shows the level of trust during the final survey.18 The Figures shows surveys of trust for selected political, civil and international institutions. It is evident that by the last survey of trust, the initially high level of trust in the president of the republic, the armed forces and the national assembly declined past the lower average, and in the case of political parties, to the lowest level. The opposite tendency, namely, a lower initial level, higher average and higher final level, is characteristic for schools, educational institutions, banks, the central bank, companies, the media, trade unions and the European Union. The third group of institutions, in which a low level of trust is relatively “stable” during both observation periods, includes the Church and clergy, the government, police, etc. These different “models” in the expression of trust in political and civil institutions illustrate the specifics of their belief, however, they are undoubtedly also an expression of the entire course (more or less) of the democratic institutionalization of Slovenian society and its development between 1991 and 2006. The position of economic operators, for example, exhibits an initial 18 The first survey was made in 1991 or later, whilst the final survey was made in 2006. 240 Niko Toš | Vlado Miheljak | Slavko Kurdija commotion within commercial structures and their gradual stabilization (the economy, banks). The gradual but increasing awareness of the population and their recognition of the opportunities provided by establishing links with the European Union, for example, becomes evident. The “model” that covers the majority of the political institutions reveals (higher) initial expectations, which are later not reaffirmed. The transition to a new political system, characterized by political plurality and the introduction of a market economy, brought with it disappointment. This model also conceals within itself the high level of post-independence enthusiasm, which particularly applies to fluctuations in the expression of trust in the armed forces, president of the republic and political parties. The low initial levels of trust in the economy is an expression of the frustrations and fears surrounding existential issues and economic development; the final low level of trust and growing distrust in the institutions of the political system – after 15 years of development – invites a critical view and warning about the nature, situation and result of the democratic political institutionalization of Slovenia. Figure 9.03 Trust in Institutions: First, Final Measurements and Average (Sum of the responses: “trust completely” + “trust considerably”)   ­ € ‚ ƒ „     †ƒ Source: Public Opinion and Mass Communication research Centre, Slovenian Public Opinion Survey, 1991–2006 Trust in institutions of the System – european Comparisons – and Slovenia 241 The different dynamics of trust hide the specifics in the processes of democratic institutionalization in Slovenia. Subjects related to an individual’s intimate environment (family, for example) have not undergone any characteristic changes. However, significant changes, which are significant from the perspective of the final status (2006), have been observed with regard to political institutions. Over the last five years, and particularly the last two, following the change in government and legislature, the process of democratic institutionalization has not taken a turn for the better. Overall trust in institutions has fallen again, and trust in political institutions has fallen dramatically. The change in government did not, therefore, result in new incentives and a “relaxed” atmosphere, as predicted, which would have been expressed by an increasing trust in political institutions. Intermediary Summaries and Conclusions For the most part, studies conducted in the 1990s in Western, Central and Eastern Europe confirm the assumption that, on the basis of cumulative positive experiences, institutions in the original, traditional democracies are valued far more highly than in countries considered to be young democracies. The results of our research places Slovenia at the top of the latter, or at the bottom of traditional democracies. The purpose of our analysis was to point out the characteristics of democratic institutionalization in Slovenia as demonstrated by the expression of trust in the institutions of the system. We have relied on the trends in the expression of trust in state and civil institutions between 1991 and 2006 (Figure 9.02). One can see that the initial level of trust in political institutions immediately after independence (1991) was relatively high – higher than the level of trust in civil institutions. Surveys conducted two years after independence (1992– 1993) show a general decline in the level of trust, defined by the opposing trends in the expression of trust in civil institutions, which increases, and state institutions, which decreases dramatically. From this period forth, we notice a gradual and steady trend of increasing trust in civil institutions and the stagnation (at a low level) with regard to state institutions. A sharp increase in the expression of trust in civil and state institutions is only recorded in 1999–2000. The trend is negative again over the next few years, during which we see a decline in both cases. However, trust in civil institutions remains at a high level, while trust in state institutions falls almost to the level of a decade earlier. Between 2002 and 2006, there is a slight increase in the level of trust (2003), which falls again by the end of the survey period (2006). At this point, the level of trust in state institutions again almost reaches the lowest point 242 Niko Toš | Vlado Miheljak | Slavko Kurdija recorded a decade earlier, while the level of trust in civil institutions declines only slightly, albeit at a high level, and remains higher than the level of trust in state institutions. Such a summary of the dynamics of trust in institutions partially confirms the hypothesis of a low but increasing level of trust in the institutions of the systems in young democracies. It does not, however, confirm the expectation of a more favorable position of state institutions vis- à-vis civil institutions during this period in Slovenia. In order to correctly interpret this general image of democratic institutionalization in the state and society in Slovenia, it would be advantageous to show which state (political) and civil institutions are related to higher or lower levels of trust, thus decisively influencing the general trend. The poor performance of state institutions is mainly the result of the expression of distrust in political parties and the national assembly, while the government, president of the republic, armed forces and police retain a higher level of trust. However, during the 15-year observation period, the latter institutions also became subject to the general dynamics of trust, which includes this negative trend. Three observed state institutions (the national assembly, the armed forces and the president of the republic) initially (1991) enjoyed a very high level of trust. Trust in the national assembly declines sharply during all subsequent observation periods and remains at a low (the lowest) level, while trust in the president of the republic and the armed forces falls to a lower level but retains a relatively high, above-average level in comparison to other political institutions. Within this group, only political parties remain at a low level, without any discernible level of trust. We shall not summarize here the detailed analyses of time dynamics or structural analyses of the expression of trust in individual institutions or groups of institutions. We can, however, confirm the hypotheses put forward at the beginning. The analysis confirms the assumption that the expression of trust in institutions depends on the knowledge, education and degree of political culture of the individual. Categories in which interpersonal trust is more easily achieved also more readily and strongly express trust in institutions. The analysis also confirms the relationship between trust and experience, that the elderly, with more social experience, are more trusting of the institutions of the system, and conform more readily, which also results in a higher level of trust. If it is characteristic of the younger segment of the population that they maintain an inchoate critical distance from political institutions, the older and oldest segments of the population display a tendency to conform to, recognize and accept the given situation, and show a friendlier demeanor towards institutions and the persons associated with them. The analyses also confirm the assumption that the expression of trust reflects ideological beliefs and the Trust in institutions of the System – european Comparisons – and Slovenia 243 political proximity of the respondents to executive authority. It is clear (not only for the 2006 period but for other surveys as well) that supporters of parties in the current ruling coalition express a significantly higher level of trust in state institutions, and vice versa, that supporters of the current opposition parties are more reserved when expressing their trust in political institutions. In both cases, this especially applies to the parliament, government, prime minister, and to a lesser degree to the president of the republic, the armed forces, police and the judiciary. The situation regarding the expression of trust cannot be said to be critical but it would be difficult to term it as satisfactory. When evaluating the general level of trust in civil institutions, especially the educational system and the economic operators under observation, the level of trust may be regarded as satisfactory and favorable. Indeed, civil institutions play an important role in ensuring the internal stability and legitimacy of the social system as a whole. The levels of trust expressed in state and political institutions depart dramatically from this favorable situation. 9.3 Trust in Institutions – European Comparative Studies19 Researching Trust in Institutions in the ESS – Classification of the Problem First conducted in 2002, the European Social Survey seeks, inter alia, to research the processes of democratization in Europe. This is achieved through a series of thematic modes and operations, a part of which is the research of trust in institutions. This has been at the core of the ESS since 2002, i.e. for all three rounds. This theme is investigated through determining trust in national assemblies, legal systems, the police and politics, the European Parliament and the United Nations. Political parties join these six institutions only in the third round (2006). The choice is thus limited to a narrow field of political institutions, which does not provide a good basis for broader comparisons with the older surveys mentioned earlier, which penetrated further into the political subsystem, as well as into the civil subsystem. Nevertheless, within its scope, the ESS provides the opportunity to analyze the processes of democratization. The survey covers most of the EU member states as well as a few non-members. The third round of the survey (ESS, 2006) includes important non-EU countries (Russia and Ukraine). The fact that the survey 19 The European Social Survey 2002, 2004, 2006 244 Niko Toš | Vlado Miheljak | Slavko Kurdija was conducted in three rounds (2002, 2004, 2006) serves as a basis for an intermediate analysis of the course of the processes of democratization in Europe. Surveys conducted thus far only cover a four-year period, which is important for recent European history20 but represents only a brief period in the restructuring of the common democratic political culture. The three rounds of the ESS cover a total of 27 countries, namely, the 25 member states plus Russia and Ukraine. A total of 21 countries participated in the first round, of which 18 were EU member states and three were in membership negotiations (Poland, the Czech Republic and Slovenia). The same countries participated in the second round (ESS, 2004). In the third round, some of previous countries did not participate21 (the Czech Republic, Greece, Italy, and Luxembourg), while a number of countries participated for the first time (two new members: Bulgaria and Slovakia, and Russia and Ukraine). The ESS has gathered a large amount of data about trust in political institutions. It provides a wide stage for an analysis of the state of democratization in individual countries, a comparative analysis of the state of democratization in all countries under observation, for uncovering the changes that took place during the period under observation (2002–2006), and – at the level of state records – recognizing the patterns of their mutual similarities and differences. We shall attempt to answer the following questions within the scope of this paper: – how is trust expressed in the six political institutions under observation (within individual countries, in all countries under observation during a specific time period, and during the three periods, individually and overall); – how is trust expressed in an individual institution, what are the trust profiles in individual countries, groups of countries and overall for all countries under observation; is there any significant departure in the trust profiles and the level of the trust in individual institutions or groups of institutions; – are we able to confirm the inferred similarities and differences between individual countries or groups of countries regarding the expression of trust in institutions; what is the result of the analysis and how can it be explained, etc. 20 During this period several Central and Eastern European countries joined the EU. 21 Primarily as a result of difficulty in securing funding for national empirical surveys. Trust in institutions of the System – european Comparisons – and Slovenia 245 The analysis will be presented in several steps. The first step is descriptive and presents the distribution of the measurement of trust.22 In addition to reviewing the frequency of distribution, we also review each country, institution and round individually, and review each vis-à-vis all countries, institutions and rounds – which we present using mean values (Table 9.02 – in annex). From this basis, we make summaries and aggregates of the results by institution in all three rounds, and of the results of the overall expression of trust in institutions in all countries (Figure 9.04). We then present the expression of trust in all institutions under observation by country, for 17 and 27 countries, depending on whether three, two or one survey was conducted in an individual country. This serves as the basis for the creation of a scale for ranking the countries, as shown in Figure 9.05. Figure 9.04 Trust in Institutions by Countries Participate in all Three Waves (Total 3 wave Mean for 6 Institutions, scale 0–10); ESS-European Social Survey 7 6 5 4 3 2 1 UN Police Total Mean Parliament Politicians Legal System EU Parliament 2002 2004 2006 Source: eSS: 2002–2004–2006 We continue by analyzing intermediate aspects of the expression of trust in the six institutions by country. The results of this analysis are presented in Figures 9.06 to 9.08. All the above is based on the calculation of mean 22 Measurements were based on an 11-grade scale (0=complete distrust, 10=high level of trust). 246 Niko Toš | Vlado Miheljak | Slavko Kurdija values and the differences between them. Thus, we also illustrate the basic tendencies for the three rounds, which can perhaps be seen as an upward or downward trend in the expression of trust. In the second step of the analysis, we factor the results of measurements by country and round, and overall. One can see that the expression of trust is consistent for countries both in individual rounds and in all the rounds overall, as well as for all levels overall, and that it is expressed in one dominant factor. The only departure from this finding are the results for Bulgaria and Ukraine, where there are two factors, the first for four national institutions, and the second for two international institutions. This suggests that these two countries occupy a common position within the group of countries under observation, which is separate from the position of the other countries. This is further investigated in the analysis. In the third step, we use the hierarchical clustering method to analyze the relationship between groupings of countries on two levels, first for the 17 countries that participated in all three rounds, then for all 27 countries participating in the three rounds. We thus come to the basic configuration of similarities and differences between the countries. The position of the traditional democracies (old EU member states), the post-transition countries (new EU member states), and Russia and Ukraine will be shown. Considering the state of democratization and the level of prosperity,23 we may expect some disparities by individual countries and the reclassification of individual western countries. We first describe the results of the hierarchical cluster method for the 17 countries that participated in all three rounds, then for each round individually. Separately, we present the results that include the countries that only participated in one or two rounds. This includes some “old” (Italy, Greece, Luxembourg) and “new” EU member states (Bulgaria, the Czech Republic, Cyprus, Estonia and Slovakia), as well as the non-members Russia and Ukraine. The results of these analyses are shown in Figures 9.01 to 9.03 and 9.04. Expression of Trust – in Six Political Institutions The expression of trust in six institutions, measured in 17 countries over three rounds is illustrated in Figure 9.04. One can see a common European profile of trust in institutions (over three rounds), with the police in first place (6.21),24 the UN in second place (5.47), followed by national parliaments 23 Based on GDP as well as the ranking of the country in the Human Development Index. 24 Mean values are based on a scale of 0 to 10, where 10 is the highest level of trust. Trust in institutions of the System – european Comparisons – and Slovenia 247 (4.80), the European Parliament (4.58) and politicians (3.83). The hierarchy of the six institutions under observation is strict and distinct, and departure from this profile by individual countries is exceptional. Such a ranking begs the question, how is democratic institutionalization taking place in Europe if the police are the most trusted institution while national parliaments rank significantly lower. Further, what is the significance of the fact that the UN enjoys a higher level of trust than national parliaments and national legal frameworks? That the European Union is still being established is expressed in the low level of trust in the European Parliament. That the low level of trust in politicians is a pan-European phenomenon is also validated. Table 9.02. shows how trust is formed and expressed in individual institutions by country. Only Russia, Ukraine and Bulgaria show a significant departure from the general definition model. This is not only expressed in the low (or lowest) overal level of trust, but also in the low (or lowest) level of trust in certain individual institutions, which are accorded a high (or highest) level of trust the other countries. For example: in Ukraine, the level of trust in the police, legal system and parliament is far below average, which also applies to Bulgaria and Russia, though to a lesser degree. A second characteristic il ustrated by the same group of countries is the higher level of trust expressed in the EU and UN. In Bulgaria, these two international institutions are, on average, subject to double the level of trust as the domestic institutions. In Russia, the European Parliament enjoys a higher level of trust than the Russian parliament, the Russian legal system and particularly the Russian police and politicians. We also mention a number of other details but remain limited mainly to the results of the third round (Table 9.02, ESS 2006). For example, the expression of trust in the police during the period under observation (2002: 6.21; 2004: 6.06; 2006: 5.74) is on the decline; however, it maintains its primacy in intermediate periods and individual rounds: the police enjoy the highest level of trust. A review of the results of the third round (2006) shows that the police enjoy the highest level of trust in Finland (8.05); this is the highest level of trust expressed in any of the six institutions under observation in any of the countries. This high ranking in Finland is not coincidental and is validated by a similarly high ranking in previous surveys (2002: 7.95; 2004: 7.96),25 as well as by the high overall level of trust in Finland in all political institutions under observation. Ukraine stands out as having the lowest level of trust in the police, or more precisely, the highest level of distrust (2.61). This is a significant departure 25 We shall not at this point discuss what contributes to the high level of trust in the police in Finland. 248 Niko Toš | Vlado Miheljak | Slavko Kurdija from the overall average for the police (by a factor greater than two), but remains at the average level of distrust for all institutions under observation in Ukraine (2.80). The highest level of distrust in Ukraine is expressed for the national parliament (2.32), the legal system (2.45) and politicians (2.04), but this general distrust is partially allayed by the expression of trust in the European Parliament (3.89) and the UN (3.75).26 As mentioned earlier, politicians enjoy the lowest level of trust (3.58). This distrust is expressed primarily in Bulgaria (1.72), Ukraine (2.04), Russia (2.51) and Hungary (2.53). Three surveys conducted in Hungary show that between 2002 and 2004, there is a sharp decline in the expression of trust in politicians (2002: 3.88; 2004: 2.68; 2006: 2.53). The European Parliament, which ranks second to last among the institutions under observation (4.63) in individual countries, enjoys an above average level of trust. This, for example, applies to Cyprus (5.84), as well as Slovakia (5.14), Bulgaria (4.76) and even Ireland, Portugal and Russia, where it is highly ranked, reaching second or third place. In Russia, for example, the European Parliament ranks much higher than the Russian parliament or legal system in terms of trust. A similar situation is that of the UN, which ranks second overall among all six institutions under observation. It scores low in Cyprus27 and Austria but high and above average in Finland, Hungary and Norway. Trust in Institutions – by Country We shall first show the ranking of the countries by the overall expression of trust over the three rounds. Figure 9.05 shows the ranking of 21 countries by overall expression of trust, which is at the axis of the Figure.28 Deviations from this average are also shown. The four Scandinavian countries are at the top of the ranking, followed by Switzerland and Luxembourg. In the middle, characterized by smaller positive or negative deviations, are Italy, the Netherlands, Greece, Hungary, Ireland, Germany and Austria. Falling below the average are the remaining “old” member states: Belgium, the United Kingdom, France and Spain. At the bottom of the ranking are Portugal, the Czech Republic, Slovenia and Poland. The second survey (2004 – Figure 9.06) only serves to bring this ranking into focus. Denmark and Finland rank highest, while Poland, the Czech Republic, 26 Which remains characteristically below average. 27 This is despite, or perhaps due to, the direct experience of the UN intervention and prolonged conflict resolution. 28 The 0 axis represents the average for the 17 countries that participated in all three rounds. Trust in institutions of the System – european Comparisons – and Slovenia 249 Figure 9.05 Trust in Six Institutions, Summarized by Countries (Mean – scale 0–10; zero point = mean on 17 Countries); Wave 1 – ESS 2002      ­  € ‚ ƒ  „€ ‚ Portugal and Slovenia rank lowest. Conspicuously, Italy and Hungary characteristically fall below the average level they had in the first round. Also characteristic is the increase in trust in Poland, the Czech Republic, etc. In the second round, a much higher level of overall trust is measured in Denmark, Belgium and Finland. The third round of measurements (2006 – Figure 9.07) confirms that citizens in the Scandinavian countries (Denmark, Finland, Norway, Sweden and the Netherlands) as well as Switzerland express an above average level of trust in politicians. Also, trust in political institutions in Ukraine, Bulgaria and Russia shifts to strong distrust. Those countries that initially ranked higher in 2004 show a tendency of increasing trust, while those that were ranked lower show a tendency of declining trust or increasing distrust. This not only applies to the post-transitional group of countries, but to the lower-ranking (according to the scale) “old” member states of the EU. By including additional groups of 250 Niko Toš | Vlado Miheljak | Slavko Kurdija countries in the survey in the last round (2006 – Figure 9.07), the differences become even more apparent. Figure 9.06 Trust in Six Institutions, Sumarized by Countries (Mean – scale 0–10; zero point = mean on 17 Countries); Wave 2 – ESS 2004 €     ­   ‚ ƒ  „€ ‚ Finally, Figure 9.08 shows the overall ranking of 27 countries on the basis of whether they participated in one, two or three rounds of the survey.29 High in the ranking, with the highest measured overall level of trust, are the Scandinavian countries (Denmark, Finland, Norway, Sweden, and also Luxembourg and Switzerland). At the bottom of the ranking, with the lowest measured level of trust or highest measured level of distrust are mostly the post-transition countries, including Ukraine, Bulgaria, Russia, etc. The 17 countries that participated in all three rounds are ranked higher, while the countries that only later joined the survey, primarily Russia, Bulgaria and Ukraine, are ranked at the bottom. 29 As before, the 0 axis represents the average for the 17 countries that participated in all three rounds. Trust in institutions of the System – european Comparisons – and Slovenia 251 The range between the highest and lowest level of trust is significant (6.47 – 2.80). Whereas the top ranks are occupied by the traditional democracies, which are characterized by a higher level of prosperity, the bottom ranks are occupied by the post-transition countries where the level of prosperity is lower. Figure 9.07 Trust in Six Institutions, Sumarized by Countries (Mean – scale 0–10; zero point = mean on 17 Countries); Wave 3 – ESS 2006 ‚ ƒ „  ­€       Considering the exceptionally high level of distrust in political institutions in some of these countries, it can be said that the legitimacy of the democratic institutions is under threat. If the top ranks are occupied by countries from Western Europe, primarily from Scandinavia, then the bottom is occupied by countries from Eastern Europe, which are bound by a common history of the Eastern Bloc, albeit each has followed its own destiny and all are still undergoing the transition process to a greater or lesser degree. In addition to the three countries mentioned above, this group at the bottom of the 252 Niko Toš | Vlado Miheljak | Slavko Kurdija ranking includes the other “new” member states of the EU, namely, Poland, the Czech Republic, Slovenia and Slovakia, as well as Portugal. Slovenia and Portugal do not share a common history with the other countries of this group, but the transition processes they underwent are similar.30 In the middle, between the topmost and bottommost ranks, are all “Western European” countries participating in the survey. This includes (according to the expressed level of trust) Estonia, at average level, and Hungary, which characteristically falls below average. Below them are ranked the United Kingdom and France, as well as Germany, Austria and Spain. A comparison of ranking (according to the level of trust) in the three rounds validates the conjecture that increasing or declining trust can be detected during the individual periods, which, however, is not illustrated in Figure 9.07. A comparison of the results of the surveys enables us to rank the countries in which the expression of trust has been increasing (Belgium, Denmark, also Spain and Finland) or declining (the United Kingdom, Hungary) between the first and third rounds. In most countries, however, there was a decline in trust between the first and second rounds, and an increase in trust between the second and third rounds (Austria, Sweden, Germany, Poland, Portugal, Slovenia). The opposite model is that in which an increase is followed by a characteristic decline in the third round (Ireland). These four models also apply to countries that participated in only two rounds. A declining trend was observed in the Czech Republic, Greece and Italy, while an increasing trend was observed in Luxembourg. When determining the increase or decline in trust over two or three periods, we must take into consideration that change can take place at either a high or low measured level of overall trust, as can be seen in the ranking of the countries (Figure 9.08). 30 In Slovenia, this followed the collapse of the Yugoslav system of self-management and the Yugoslav federation; in Portugal, it followed the establishment of a democratic system 15 years earlier, however, its economy significantly lagged behind the other “old” member states of the EU. Trust in institutions of the System – european Comparisons – and Slovenia 253 Figure 9.08 Trust in Institutions, by all 27 Participating Countries (Total Mean for Six Institutions, scale 0–10); ESS – European Social Survey ‡ˆ‰  Š  € Š „ †     ­ € ‚ € ƒ‰€ ‚ ƒ  „ € € Source: eSS: 2002 – 2004 – 2006 Trust in Institutions and “Satisfaction” Analyses of the relationship between trust in political institutions, “trust in people” and “satisfaction” with conditions in the system help to explain the cause of these phenomena. On the basis of the results of the ESS 2006, we looked into the relationship between the overall expression of trust in seven political institutions31 and a group of intervening variables: trust in people, satisfaction with life, satisfaction with economic conditions in the 31 Including political parties. 254 Niko Toš | Vlado Miheljak | Slavko Kurdija country, satisfaction with the functioning of the government, and satisfaction with the functioning of democracy. The survey, conducted through a sample taken from all the countries combined, confirms the relationship between the level of trust in the institutions of the system and the above-mentioned intervening variables. They may, therefore, be referred to as predictors of the expression of trust. There is a high level of correlation. One can see that the political dimension holds primacy over the economic dimension. This is most apparent in the close relationship between trust in institutions and satisfaction with democracy.32 All other surveys confirmed the same, albeit with a lower intensity. This also applies to the survey on trust in institutions vis-à-vis the expression of trust in people.33 Figures 9.09 and 9.10 show how these relationships are manifested in individual countries. Figure 9.09 Trust in Seven Institutions by Satisfaction with Democracy The relationship between the index of trust in institutions and the “predictors” is also confirmed at the level of national records. However, there are characteristic differences in the level of trust in institutions and the predictors between countries. Figure 9.09 illustrates the diagonal relationship between the measurement of the variables and the corresponding ranking 32 (Pearsons corr. c. 0.61; sig. 0.000). (corr. c. 0.329; sig. 0.000). 33 Corr. c. 0.329; sig. 0.000. Trust in institutions of the System – european Comparisons – and Slovenia 255 of the country. Scandinavian countries, as well as Switzerland and Cyprus, which rank highly on the index of trust in institutions, are characterized by a high level of satisfaction in democracy. In contrast, countries with a low(er) level of trust in institutions are also characterized by a low(er) level of satisfaction with democracy. Slovenia, Poland, Portugal, Russia and Hungary are to be found at the bottom left of the figure, below the average level. In the middle part of the figure, hovering at about or slightly above average, are the remaining Western and Eastern European countries. The figure shows the position of Estonia and Slovakia at the average level together with the United Kingdom and France. At the bottom of the expression of satisfaction with democracy are Ukraine and Bulgaria. Figure 9.10 Trust in Seven Institutions by Satisfaction with Life in General The survey on the relationship between the index of trust with “satisfaction with the government” also confirms expectations, only that there is a greater dispersion with regard to conditions in individual countries. The trend, however, is the same as above (see Figure 9.09): at the bottom are Bulgaria and Ukraine; in the middle are Poland, Hungary and Portugal. Slovakia, Estonia, Slovenia and Russia34 stand out as expressing a higher level of satisfaction in the government; at the very top is the group of Scandinavian countries. 34 2006: the Putin era. 256 Niko Toš | Vlado Miheljak | Slavko Kurdija Figure 9.10 shows the ranking of the countries with regard to trust in institutions and general satisfaction with life. The diagonal classification of countries remains the same, with only minor deviations. Again, Bulgaria and Ukraine are at the bottom of the group while the Scandinavian countries and Switzerland are at the top right. Two diametrical y opposed groups emerge: Denmark records the highest level of trust in institutions and satisfaction with living conditions; Ukraine and Bulgaria are characterized by the lowest level of trust in institutions and satisfaction with life. Portugal, Hungary, Poland and Russia are stil at the bottom (left) of the scale; Slovenia and Estonia rank in the middle. The survey on satisfaction with economic conditions confirms the rectilinear relationship. The variables included in the analysis help to explain the origin of trust in political institutions; clearly, we cannot exclude the possibility of an opposite (reverse) effect. Survey of the Relationship between Countries – Characteristic Groups The final step in analyzing the survey results is based on the hierarchical clustering method.35 The descriptive analysis presented above provides an indication as to how and where individual countries rank with regard to the measured overall level of trust both for each round as well as for all of the three rounds. However, this only provides a superficial image; it can be further analyzed by ranking groups that comprise all individual measurements for each institution at every round and for each country. This means that the ranking not only shows the hierarchy according to the “quantity” of overall trust expressed, but that it also presents all the specific profiles that emerge from the measurement of trust in individual institutions for each round. In such a “quantitative” hierarchy, it would seem that Estonia and Hungary, typical post-transition countries, rank quite highly with the more developed countries in the individual rounds. They are, however, typical post-transition countries according to their political and economic status. 35 In the hierarchical clustering method, the calculation of distances between units creates groups or clusters. The graphically presented dendogram shows the process of separating the units or groups of units. We can observe the ranking of the units at each step or level. The proximity of the units, or more precisely, the length of the line connecting the units, shows the distance (difference or similarity) between the units. Units connected in the first step are most similar to each other, while those connected by the longest arch differ the most. The basis for our calculation of the distances was the square of the Euclidian distance, while the connections are based on the Ward method. The classification criteria are the six above-mentioned criteria – assessment of trust in parliament, the legal system, police, politicians, the EU Parliament and the UN. The classification is thus based on similarities in the assessments for trust in the said institutions in the individual countries. Trust in institutions of the System – european Comparisons – and Slovenia 257 We used the hierarchical clustering method on the three groups of data acquired in the three rounds of the survey. The analysis only covers the 17 countries that participated in all three rounds. The reason for this limitation is that the results of analyses from all three time periods need to be compared and the trends identified. Figure 9.09 shows the results of the hierarchical clustering method from the first round [ESS, 2002]. At the primary level are five major groups of countries (S1–S5). The most obvious relationships are those between: S1 Portugal, Slovenia and Poland; S2 Spain, France and Belgium: S3 Switzerland and the Netherlands; S4 Germany, Italy, Austria and the United Kingdom; and S5 Norway, Sweden Denmark and Finland. The second group (S2) also includes Hungary. On a higher relationship level, S3 and S4 combine to form a new unit that includes Switzerland, the Netherlands, Germany, Italy, Austria and the United Kingdom, and S1 and S2 combine to form a new unit that includes Portugal, Slovenia, Poland, Spain, France, Belgium and Hungary. Figure 9.11 Trust in Institutions; 17 Countries (ESS – European Social Survey) Wave 1 – ESS 2002      After the second survey [ESS, 2004], the hierarchical clustering method reveals four distinct groupings within the set of 17 countries: the first (S1) includes Belgium, Spain, Ireland and the Netherlands. The second (S2) includes 258 Niko Toš | Vlado Miheljak | Slavko Kurdija Austria, Germany, France and the United Kingdom. The third group (S3) includes Portugal, Slovenia, Hungary and Poland, while the fourth group (S4) includes Denmark, Finland, Norway, Sweden and Switzerland. It is apparent from this classification that the Central European, post-transition countries (Poland, Hungary, Slovenia) and Portugal fall into one group regarding the expression of trust in political institutions. An analysis of results of the third round [ESS, 2006] shows an even more distinct clustering. The first (S1: Belgium, Ireland, Spain) and second groups (S2: Austria, Germany, France, the United Kingdom) are related at the highest level. The third group (S3) includes the Central European, post-transition countries (Hungary, Slovenia, Poland – and Portugal). Finally, the fourth group (S4) includes the Scandinavian countries (Denmark, Finland, Norway, Sweden, the Netherlands) and Switzerland. This profile is similar to that of the second round. Figure 9.12 Trust in Institutions; 17 Countries (ESS – European Social Survey) Wave 2 – ESS 2004      Finally, we have applied the hierarchical clustering method to all the countries that participated in the survey (2002, 20004, 2006), regardless of the number of rounds. Figure 9.12 shows the classification for the 27 countries. The profile reveals four distinct groups. The first group (S1) includes all the Scandinavian countries as well as Switzerland and Luxembourg. These are Trust in institutions of the System – european Comparisons – and Slovenia 259 countries with a strong tradition of democracy. The second group (S2) includes Ireland, Italy, Belgium, Spain, Austria, Germany, France and the United Kingdom, as well as Cyprus and Greece. These are developed, Western European and Mediterranean countries with differences in the conditions and course of democratization. In a group of their own (S3) are Bulgaria and Ukraine. Finally, the fourth group (S4) comprises all post-transition countries (Slovenia, Slovakia, Estonia, Hungary, the Czech Republic, Poland, Russia) and Portugal. Figure 9.13 Trust in Institutions; 17 Countries (ESS – European Social Survey) Wave 3 – ESS 2006      The countries that fal in the first group (S1) are characterized by a high level of expressed overal trust and smal deviations from the overal high level of trust in institutions. It is characteristic of these countries that the expression of trust in individual institutions never transforms into a prevalent distrust. The same applies to the second group (S2), however, at a lower level of intensity and with more distinct fluctuations (including downward) between the three observation periods. The third group (S3) is characterized by a general low level of overal trust in political institutions and a low level of trust in al domestic institutions, which is partial y al ayed by a higher level of trust in the EU and the UN. Final y, the fourth group (S4) includes all Central European, post-transition countries, as wel as Russia and Portugal. In terms of the expression of trust, Estonia (single survey) and Russia 260 Niko Toš | Vlado Miheljak | Slavko Kurdija (single survey) represent the top and bottom levels, respectively, of trust in this group. Between them are countries characterized by a significant decline in the level of trust from one round to the next (Hungary, the Czech Republic), as wel as countries with significant fluctuations (increase, decrease) in the expression of trust ( e.g. Poland and Slovenia). The group comprises countries with different levels of material prosperity and includes Portugal as a “Western” country with a long tradition of democracy but a relatively low level of economic development. Figure 9.14 Trust in Institutions, 27 Participating Countries in Three Waves (ESS 2002 – 2004 – 2006) Merged Data Set (ESS – European Social Survey)  ­ €  ‚   ƒ  €„­   „ Needless to say, the il ustration of the characteristics of the countries in the four groups provides only a superficial explanation of the causes. Among the circumstances that influence the development of a democratic culture, trust Trust in institutions of the System – european Comparisons – and Slovenia 261 in the democratic system as a whole and trust in individual institutions are the level of general development of the society, economic development and the development of social services that serve the needs of the population and guarantee their quality of life. It is sufficient to point out the relationship between the grouping of the countries on the basis of the hierarchical clustering method and their rank in the “human development index”36 for the purpose of explaining these aspects. A comparison between the classification of individual groups with the classification of the participating countries in the human development index shows that the first group (S1), which can be referred to as the “Viking” group, ranks highest on the human development index (second to 18th place); this is fol owed by group S2, namely, the “Western” and “Southern” European group, which has a high ranking in the human development index (fifth to 28th place). Next is the post-transition group (S4), which is ranked in the middle and bottom of highly developed nations (27th to 67th place), and final y group S3 comprising the two countries that are characterized by transitional problems and are ranked lowest in the human development index, at the transition between highly developed and developing countries (53rd to 76th place). Obviously, on this basis, we cannot come to a firm conclusion about the relationship between the expression of trust in political institutions and the quality of individual and social life. We have less of a basis stil in determining the relationship between the causes. However, it is apparent that the survey and analysis contribute to explaining the phenomenon. The stepwise hierarchical clustering method after the individual rounds and the final overal survey of the relationship between al countries in al rounds conceals some of the first impressions from the classification (Figure 9.08). Primarily the first impression that, in terms of the expression of trust, Estonia is closer to Belgium and Germany and Hungary is closer to Spain and the United Kingdom than to their post-transition partners (Poland, Slovakia, the Czech Republic, Slovenia), which is unexpected. We would sooner have expected this of Slovenia, which after 1990 was considered more highly developed and which went through the transition processes with fewer problems. We would have expected Slovenia to be in the same group as Portugal and Greece. Such expectations are bolstered by the results of surveys conducted in the 1990s. Nevertheless, Slovenia ranks among the post-transition countries, together with Poland, the Czech Republic, Hungary, Slovakia and Estonia. Analyses should not, therefore, be based solely on quantitative indicators (level of overall trust), but on the profiles of distrust in the participating countries. Likewise, Switzerland and Luxembourg stand out by fal ing into the “Viking” block. 36 Human Development Index 2007/2008. 262 Niko Toš | Vlado Miheljak | Slavko Kurdija Switzerland, therefore, is closer to the Scandinavian countries and fal s under a group that is characteristical y different from other Western European countries. 9.4 Conclusions We determine in the introduction that the trust of the people in the system forms the basis for democratic institutionalization. This is pointed out by Almond and Verba, who have also investigated the relationship between the expression of trust and the legitimacy of the system. Lipset and Schneider suggest that gaps in the level of trust in the most important political institutions result from increasing criticism of the legislative and executive branches of state, and particularly of political parties. Initially, we presume that the traditional democracies will display a high level of trust in institutions and that the new democracies will show a gradual rise in the level of trust. We also assume that cultural, economic and political conditions will be significant in the development of democratic institutions, and therefore, expressions of trust will manifest differently in each particular system. We present our analysis in two parts. The first is based on surveys of trust from Western and Southern Europe, some from Central and Eastern Europe, and particularly those from Slovenia. We present the trends in the expression of trust and rank the countries accordingly; the results of the survey in Slovenia are analyzed in detail regarding trust in institutions, the political and civil subsystem, and a detailed account of the timeline of the expression of trust between 1992 and 2006 is given. A longitudinal view of the expression of trust in general and in Slovenia in particular reveals oscillations, which result from both people’s past experiences and also their high expectations for improvement. In Slovenia, for example, the “subject” of trust is the renovation of the entire system. It becomes apparent that the expression of trust depends increasingly on people’s expectations of individual institutions. The fewer expectations fulfilled, the greater the dissatisfaction; this is expressed as a decline in the level of trust in individual institutions of the system. The analysis in the second part emonstrates that trust in institutions is not expressed when expectations are not met. Easton has determined that the level of support or legitimacy of the political community, expressed as a sense of community, is formed through trust in political institutions, in civil institutions and in other people. The second part of the analysis also shows that people’s expectations are left increasingly unfulfilled. An investigation in the 1990s asked what level of “trust” is required for a democratic society to function normally and efficiently. It is Trust in institutions of the System – european Comparisons – and Slovenia 263 difficult to determine the level of trust necessary for an individual society to be deemed democratic. When, then, is the democratic deficit cancelled? In Slovenia, where distrust does not yet undermine the foundations of the democratic political system, this deficit is most clearly seen in the trust in the legal and judiciary systems, but it also applies to the executive, parliament and especially to political parties. Intermediate analyses show that, following an initial wave of enthusiasm after the introduction of the democratic system (1990–91), the zeal began to decline, not only for political parties but for parliament and the government as well. We also demonstrate that trust in the civil society has increased during the same period. Only trust in the Church, clergy and political institutions declines or remains at the initial low levels. An analysis of the data from Slovenia37 confirms a series of assumptions on social and structural factors that influence the expression of trust. Thus, the correlation between the conditions of primary socialization, age, education and philosophical and political orientation is confirmed. Older people, for example, express a higher level of trust in political institutions, which is not only the result of their greater experience, but also of their greater tendency to conform. Generally, the better educated and more politically active express a higher level of trust in institutions. The expression of trust is also linked to “affiliation” with those in power; restraint in the expression of trust is characteristic of the opposition. All this is in accordance with expectations. The second part of our analysis, based on the results of the European Social Survey, focuses on the attitudes towards six or seven institutions, in three consecutive surveys,38 for several participating countries. In addition to countries in the European Union, the survey includes Russia and Ukraine. One can see that of al the institutions included in the survey, the police rank highest, followed by the UN, the legal system, national parliaments, the European Parliament and, at the bottom, politicians and political parties. We describe in detail the differences in the expression of trust between countries. At the top of the scale are the Scandinavian countries, in the middle are the transition countries, and at the bottom, Bulgaria and Ukraine. It is apparent that the expression of trust in political institutions is a function of a developed democracy – and vice-versa. This trust in institutions reflects a developed democratic culture. This particularly applies to institutions of the political system. It is difficult to speak of a democratic system in countries where trust in political institutions has failed to develop or where there is a stronger expression of distrust. Again, we ask what level of trust in the system is required before we can speak of it 37 The analysis is based primarily on the Slovenian Public Opinion Survey 2006. 38 ESS 2002, 2004, 2006; participation of 17 to 23 countries, a total of 27 countries. 264 Niko Toš | Vlado Miheljak | Slavko Kurdija as being democratic. The analysis, presented in several steps, always shows the same ranking of the countries, based on the level of democratic culture and the democratic system. The question is whether we may consider countries in which there is a high level of distrust in parliament and the legal system, as wel as the police and political parties, as democratic states. This especial y applies to Bulgaria and Ukraine. It is, however, not limited to these two countries – there are several other countries also in the early stages of democratic transition that can be considered pre-democratic. Our analysis, however, do not include these latter countries. We find both similarities and large differences between countries regarding the expression of trust in individual institutions and the group of institutions as a whole. Whereas there is a strong emphasis on national institutions in one cluster of countries, a second also regards international institutions (the EU, UN) very highly, while a third only shows a high level of trust in the international institutions. By looking at the first rankings and especial y the clustering of countries into groups we could see the relationship between the ranking of countries and their place in the Human Development Index. This confirms the assumption that the democratic climate in a society – namely, trust in institutions – develops hand-in-hand with the development of democratic and economic conditions. The cluster analysis shows that countries that share similar characteristics are ranked together. Two large groups of countries emerge. In one group are the developed European countries with a long tradition of democracy; while in the other the group are the post-transition countries and Portugal. Despite mutual differences, countries in both cases show a high level of internal similarity, some with a developed democratic nature, others with a tendency towards it. We can, once again, confirm that trust in the system is the core of democratic institutionalization. Distrust (which is more-or-less characteristic for post-transition countries, especially Bulgaria, Russia and Ukraine) is not only an expression of unfulfilled expectations, but also of the disharmony between the prevailing individual and collective values, between institutions as the agents of collective values and the agents of institutional power. If the centres of power are not intimately linked to the values of the institutions, we can expect an unstable social construct, greater uncertainty and a decline in social activity. A positive response to such a situation would be to initiate the process of re-institutionalization and to replace the elite that occupies institutional roles. These processes are placid in developed democracies and more extreme in undeveloped democracies. Trust in institutions of the System – european Comparisons – and Slovenia 265 TABle 9.01 Results of Surveys of Trust in Institutions During 1991–2006 (Sum of the Responses “Trust Very Strongly” and “Trust Considerably” – in %) 1991/2 1992/3 1993/1 1994/4 1995/1 1996/1 1997/1 1998/2 1999/4 2000/1 2001/2 2002/1 2006/2 2003 /3+4 2005 /1+2 verage A 1991-2006 god 30,3 28,1 26,0 29,2 32,1 29,3 – 33,5 35,3 44,0 41,8 33,8 38,9 42,2 39,0 34,5 Neighbours 38,0 38,9 33,9 38,8 37,2 35,2 28,9 45,4 47,0 50,3 48,6 43,2 53,0 52,2 – 42,2 Family and relations 87,9 88,9 87,8 89,4 89,7 86,9 86,7 89,0 92,0 93,6 91,0 88,1 94,1 94,3 94,6 90,3 Church and priest 25,7 20,3 17,4 19,4 21,1 22,2 17,4 21,5 23,6 30,1 26,8 21,4 23,8 25,2 23,3 22,6 Television, newspapers, radio 25,5 23,1 19,0 23,7 25,7 28,2 31,8 34,7 43,7 52,2 46,1 36,1 45,1 38,6 35,9 34,0 educational institutions 52,7 56,7 56,7 67,4 71,6 68,8 69,8 76,1 73,0 82,7 80,3 77,2 82,4 77,2 80,3 71,5 Trade unions 16,2 13,1 11,5 12,0 14,5 14,2 12,8 29,0 29,2 39,9 36,3 40,6 42,6 44,4 39,4 26,4 Political parties 11,3 7,7 3,2 4,7 4,5 4,4 6,4 4,3 6,0 13,6 9,4 6,3 8,9 11,4 7,6 7,3 National Assembly 33,6 20,1 15,0 15,0 10,1 11,2 13,0 9,1 11,7 23,6 17,7 14,6 21,5 20,7 16,8 16,9 President of the republic1 67,8 59,1 46,6 45,3 36,3 40,2 41,7 44,0 44,8 59,2 55,6 45,9 41,0 41,0 37,3 47,1 government 32,5 34,4 13,6 14,3 27,9 28,7 27,6 26,3 29,7 43,9 41,3 29,7 34,9 35,0 29,9 30,0 Prime Minister – – – – 32,4 34,7 30,9 33,0 36,1 48,3 48,6 36,2 37,4 35,0 31,1 36,7 Armed Forces 69,0 59,8 41,8 48,7 29,0 33,1 31,6 35,2 39,3 53,3 45,9 38,9 55,4 47,8 38,0 44,5 Judiciary 34,7 31,4 24,5 26,2 25,7 24,3 29,6 33,2 31,6 45,3 41,7 35,6 34,4 32,6 37,5 32,6 Police 37,7 31,2 20,7 27,9 28,3 34,0 33,5 36,8 37,3 53,1 46,9 40,0 44,8 40,8 41,4 37,0 Banks 11,6 21,4 27,7 35,2 40,2 44,2 46,4 52,7 54,6 65,5 65,5 58,0 54,9 59,1 59,4 46,4 Economy, companies 16,8 15,6 13,8 25,8 28,8 28,8 26,8 32,8 37,1 53,2 45,3 42,4 50,0 50,8 51,0 34,6 Slovenian tolar – – – – 55,2 53,6 52,6 63,3 61,8 69,2 69,8 64,6 64,4 64,2 62,8 62,0 Central Bank – – – – 45,8 48,8 50,4 59,6 61,5 68,7 68,2 60,0 57,9 61,9 – 58,3 NATO – – – – – – 28,2 – 42,8 44,7 38,4 31,5 40,1 42,5 34,9 37,9 eu – – – – – – 33,4 – 47,6 41,0 41,8 42,2 55,8 55,2 52,8 46,2 uN – – – – – 37,2 – 48,1 51,0 49,9 41,6 52,3 51,3 44,1 46,9 Total average all 37,0 34,4 28,7 32,7 34,5 35,3 35,1 40,0 42,4 51,2 48,0 42,2 47,0 46,5 42,9 Totall 14 institutions 37,4 34,5 28,5 32,5 32,4 33,5 33,9 37,5 39,5 50,7 46,4 41,1 45,3 44,2 42,3 Source: S POS 1991–2006; Institute for Social Sciences, FSS, University of Ljubljana 1 1991/2: Presidency of the republic of Slovenia 266 Niko Toš | Vlado Miheljak | Slavko Kurdija TABle 9.02 Trust in Institutions by Countries in Three Waves: 2002– 2004–2006 (Mean Scale 0–10); ESS – European Social Survey Mean (total) 3 waves 4,86 4,99 5,62 4,87 6,47 4,85 6,20 4,59 4,63 4,66 5,18 5,29 5,88 3,95 4,18 5,59 4,30 5,08 Mean (total) 1 4,84 5,12 5,65 4,78 6,52 4,97 6,26 4,62 4,54 4,31 5,1 5,45 5,92 3,96 4,21 5,64 4,46 5,07 5,08 UN 4,70 5,28 5,38 4,74 6,51 5,07 6,48 5,12 5,03 5,45 5,71 5,45 6,77 5,33 4,92 6,41 5,07 5,45 5,47 EU- Parliament 3,97 5,15 4,76 4,07 4,96 5,03 4,99 4,37 3,49 4,96 5,26 4,78 4,74 4,80 4,55 4,49 4,98 4,61 4,58 Politicians 3,26 4,36 4,94 3,26 5,61 3,50 4,95 3,29 3,41 2,53 3,82 5,04 4,43 2,10 2,51 4,46 3,21 3,77 3,83 Police 6,30 5,94 6,93 6,63 7,83 6,04 8,05 5,71 6,00 5,12 6,15 6,18 7,16 5,01 5,13 6,54 5,01 6,23 6,21 Institutions (2006) Legal system 5,97 4,89 6,24 5,61 7,45 5,01 7,05 4,90 5,00 4,34 5,02 5,75 6,55 3,75 4,03 6,04 4,17 5,41 5,36 Parliament 4,91 4,99 5,76 4,22 6,40 5,00 5,99 4,33 4,20 3,35 4,82 5,34 5,65 2,68 3,81 5,62 4,22 4,76 4,80 Mean (total) 4,75 4,94 5,56 4,78 6,47 4,98 6,24 4,53 4,62 4,53 5,28 5,14 5,80 3,60 3,95 5,38 4,15 5,00 5,08 UN 4,50 5,10 5,36 4,94 6,65 5,04 6,58 4,70 5,19 5,68 5,81 5,41 6,87 5,05 4,72 6,32 4,57 5,43 5,47 EU- Parliament 4,02 4,98 4,61 4,18 4,83 5,05 5,00 4,31 3,55 5,22 5,37 4,61 4,55 4,26 4,04 3,95 4,53 4,51 4,58 Politicians 3,25 4,24 4,77 3,23 5,59 3,68 4,88 3,49 3,59 2,68 3,92 4,69 4,24 1,92 2,06 4,19 3,10 3,73 3,83 Police 6,18 5,78 6,86 6,48 7,94 5,91 7,96 5,66 6,12 5,17 6,59 5,97 7,13 4,58 5,06 6,49 4,71 6,20 6,21 Institutions (2004) Legal system 5,83 4,83 6,14 5,54 7,21 4,72 6,90 4,77 5,12 4,43 5,21 5,50 6,35 3,01 3,94 5,77 3,85 5,29 5,36 Parliament 4,77 4,68 5,52 4,21 6,29 5,09 6,01 4,27 4,29 3,63 4,71 4,67 5,42 2,40 3,72 5,35 4,13 4,66 4,80 Mean (total) 5,00 4,91 5,67 5,04 6,40 4,60 6,13 4,64 4,75 5,16 5,12 5,28 5,91 4,24 4,49 5,75 4,29 5,16 5,08 UN 4,54 5,10 5,42 5,18 6,54 4,71 6,46 4,61 5,31 5,99 5,68 5,41 6,76 5,64 5,37 6,58 4,90 5,54 5,47 EU- 1 Parliament 4,21 4,88 4,81 4,52 4,84 4,82 4,88 4,39 3,64 5,67 5,1 4,72 4,68 4,75 4,85 4,02 4,65 4,64 4,58 Politicians 3,52 4,28 4,93 3,50 5,47 3,37 4,78 3,63 3,79 3,88 3,75 4,87 4,58 2,72 2,82 4,72 3,07 3,99 3,83 Police 6,44 5,64 6,80 6,73 7,90 5,43 7,95 5,89 6,04 4,91 6,53 5,82 6,99 4,95 5,13 6,76 4,89 6,21 6,21 Institutions (2002) Legal 1 system 6,08 4,39 6,19 5,73 7,13 4,31 6,75 4,83 5,03 5,1 5,14 5,38 6,33 3,68 4,26 6,06 4,28 5,38 5,36 Parliament 5,10 4,99 5,75 4,47 6,18 4,83 5,79 4,45 4,68 5,00 4,43 5,22 5,70 3,48 4,44 5,92 4,04 4,97 4,80 ermany nited Countries participate in all 3 waves Austria Begium Switzerland g Denmark Spain Finland France u Kingdom Hungary ireland Netherlands Norway Poland Portugal Sweden Slovenia Total Mean (total 3 waves) Trust in institutions of the System – european Comparisons – and Slovenia 267 Continuing TABle 9.02 Mean (total) 3 waves 3,26 4,00 5,38 4,96 5,03 5,06 5,63 3,59 4,55 2,80 4,98 Mean (total) 3,26 5,38 4,96 3,59 4,55 2,80 4,86 4,98 UN 4,93 4,28 5,57 4,30 5,32 3,75 5,29 5,35 EU- Parliament 4,76 5,84 5,33 3,92 5,14 3,89 4,63 4,66 Politicians 1,72 4,36 3,51 2,92 3,57 2,04 3,58 3,73 Police 3,89 5,63 5,54 3,38 4,71 2,61 5,74 6,00 Institutions (2006) Legal system 2,46 6,08 5,12 3,81 4,22 2,45 5,06 5,22 Parliament 2,18 5,76 4,55 3,38 4,20 2,32 4,48 4,66 Mean (total) 3,88 4,84 4,85 5,73 4,93 4,98 UN 4,92 3,96 5,19 5,50 5,30 5,35 EU- Parliament 4,38 5,34 4,88 5,22 4,59 4,66 Politicians 2,73 3,59 3,23 5,18 3,68 3,73 Police 4,23 6,03 6,36 6,47 6,06 6,00 Institutions (2004) Legal system 3,72 5,38 4,92 6,14 5,20 5,22 Parliament 3,19 4,69 4,41 5,76 4,59 4,66 Mean (total) 4,29 5,21 5,32 5,52 5,16 4,98 UN 5,33 4,61 5,62 5,10 5,46 5,35 EU- Parliament 4,67 5,75 5,54 5,03 4,76 4,66 Politicians 3,22 3,46 3,54 4,75 3,94 3,73 Police 4,98 6,43 6,66 6,43 6,21 6,00 Institutions (2002) Legal system 3,81 6,27 5,49 6,23 5,42 5,22 Parliament 3,62 4,83 4,83 5,68 4,94 4,66 ep.r stonia reece ussia kraine Countries participate in all 3 waves Countries participate in 1 or 2 waves Bulgaria Czech Cyprus e g italy luxembourg r Slovakia u Mean (total) Mean (total 3 waves) 10 Types of Non-Electoral Political Participation in Europe Klára Vlachová | Tomáš Lebeda 10.1 Introduction Is a high rate of political participation by the public good or bad for democracy? This question has consistently been a topic of interest in political sociology, and political sociologists have developed three key theories of political participation: the theory of democratic elitism, the rational choice theory, and the theory of participatory democracy [ cf. Faulks, 1999]. The theory of democratic elitism and rational choice theory are instrumentalist theories of participation, interpreting political participation as a means of attaining political aims. Supporters of the theory of democratic elitism, such as Schumpeter [2004], accord greater importance to the maintenance of political stability than upholding democracy and find the enlightened leadership of elites preferable to broad political participation by the masses, due to the apathy and lack of education among the latter. Rational choice theory takes a similarly minimalist view of political participation [Downs, 1957; Olson, 2000] and regards a low level of political participation as a sign of rationality on the part of citizens, since an individual has little more to gain from participating in politics than from not participating. If citizens are to be encouraged to participate, they much be convinced that the benefits of participating outweigh the efforts of doing so. Conversely, the theory of participative democracy considers the public’s political participation as crucial to democracy. Political activity is not just a means of attaining certain ends, as it actual y helps strengthen civil society and teaches citizens about the art of governance. According to this theory, democracy flourishes only when citizens are interested in politics, informed, and political y active. A lack of interest and political apathy can lead to a decline in the quality of democratic governance and make it easier to usurp power. But in modern democracies, far from the majority of citizens are interested in politics, and not all of them participate. Many authors [ e.g. Almond and Verba, 1963; Milbrath, 1965] agree that a high rate of political participation and interest in politics is not necessarily good for democracy. According to Almond and Verba’s “myth of civic competence” (1963), participation should be moderate in form and frequency. In their view, what is important in modern democracies is not politically active citizens but politically competent citizens who believe that if they wish, they can have an impact on the political process. Too much or too little participation pose equal threats to the system’s delicate balance. However, the fact is that there are no firm criteria with which to evaluate levels of political participation [McDonough, Shin and Moises, 1998]. Western democracies have shown that they are much stronger than many 272 Klára Vlachová | Tomáš lebeda analyses have indicated. In their early stages, the fragile post-war democracies managed to survive low rates of political participation, and later on, all Western democracies survived the wave of protest movements at the end of the 1960s, which also posed a threat to their existence. Some authors are currently drawing attention to the decline in the level of election participation and the limited public involvement in political parties and politically oriented groups, and they see this as the malady of modern democracies. Others, such as Dalton and Kuechler [1990], Topf [1995], Kaase and Newton [1995], McAdam, McCarthy and Zaid [1996], Smith [1997], Beck [1997], and Norris [1999], believe that these conventional forms of political participation are indeed on the wane, but as they recede, new and previously unconventional forms of civic activities that focus more on post-material values, such as peace, the environment, and animal rights, will rise. Western versus Central and Eastern Europe Many scientists studying the transformation in Central and Eastern Europe were concerned that the new, post-communist democracies would face political apathy and a low level of political participation, which, in the long term, would be detrimental to their development as democracies. While the first elections in the post-communist societies saw participation rates of 80-90% and a dramatic increase in the number of political parties, researchers soon identified something they began referring to as the “post-honeymoon effect” – a subsequent decline in election participation, in membership in political parties and civic associations, and in the public’s interest in politics [Dalton, 2000; Inglehart and Catterberg, 2002]. The low level of political participation and interest in politics in post-communist societies are considered to be a result of the public’s experience with communist politics, a weakened civil society, and a low level of interpersonal trust [Putnam, 1993]. The totalitarian repression by the communist regimes had a much more devastating effect on civil society than did the repression experienced under authoritarian regimes.Up until 1989, most citizens in communist countries had experience with involuntary politics and civic participation in interest groups, communist parties, or participation in public demonstrations. A smaller number had participated in protest events (rallies, strikes, petitions) against the government or the system of government. But in these countries few had experience of voluntary, conventional political participation. While the compulsory political participation found in the totalitarian regimes may be a source of people’s aversion to conventional participation in the new democracies, mass membership in communist parties and interest groups may Types of Non-electoral Political Participation in europe 273 equally have served as a form of socialization in political participation that could have positive repercussions in the democratic system of governance [Bahry and Silver, 1990; Letki, 2004]. Forms of Political Participation in Democratic Countries It is not possible to lay down a complete list of all forms of political participation. In the 1960s, Milbrath [1965] put together a relatively exhaustive hierarchy of forms of conventional political participation (Table 10.01), which encompasses most political activities that are typically observed in a normal democratic process. In this hierarchy, activities differ in terms of the amount of time and energy they require. The higher up the hierarchy an activity is located, the more time and energy it requires. And the higher up the hierarchy an activity is located, the fewer the number of people who participate in it. Whoever takes part in activities listed at the top of the hierarchy also tends to participate in activities at lower levels. TABle 10.01 A Hierarchy of Forms of Participation (Milbrath 1956) Holding public and party office Being a candidate for office Soliciting political funds gladiatorial Activities Attending a caucus or a strategy meeting Becoming an active member in a political party Contributing time in a political campaign Attending a political meeting or rally Making a monetary contribution to a party or candidate Transitional Activities Contacting a public official or a political leader Wearing a button or putting a sticker on the car Attempting to talk another into voting a certain way initiating a political discussion Spectator Activities Voting exposing oneself to political stimuli In the 1970s, Marsh [1977] drew up a scale of unconventional forms of political participation, which encompasses activities ranging from signing petitions to participating in demonstrations, occupying buildings, destroying of property, and personal violence. Marsh set these activities in the context of “pressures for change” and used them to measure protest potential. 274 Klára Vlachová | Tomáš lebeda Political participation is a function of stimuli, personal factors, the political environment, and social position [Milbrath, 1965], and the level of political participation in a state is affected by both social factors on the micro-level (political attitudes – interest, trust, internal and external political efficacy, and personal characteristics – age, sex, education, social status, religiosity), and macro-level (the duration of democracy in a country and its type, how economically advanced a state is, its political system, how easy or difficult it is to participate). At present, the most common form of political activity in democratic states is voting. This is a specific form of political behavior in that it only takes place occasionally and it is strongly affected by the mechanisms of social and legal control [Marsh and Kaase, 1979]. For this reason, it is no longer ranked with conventional forms of political participation. Other non-electoral forms of political engagement, especially ones connected with political parties, are fewer in number. It is these less common forms of political participation, their types and incidence that we are interested in here. But some caution is necessary when comparing non-electoral participation in different countries, cultures, and sub-cultures, as the different forms of political participation are not always rooted in the same way in each place. What appears to be a natural form of political activity for the middle class need not be natural for the working class. A regular form of political activity in Protestant countries may not be as common in Catholic countries. Moreover, dramatic changes are currently occurring in the patterns of political activity. Political activities that were formerly regarded as unconventional are now often mainstream – for example, signing petitions [Norris, 1999] – while, conversely, what older studies indicated as conventional forms of participation may now be unconventional (our analyses indicate an example of this to be the practice of wearing or displaying campaign badges or stickers). In this study we will attempt to show whether any specific types of non-electoral political behavior can be observed in European democracies, how the level of different non-electoral political activities varies between European countries, and how political participation differs on the basis of some selected micro- and macro-factors. 10.2 Data We used data from the first wave of the European Social Survey for our analysis. The first wave of the ESS involves twenty-two countries, but given that the analysis is intended to focus solely on European countries, Israel has been excluded from the sample. The ESS looked at ten forms of non-electoral political activities. The survey question read as follows: There are different ways of trying to improve things in [country] or help prevent things from going wrong. During the last 12 months, have you done any of the following?1 Firstly … READ OUT Contacted a politician, government or local government official. Worked in a political party or action group. Worked in another organization or association. Worn or displayed a campaign badge/sticker. Signed a petition. Taken part in a lawful public demonstration. Boycotted certain products. Deliberately bought certain products for political, ethical or environmental reasons. Donated money to a political organization or group. Participated in illegal protest activities. 10.3 Results The average level of overall non-electoral political participation (all forms studied) varies considerably across Europe. Table 10.02 and Figure 10.01 show how the countries in the study differ from or resemble one another on the basis of the ten non-electoral political activities listed above. Similarities can be found in the average levels of overall (non-electoral) political participation in four geographic regions of Europe – Northern Europe or Scandinavia (Sweden, Norway, Finland), Western European states (Germany, Denmark,2 1 Response categories: Yes, No, Don’t know 2 Denmark tends to be geographically ranked with the Scandinavian countries. However, the political activity of the Danes corresponds more with the average participation observed in Western European countries. Its “continental” position goes hand in hand with a similar level and structure of political participation to that in Western European countries. 276 Klára Vlachová | Tomáš lebeda France, Belgium, the Netherlands, Luxembourg, the UK, Ireland, Switzerland, and Austria), the Mediterranean region or Southern Europe (Spain, Portugal, Italy, and Greece), and the post-communist countries of Central Europe (Czech Republic, Poland, Hungary, Slovenia). The countries with a higher level of overall non-electoral participation are primarily northern countries, especially Norway and Sweden, and even Finland. A high level of participation is also typical of Switzerland. The other Western European countries are very similar, and, with the exception of the Netherlands, the level of participation is above average. Conversely, the lowest average level of overall participation is found in the post-communist countries of Central Europe – in Poland, Hungary, and Slovenia – and in the post-authoritarian countries in the southern Mediterranean – Greece, Portugal, and Italy. The Netherlands, the Czech Republic, and Spain form a curious trio – three countries with a very similar level of participation, but in each case, a level unlike that of the rest of their regions. Compared to the other Western European countries, participation in the Netherlands is lower and overall, below average. The Czechs and the Spanish, on the other hand, participate substantially more than the other post-communist and southern European countries respectively. However, when we look at individual activities (Table 10.02) we find that the structure of participation in Spain differs considerably from the structure in the Czech Republic and the Netherlands. In Spain, participation in public demonstrations is strong and substantially contributes to its above-average level of participation compared to other Mediterranean countries. On the other hand, the structure of activities that Czechs participate in is much more like the participation structure observed in Western Europe than that in other post-communist countries. Figure 10.01 vividly indicates the size of the differences in the average level of overall non-electoral participation. The highest participation level in Norway is four times the lowest level of participation, which was observed in Hungary. Overall non-electoral participation is much lower in the post-communist and Mediterranean areas than in the Western European and particularly the Scandinavian countries [ cf. Barnes and Simon, 1998; Letki, 2004]. The structure of participation in Denmark is closer, for example, to the structure of the participation in neighbouring Germany than to that in Norway or Sweden. Types of Non-electoral Political Participation in europe 277 TABle 10.02 Participation in Non-electoral Political Activities uK 18.2 3.4 9.2 7.8 9.8 32.3 40.0 26.1 4.4 0.8 Switzerland 17.4 7.8 17.3 18.1 9.4 44.6 39.4 31.4 7.9 1.9 Sweden 16.5 5.0 24.6 6.5 10.7 55.2 40.8 32.5 6.4 0.8 Spain 12.4 6.1 16.7 5.3 9.8 11.6 24.2 8.0 17.5 1.7 Slovenia 12.2 3.5 2.3 6.5 2.1 9.6 11.8 5.1 2.7 0.8 Portugal 12.3 4.2 4.2 3.9 6.5 6.9 7.3 3.4 4.3 0.3 Poland 9.6 2.9 5.8 9.2 2.9 9.8 6.9 3.6 1.3 0.2 . Norway 23.8 9.4 28.1 11.5 22.8 36.7 37.3 20.3 9.0 0.8 Netherlands 14.5 3.4 23.1 7.9 3.8 25.7 22.4 10.4 2.9 0.4 luxembourg 18.4 3.7 16.4 15.0 5.2 30.0 29.0 15.8 20.9 3.6 ireland 22.1 4.7 13.8 10.5 9.4 24.6 27.6 13.6 7.1 0.8 Hungary 14.7 2.9 2.8 1.7 3.2 10.5 4.2 4.8 3.7 0.8 greece 14.6 4.8 5.7 2.3 2.7 6.6 4.7 8.5 4.5 1.5 germany 12.8 3.9 17.8 9.2 5.8 39.2 30.4 26.1 10.6 1.2 France 17.6 4.9 17.6 3.0 11.5 28.1 34.8 26.6 17.9 0.8 Finland 24.4 3.4 30.7 6.9 15.7 41.8 24.0 26.8 2.0 0.3 Denmark 17.8 4.1 17.3 9.1 5.1 43.8 28.2 22.9 8.33 1.1 Czech republic 22.8 4.7 15.1 11.6 5.0 22.6 16.1 10.8 4.6 1.4 Belgium 17.8 5.4 23.2 9.5 7.4 27.0 33.9 12.8 8.4 2.4 Austria 18.5 10.3 19.1 11.4 7.9 29.7 27.3 21.5 9.6 1.3 ork for a political ork in another org. Contacting politicians / officials W party/ activist group W or association Donating money Displaying of badge/ sticker Purchase of a product Signing of a petition Boycott of products Taking part in a lawful demonstration Participation in illegal protest Note: Percentage of people who indicated that they have done the given activity in the 12 months prior to the survey 278 Klára Vlachová | Tomáš lebeda Figure 10.01 A Comparison of Average Levels of Overall Non- Electoral Participation in the Analyzed Countries ­     Note: Percentage of people who indicated that they have done the given activity in the 12 months prior to the survey. The percentage of wholly apathetic citizens tends to outweigh the percentage of active citizens in the Mediterranean and post-communist regions, with one exception – the Czech Republic. The proportion of people in a society Types of Non-electoral Political Participation in europe 279 that are wholly apathetic compared to the proportion that participate to at least some degree has a much stronger effect on the average level of overall participation than the level of participation among active people does. The aim of our analysis is not just to compare the average level of overall participation in the countries surveyed, but also to determine whether it is possible to find latent types of non-electoral political participation that match the types described in the literature. We used exploratory factor analysis on the sample to analyze the ten items listed above and extracted three factors (Table 10.03). The first factor is strongly correlated with the items “contacted a politician/official”, “worked in a political party or action group”, “worked in another organization or association”, “donated money to a political organization or group”, and “wore or displayed a campaign badge/sticker”. The second factor strongly correlates with the items “signed a petition”, “boycotted certain products for political reasons”, and “bought certain products for political reasons”. The third factor strongly correlates with the items “took part in a lawful public demonstration”, “participated in illegal protest activities”, and also with the item “wore or displayed a campaign badge or sticker”. This item correlates with the third factor even more strongly than with the first factor. The items that correlate with the first factor (1 to 4) and with the third factor (5 to 7) can be described as active displays of participation. A citizen must make a decision beforehand to engage in these forms of participation and expend a certain amount of energy when participating. While the first factor encompasses items that represent conventional forms of participation, which can be undertaken in a relatively private, discreet manner, the third factor involves activities that are clearly public and manifest in character, requiring a person to stick out their neck and publicly declare their position on a matter. Conversely, the second factor (items 8 to 10) represents activities that are of a more passive nature. They do not usually require that a person invest much energy in doing them and they often arise out of circumstances conducive to participation ( e.g. we came upon a place where people were signing a petition; we were able to choose from a number of types of goods and to boycott some of them and/or for political, ethical, or environmental reasons to buy a particular item). The second and third factors, for the most part, comprise items that represent more or less unconventional forms of political activities. In addition, the item “wore or displayed a campaign badge/sticker” comes across as ambivalent – as both conventional and unconventional. For the purpose of further analysis, we built sum scales out of the items that correlate with these factors. They were created by adding together the activities that fall under each of the given types of participation for each respondent. 280 Klára Vlachová | Tomáš lebeda In accordance with previous research, the first of them was defined as the active conventional type of participation (items 1 to 4), but neither of the two other scales matched the original definition of unconventional participation. Therefore, we defined the second scale as the active demonstrative type (items 5 to 7) and the third as the passive type (items 8 to 10). TABle 10.03 Types of Participation Factor 1 Factor 2 Factor 3 1 Contacting politicians/officials 0.663 0.145 -0.050 2 Work for a political party/ activist group 0.692 -0.078 0.241 3 Work in another organisation or association 0.572 0.204 0.161 4 Donating money 0.593 0.104 0.023 5 Displaying of badge/sticker 0.417 0.138 0.495 6 Taking part in a lawful demonstration 0.189 0.181 0.696 7 Participation in illegal protest -0.063 0.004 0.758 8 Signing of a petition 0.224 0.556 0.319 9 Boycott of products 0.060 0.812 0.096 10 Purchase of a product 0.119 0.823 -0.010 % explained variance 18.8 17.8 15.0 Note: Factor analysis, varimax rotation The reliability of these scales is below the recommended value of 0.75. For the scale of conventional participation, Cronbach’s alpha is equal to 0.54; for the scale of passive participation it is 0.64; and for the scale of demonstrative participation it is 0.47. Clearly the items corresponding to non-electoral political participation that were used in the research are unable to provide fully reliable measurements for the patterns of political participation for such a large number of European countries. Factor analyses conducted in individual countries in particular show that in five of the countries studied, i.e. Portugal, Sweden, the UK, Greece, and Spain, there are two latent factors of political participation, and in the other countries there are three. Yet, not in every country do the latent types of participation observed correspond to the types found at the European level. Nevertheless, we have decided to use the types of participation we devised as variables in the analyses, since at this point we have no better tool at our disposal for comparing political participation in European countries. Types of Non-electoral Political Participation in europe 281 Analysis of variance confirms that the average level of conventional, manifest, and passive participation varies between countries with greater statistical significance than it does within the individual countries.3 But there are no areas in Europe that can be identified as homogenous regions in terms of overall participation and types of participation, or that, as such, can be distinguished as differing from one another. The three types of non-electoral political participation identified in the countries surveyed are present in varying proportions (see Figure 10.02). In Western European and Scandinavian countries, the passive type of participation clearly predominates over the other two forms. The strongest predominance is in Sweden, followed by the UK, Germany, Denmark, and Switzerland. The passive type of participation is less prevalent in the post-communist countries and even less so in the Mediterranean countries. In Greece, Portugal, and Poland, the passive type of participation is so weakly represented that even the conventional type is more dominant. The Mediterranean area is also interesting for the higher than average proportion of demonstrative type of participation. A similar trend, in a more moderate form, can also be documented in France, which is renowned for its protest activities. The manifest type is also strongly represented in Luxembourg and Norway. What affects the level and types of participation in a given society? As noted above, on the one hand, it is the social micro factors (political attitudes and personal characteristics) and on the other hand, the social macro factors (economic, socio-demographic, socio-cultural, and political characteristics of the given society). In the following section, we will look at some of the micro and macro factors that have a significant effect on overall political participation and individual types of participation. 3 Conventional participation: F=63.6, manifest participation: F=60.0, passive participation: F=283.7, all are statistically significant at 0.001. 282 Klára Vlachová | Tomáš lebeda Figure 10.02 Types of Participation in the Analyzed Countries4 ­     4 In its initial form, the scale of conventional participation is a four-point scale, while the manifest and passive scales are only three-point scales. For the purpose of this comparison they were transformed into scales with the same number of points. The average values for individual states were comparatively expressed in percentages, which are presented in Graph 4. 10.4 A Micro Analysis of Political Participation The first important factor that significantly affects al types of participation is interest in politics. This factor is measured by two variables: 1) a subjectively declared interest in politics,5 and 2) how often the respondent discusses politics and current events. A subjectively declared interest in politics is the strongest variable to correlate with the overal level of political participation (see Table 10.04). In al the countries studied, we find very strong and statistical y significant correlations. The highest correlation is recorded for the UK (rs = 0,391**),6 France, Portugal, Italy, and Austria. In two-thirds of the countries, this factor is more significant than the variable of discussing politics. This second variable is also very significant, but the strength of the correlation shows greater variation between countries. The factor of how often respondents discuss politics shows the strongest effect in Spain (rs = 0,449**), the UK, Switzerland, Italy, and France. Both a subjective interest in politics and how often a person discusses politics has a strong effect on the passive and the conventional types of participation; even their effect on the manifest type is of some significance. Also of significance for the overal level of political participation is where respondents identify themselves on a left-right political scale. Participation, especial y the manifest and passive types, increases as respondents place themselves further left on the scale. Activities like demonstrations, boycotting goods, or signing petitions are ones that, in Western democracies, left-wing citizens tend to engage in more often and do so in support of issues promoted in the programs of left-wing parties and movements. A left-wing political orientation is strongly connected with the increase of passive participation in advanced Western European countries: in Switzerland, (rs = -0,209**), Germany, Denmark, and Austria. The case of manifest participation is similar, with the strongest correlations recorded for France (rs = -0,248**), Austria, Germany, Luxembourg, Denmark, and Spain. This kind of correlation is not, however, found in the post-communist countries. In Hungary, the Czech Republic, and also Poland, but to a lesser degree, the overal level of participation, especial y the manifest (Hungary rs = 0,159**) and passive types of participation, 5 Large differences can be found between European countries regarding the question of interest in politics. The biggest interest in politics is expressed by the Dutch (66%), followed by Germans, Danes (63%), and the Swiss (61%), while the lowest interest in politics was declared by Czechs, Greeks, and Italians (32%), and Portuguese (36%). 6 The Spearman coefficient is used throughout the micro analysis to measure correlations, which are marked as rs . When talking about a group of countries among which a correlation is detected, then the value of the correlation coefficient is only indicated for the strongest or, conversely, the weakest. 284 Klára Vlachová | Tomáš lebeda correlates with a declared right-wing political orientation, or, as in the case of Slovenia, there are no correlations. This kind of phenomenon cannot be found in almost any other country. In Hungary, after the parliamentary elections in 2002, it was mainly citizens with a right-wing orientation that took part in manifest political activities – they took part in demonstrations to demand a vote re-count, as the defeated right-wing coalition was unwil ing to accept the outcome of the elections. In Central and Eastern Europe, less experience with democracy combined with a system that does not always work satisfactorily leads many citizens with a right-wing orientation to engage in manifest and passive political activities. In addition, there is no ruling out the possibility that in this region there is a persistent aversion to left-wing political views, as the term “left-wing” is tainted with notions of the undemocratic method of governance associated with the previous, communist regimes. The level of religiosity, whether subjectively declared or measured on the basis of how often a person attends church, is often connected with the level of political participation. The exact nature of this correlation depends, however, on the type of society. In predominantly Protestant societies, the correlation with political participation is positive (the UK, Norway, Sweden, the Netherlands, Switzerland). Conversely, in countries with a higher proportion of Catholics, people who are very religious and attend church often participate less than others politically (Spain, Italy, Portugal, Greece, Ireland, Luxembourg, Poland). This is most pronounced in the case of the manifest and passive types. In other countries, religiosity does not play a role in the level of political participation, or the correlations with individual types of participation are variable. Religious affiliation proves to be more revealing. It has an effect not just on the level of participation but also on activity within each of the three types. Catholics, on the whole, participate less than Protestants and people with no religious affiliation. Catholics are the least active with regard to the passive type of participation. This is most pronounced in predominantly Catholic countries: in Portugal (rs = -0.167**), Spain, Ireland, Italy, and elsewhere. The correlation between the demonstrative type of participation and Catholicism is also negative: in France (rs = -0,153**), Italy, Austria, and Portugal. On the other hand, Protestantism has a positive effect on the overall level of participation and it correlates most with the passive and manifest types. What is at the root of this observation? Protestant churches are evidently more effective at socializing their members to engage in political and civic participation outside the church [Jones-Correa and Leal, 2001]. However, this is just a hypothesis that requires more study, since in Denmark, for example, Protestantism has a slightly negative effect on these two types of participation. Types of Non-electoral Political Participation in europe 285 The strongest socio-demographic factor across Europe is the highest completed level of education. In every country, without exception, higher levels of education coincide with higher overall participation, though understandably to varying degrees – from France (rs = 0,370**), Poland, the Netherlands, and Belgium through to the Czech Republic and Greece (rs = 0,195**). The strongest correlation between the level of education and participation occurs with the passive type, followed by the conventional type. The most significant exception, where the level of education has a much stronger correlation with manifest participation (rs = 0,255**) than with conventional participation (rs = 0,181**), is in France. But even there the correlation with passive participation remains strongest (rs = 0,364**). The only other similar case is in Greece. It has long been known that higher levels of education correspond with higher levels of political participation. More educated people are more responsive to political stimuli, are more interested in politics, understand political information better, and tend to have more of a desire to influence the political process, as they have more information, time, and energy at their disposal to do so. In contemporary democracies, the education process is, moreover, regarded as a ritual through which an adult becomes a member of the political community. The age at which citizens obtain their political rights usually coincides in these societies with the age at which a person completes the attested level of education regarded as the relatively most important ( e.g. secondary school diploma) [Kamens, 1988]. Another variable with a significant effect is the level of household income. In every country, without exception, people with higher household income participate in politics more often. This association is strongest in Portugal (rs = 0,263**), France, Spain, Luxembourg, the UK, and Ireland. The level of household income is positively connected with the passive type of participation and conversely has the weakest or even no connection with the protest type. More educated people tend to have higher incomes and are also the group that are politically more active. But a higher income is also an important resource that makes it easier to engage in one of the forms of passive participation – the purchase of goods for political, environmental, or humanitarian reasons. Age is a variable that has a significant effect mainly on the manifest and passive types of participation. There are, however, exceptions. The manifest type of participation is dominated by young people. Examples are Luxembourg (rs = -0,292**), Spain, Denmark, the Czech Republic, France, and Italy. The passive type of participation also typical y involves the younger generation the most, which is especial y evident in Spain (rs = -0,292**), Finland, the Czech Republic, Norway, and Sweden. This is not surprising given that it is a modern 286 Klára Vlachová | Tomáš lebeda form of participation and one that the younger generation is more familiar with. On the other hand, the conventional type is usual y independent of any age connection, or it increases slightly with age. The size of the community in which a person lives has a slight effect on the passive type of participation in the majority of the countries surveyed. This is most notable in Finland (rs = 0,199**), France, Switzerland, and Spain. The urban population is more active in the passive type of participation than the rural population, largely because it is easier to engage in passive forms of participation in cities (there is a broader range of products to purchase or boycott, people are more likely to come across petitions, e.g. in public areas). The same is true of manifest participation – it is easier to organize demonstrations in cities than in the countryside. TABle 10.04 Factors Affecting Political Participation in Europe Political participation overall conventional manifest passive interest in politics 0.389** 0.294** 0.188** 0.335** how often discusses politics 0.377** 0.267** 0.212** 0.328** left-right political scale -0.122** -0.052** -0.155** -0.122** level of religiosity -0.118** -0.008** -0.089** -0.125** how often a attends church -0.096** -0.008 -0.063** -0.116** Catholic (roman or greek) -0.191** -0.067** -0.061** -0.223** Protestant 0.111** 0.063** -0.031** 0.132** without religious affiliation 0.111** 0.024** 0.079** 0.125** sex -0.042** -0.098** -0.036** 0.022** age -0.085** 0.008 -0.123** -0.090** size of the community -0.074** -0.014* -0.086** -0.084** education 0.363** 0.235** 0.160** 0.345** income 0.262** 0.153** 0.072** 0.264** ** The correlation is significant 0.01 level. The Spearman coefficient was used. N = 39 860, weighted by population size The questions used can be viewed in the ESS questionnaire, which is available for downloading at: http:// www.europeansocialsurvey.com in the section Questionnaire – Main questionnaire – round 1. Minor differences can also be found between men and women. If there are any statistically significant relationships, then it is usually that men tend traditionally to be more active than women. This predominance is most pronounced in the case of conventional participation and most notably so in the Czech Republic (rs = -0,170**), Italy, and Austria. Activities ranked Types of Non-electoral Political Participation in europe 287 as conventional types of participation are the ones that demand the most time and energy. That is why they tend to be dominated by men, who, in many European countries, still devote much less time and energy to looking after the family. On the other hand, the only type of participation in which women dominate in some European countries is the passive type. This can be explained by the fact that women tend to do more of the shopping, allowing them more of an opportunity to purchase or boycott goods for political, environmental, and other similar reasons. It is with reference to the purchase of goods that the difference between men and women is strongest. Examples are found in Scandinavia, for instance, in Finland (rs = 0,178**) or Sweden. Some of the correlations indicated above may be strong in some countries, while in others they might not appear at all. When examining the “European” population of all twenty-one countries as a whole, some correlations may appear weaker. Conversely, others are much stronger from a European-wide perspective than they are in individual countries, because they derive from the differences between individual societies. Table 10.04 contains a summary of the correlations for the population of all European countries combined. 10.5 A Macro Analysis of Political Participation We used aggregate data to map the key factors affecting the variance in the level of political participation between individual countries. Individual countries were entered into the analysis as separate cases. The average values of overall participation and the average values of all three types of participation were the explained variables. As explanatory variables we selected objectively measurable indicators drawn from international comparative statistics (see Table 10.05 for the sources). These include macro-economic, socio-demographic, socio-cultural, and political indicators. We managed to identify six factors that affect the variance in the level of political participation between countries. These cannot be observed separately because they are strongly intertwined. The most important factor connected with the level of participation in individual countries is the length of democratic experience. We measured this using two constructed variables: 1) the duration of an uninterrupted democratic system in the country, and 2) the duration of overal democratic experience in a country; the latter adds to the first the duration of democratic systems in the country from earlier periods, which were interrupted by various undemocratic periods. The strongest correlations for both variables are with overall participation, and they also show exceptionally strong correlations with the passive and 288 Klára Vlachová | Tomáš lebeda conventional types of participation. Countries that are part of the third wave of democratization – the post-communist and Mediterranean countries, with the exception of Italy – show the lowest values in relation to overall and passive participation. Conversely, the highest values are recorded for countries with a long democratic tradition. The strongest correlation is found for total democratic experience, which suggests that a certain intergenerational “social memory” is at work here. This would explain, for example, the higher level of participation in the Czech Republic compared to other post-communist countries. TABle 10.05 Factors Affecting the Level of Non-Electoral Political Participation in Individual Countries Political participation Data overall convencional manifest passive N source Duration of uninterrupted democracy 0.826** 0.709** 0.529* 0.810** 21 5 Total democratic experience 0.857** 0.762** 0.534* 0.834** 21 5 Per capita gDP (PPP) 0.694** 0.622** 0.693** 0.602** 20 2 Human Development index - HDi 0.787** 0.668** 0.632** 0.742** 21 3 Number of household members -0.716** -0.511 0.062 -0.792** 15 2 Percentage of employees in the agricultural sector -0.653** -0.495* -0.517* -0.633** 20 1 Percentage of employees in the services sector 0.679** 0.568** 0.517* 0.636** 20 1 Spread of the internet 0.666** 0.592** 0.137 0.715** 21 1 Percentage of Catholics -0.541* -0.595** -0.066 -0.551* 20 1 Percentage of Protestants 0.655** 0.586* 0.394 0.652** 17 1 interest-group pluralism -0.718** -0.801** -0.127 -0.655** 17 4 individual countries were entered as cases into the analysis. The Person correlation coefficient was used ** The correlation is significant at 0.01 level. * The correlation is significant at 0.05 level. 1 – CIA, World Factbook, 2002 http://www.cia.gov/cia/download2002.htm 2 – Eurostat, http://epp.eurostat.cec.eu.int 3 – United Nations Development Program, Human Development Report, http://hdr.undp.org/statistics/data 4 – Lijphart (1999) p. 177, p. 80–81. 5 – Variables were constructed from: Dahl (1995) p. 211–219 and Huntington (1991) p. 13–26. Another significant factor is the economic development of a country, which was measured using per capita GDP at purchasing power parity. The Types of Non-electoral Political Participation in europe 289 populations of wealthy and economically advanced countries show a higher level of participation both overall and in the individual types of participation. The strongest correlation is between a country’s economic development and the manifest type of participation, but the correlation with the other two types of participation is also very strong. Economic development is linked to another factor – the social development of a country. The human development index is used as the variable for this factor in the analysis.7 Here again, the populations of more developed countries participate more often, and the range of forms of participation is wider. The fourth factor is the traditional or, conversely, the modern character of society. We attempt to identify the signs of a more traditional society on the basis of a higher proportion of the population employed in agriculture and a higher number of members living within a single household.8 More traditional societies on the whole participate less, especial y in connection with the passive type of participation. Conversely, modern societies, which have the opposite features, as wel as more widely spread Internet access and a higher proportion of employees in the services sector, exhibit a high level of participation, especial y the modern form of participation, i.e. the passive type. An important factor that we have identified at the micro level in individual form is the religious composition of society. The higher the percentage of the population that is Catholic, the lower the level of overal participation. But even stronger than this factor’s correlation with overal participation is its correlation with the conventional type of participation, though there is no correlation at al evident with the manifest type of participation. Conversely, the higher the percentage of Protestants in the population,9 the higher the level of passive, overal , and, though less significantly, even conventional participation. Again, only in the case of manifest participation is there no discernible effect of Protestantism. Manifest participation is, thus, the only type that is unaffected by religious affiliation. The last factor is the predominant form of interest mediation in society. This factor was measured using the degree of interest-group pluralism [see Lijphart, 7 The HDI measures the average levels of three basic dimensions of human development: life expectancy, education, and the standard of living (for details, see the Web site of the United Nations Development Programme: http://hdr.undp.org/statistics/data/indic/ indic_8_1_1.html). 8 Values for the average number of members sharing a single household were only available for 15 of the 21 countries. 9 Figures for the percentage of Protestants in the population were not available for the post-communist countries. 290 Klára Vlachová | Tomáš lebeda 1999: 177).10 This variable is defined along a continuum running between the neo-corporatist and pluralist models of interest mediation. The greater the pluralism of interest groups, the greater the degree of pluralist interest mediation in society. Conversely, low levels of interest-group pluralism are associated with a neo-corporatist model of interest mediation. Participation is strongest in systems with neo-corporatist mechanisms. As the pluralism of interest groups increases, participation declines. The correlation is strongest for the conventional type of participation. We believe that interest groups may assume the function of individual political participation in systems with greater interest-group pluralism. The competing interest groups, organizations, and lobbies that are typical for the pluralist model thus take over the function of articulating interests or even the role political participation itself. This may reduce the need for isolated citizens to engage in individual activities. Our conclusion is also supported by the fact that this connection is most pronounced for the conventional type of participation, which has the closest ties to the decision-making process. This connection is also strong and significant in the case of the passive type. But there is no correlation in the case of the manifest type. Highly organized and monopolized interest groups (especial y unions and employer and professional associations), which in neo-corporatist systems often co-participate in the process of developing government policy, ensure that their broad membership bases respect the agreed compromises and do not actively oppose them. That is why there is no correlation between neo-corporatism and the manifest type of participation. 10.6 Conclusion In contemporary democratic Europe, we can identify active and passive types of non-electoral political participation, while the active form can be further broken down into the conventional and the unconventional/manifest sub-types. Conventional participation signifies traditional, peaceful forms of political activity, from contacting politicians to working for a political party or some other politically active group, to financially supporting such activities. This type of activity requires an active and targeted approach, but it need not be manifested publicly. On the other hand, the second type of active participation – the manifest type – involves a form of participation that is 10 Lijphart’s values for this factor do not include post-communist countries. Therefore, the Czech Republic, Poland, Hungary, and Slovenia were not included in this analysis. Types of Non-electoral Political Participation in europe 291 expressed publicly and openly, from the wearing or displaying campaign badges or stickers, to legal demonstrations, to even illegal protest activities. The manifest type of participation is the weakest type across Europe, and it is not a regularly practiced type of political participation. A new form of political participation that we can identify in contrast with earlier studies is the unconventional, passive type. This covers activities such as boycotting or purchasing goods for political, ethnical, or environmental reasons. According to our analyses, signing petitions can also be identified as a form of this activity. The targeted boycotting or purchasing of goods is a new form of participation that is becoming more widespread in economically advanced societies, and it allows people to respond to political issues in a globalized world ( e.g. the boycotting of American goods in Europe after the start of the military campaign in Afghanistan and Iraq,11 or the boycotting of French goods in the United States after France criticized these military campaigns). Signing petitions has long been a marginal and unconventional form of political activity, but today it is a frequently employed form of political participation. The comparative level of political participation in individual societies is uneven. Those who participate more are people who have better resources for engaging in such activity – a greater interest, more information, higher education, higher income, and more time. Political participation is skewed towards individuals who are privileged in these areas. Overall political participation varies considerably between European countries. In the Scandinavian countries, participation is the highest, and it is also high in Western Europe. Lower participation levels are observed in the Mediterranean and in the post-communist countries. A high level of political participation is observed in modern, rich, economically advanced countries with a long democratic tradition and a firmly established democratic system. But in these advanced societies, it is the unconventional, passive type of participation that predominates. The question is whether a deeply embedded democratic system is the outcome of a high level of systematic participation among citizens, or whether, on the contrary, a democracy that functions long and well is what leads citizens to engage more in political activity. 11 In Germany, for example, signs could be seen in restaurants informing customers that under the current political situation they would not be offering Coca-Cola. Part II – The Visual Road to Comparative Research Part II: The Visual Road to Comparative Research Karl H. Müller | Niko Toš The first road to comparative research was centered on statistical analyses and on comparative descriptions of living conditions, attitudes and the like. This road, being the most widely used, was characterized as the traditional way for comparative investigations on the basis of micro-social data. The second road to comparative research is a path which, so far, was not considered a research trajectory altogether, but as a necessary side-step within the first road. Already along the first road diagrams are widely used to display the results of statistical analyses. In sharp contrast to a mere auxiliary instrument for displaying statistical information, the second road to comparative research builds on data visualization as an instrument and as an object for research. Thus, the second approach transforms a set of comparable data into a visual pattern with which a competent observer interacts and, moreover, which serves as a heuristic basis for the formation of new hypotheses. Here, the part of statistical analyses is substituted with a complex interaction between a visual data pattern on the one hand and a competent social scientist on the other hand. It must be noted that the first and the second road to comparative research are not strictly independent from each other. In many instances it is useful to use the second road first in order to get quick insights into interesting or significant characteristics of a data set. Afterwards, it might be advisable to use statistical techniques in order to substantiate the results of visual interactions further. In a nutshell, the second road for comparative research can best be utilized as an exploratory way of data analysis while the first road is the appropriate form for testing the statistical viability and significance of micro-social hypotheses or theories. Figure II.01 summarizes the main ingredients for the second road to comparative research. Figure II.02 presents an overview of a new online program under the name of WISDOMIZE which has been used for the production of diagrams in Part II. 296 Karl H. Müller | Niko Toš Figure PArT ii.01 The Second Road to Comparative Research Methods for Pattern Formation Comparable Comparative Micro-Data Visual Results Visual Competent Pattern Observer Figure PArT ii.02 The Overall Structure of WISDOMIZE XML FILE GRAPHIC OUTPUT EXPERT MODE OUTPUT WIZARD MODE XML EXCHANGE COMPILING DATA STORAGE CALCULATION DRAW The Online-Program WISDOMIZE in a Nutshell Armin Reautschnig The graphics you will see as you walk along the second road to comparative research in this book are data visualizations, made especially for this project by using a new tool called WISDOMIZE. WISDOMIZE is a web application for the generation of data-based graphics. These graphics consist of graphical prototypes and are created online and on demand. The application can be found through the URL http://www.wisdom.at/wisdomize Thus, the main function of WISDOMIZE lies in the creation of data-based graphics which use socio-economic micro-data from regional, national or European surveys or panels as their reference domain. WISDOMIZE was developed using Ruby on Rails, a web application development framework written in Ruby. It uses mysql-databases for data storage and serves the generated graphics in SVG (Scalable Vector Graphics) format. The settings used to create a graphic are saved in the application and can be exported as XML-data and saved locally. The application follows the structure of the Input-Processing-Output-(Storage) Model (IPO+S Model) and has a client-accessible front-end, running in a web browser, while the server-side back-end handles the creation of the requested graphics. The parameters of the graphics are passed on to the program through a graphical user interface (GUI). Figure Part II.02 shows the inner structure of the application and the functional relationships between the individual components. The tool interfaces will be explained below. The Two Interface Modes The GUI is equipped with two modes of operation that are oriented to the different needs of two groups of users. – In the “Wizard Mode”, graphics are compiled in a step-by-step process using pre-defined components. The creation of a graphic is thus reduced to simply a few mouse-clicks. – In the “Expert Mode” the components of the graphic can be newly defined or expanded before the actual compilation occurs. The possibilities for controlling the graphics are thereby increased, but the use of the program is, by the same token, made more complex. The creation of the components are saved in the application and can be shared with other users. In this way, the capacity of the application increases through use itself. By providing two interface modes, broader application areas can be uncovered and more user groups can be served. 298 Armin reautschnig With respect to the client-side, the GUI runs in a web browser and is connected to the back-end of the program through the internet. The settings from the interface are analyzed server-side and used to generate the graphics. The output format of the graphic sent to the front-end is the text-based vector format SVG (Scalable Vector Graphic). The graphics can thus be scaled without any reduction in quality and, therefore, are also suitable for print publishing. Figure PArT ii.03 Available Choices for the Wizard-Mode of WISDOMIZE GRAPHIC VARIABLE Static Patterns B4 TRSTPRL Dynamic Patt... B5 TRSTLGL Complex Patt.. B6 TRSTPLC B7 TRSTPLT B8 TRSTPRT COUNTRY DATASET STATISTIC Particip...ESS 1 ESS 1 Deviation Particip...ESS 2 ESS 2 Particip...ESS 3 ESS 3 Inters...ESS 1-3 ESS 1–3 Inters...ESS 1-3 ESS 1–2 SUB-POP 4 Age Goups... 3 Education G. Gender Basic Populati.. In the “Expert Mode” an additional function – the XML file output – is available. The individual parameters of the graphic are saved in XML format and can be accessed locally as data. These “wisdomize exchange files” assist with documentation, archiving and retrieval of the compilation processes. The definition of the parameters occurs, as already mentioned above, through the graphical interface of the application. The graphic below exhibits Part ii: The Visualization Approach to Comparative research 299 the predefined components of the interface in “Wizard Mode”. These components can be defined in “Expert Mode” by the user on the level of variable of the data set. The graphical prototypes cannot be modified. The “visual library” of the tool comprises three areas. Graphical Prototypes The graphical prototypes of WISDOMIZE have been separated into three groups, namely in static, dynamic and complex prototypes. Static Prototypes: They present the value of the variables for an entire population or multiple subpopulations for one point in time. Several of these prototypes have been developed. They include, inter alia, simple diagrams, arranged as rectangles, deviations from the mean value of a variable, deviations from a user-defined mean. Static patterns can be created for a single variable up to seven variables in the Wizard Mode and for more variables (eight and higher) in the Expert Mode. Dynamic Prototypes: They present the values of the variables for an entire population or multiple subpopulations for several points in time. Again, various prototypes are available in the WISDOMIZE library of prototypes. One of the basic differences to the static prototypes lies in the overall arrangement. Dynamic patterns are ordered along a vertical axis, whereas the static patterns have been arranged around a horizontal axis. Complex Prototypes: They present the values of a large number of variables for an entire population or multiple subpopulations for one point in time. Several prototypes are available which all share as their common feature that they are able to use a large number of data and they transform them into a visual pattern which can be used for exploratory visual data analyses by competent users. Through the publicly accessible web application introduced here, a method for complex exploratory data analysis will become available. These graphics are no longer merely cosmetic renderings of results, but can also enable the recognition of correlations, of clusters and the formulation of hypotheses. The data-bases, so far, consist of the European Social Survey (ESS), and will be expanded through additional data sets (e.g. ISSP, European Value Studies, etc.). A uniform documentation of data sets according to the DDI 3.0 guidelines of the Data Documentation Initiative is essential for this process. What follows now, is a more detailed description of the diagrams which have been produced with the help of WISDOMIZE and which can be found in chapters 11 to 13 of the book. 11 Static Patterns Armin Reautschnig Here, we have used mainly a single graphical prototype, namely deviations from a user-defined mean1. With this graphic prototype, the values of one or more variables can be presented for the entire population or diverse subpopulations at a given point in time. These single units of information are presented as rectangles, originating on a horizontal axis. The width of the rectangle is determined by the number of units to be represented and by the available width in the graphic. According to these relationships, either the output size of the graphic is determined by the number of units to be presented or the number of units is limited by the possible output size of the graphic. Figure 11.01 The Prototype for Static Patterns: Deviations from a Self-Defined Mean 1 Obviously, WISDOMIZE has more graphical prototypes available for static patterns. For the book however, just a single prototype has been used. 304 Armin reautschnig The criterion for the output size or a limitation of the units is the readability of the graphic, which is in turn strongly linked to the output medium. Varying ratios result from the different resolutions of, for example, print and computer monitors, and must always be fine-tuned. The height of the rectangle corresponds to the processed values. If the values of multiple variables for an entire population are to be illustrated, they are separated from one another with blank spaces. If the representation is also to include subpopulations, the values of the individual subpopulations are shown separated according to variable. Here, once again, spaces are inserted to separate the subpopulations visually. The scale of the graphic is linear with an integer-based gradation. The range of the scale is variable and is determined by the quantity of the maximum processed value of the entire graphic. Thus, the area of interest is brought to the fore. In order to further increase readability, the rectangles are colored differently depending on whether they have a positive or negative value. The diagrams in this book restrict the number of units to 28 which is due to the fixed size of the book layout and the resolution of the book print. Such a unit can be filled at the maximum with seven variables and four subpopulations. In the book, the values of the variables are presented as deviations from the mean of the scale from zero to ten, i.e., 5. The rectangles of the units whose processed value is above the average are shaded in light grey, while those below the average are dark grey. 11.1 Working Conditions grAPHiCS 11.1.01 – 11.1.08 VAriABleS +2 A B C D E F G +1 -1 -2 A Allowed to be flexible in working hours B Allowed to decide how daily work is organised C Allowed to influence job environment D Allowed to influence decisions about work direction E Allowed to change work tasks F Get similar or better job with another employer G Start own business grAPHiCS 11.1.01–11.1.08 grOuPS: geNDer Age eDuCATiON a b c d e f g h i a Male c 15 – 29y b Female g Not completed primary education d 30 – 44y Primary or first stage of basic e 45 – 60y Lower secondary or second stage of basic f 60y + h Upper secondary Post secondary i First stage of tertiary Second stage of tertiary 306 Armin reautschnig Figure 11.1.01 Working Conditions in General (CEE) D PL CZ CH A I SLO H Static Patterns 307 Figure 11.1.02 Working Conditions According to Three Levels of Education (CEE) D PL CZ CH A I SLO H 308 Armin reautschnig Figure 11.1.03 Working Conditions According to Four Age Groups (CEE) D PL CZ CH A I SLO H Static Patterns 309 Figure 11.1.04 Working Conditions According to Gender (CEE) D PL CZ CH A I SLO H 310 Armin reautschnig Figure 11.1.05 Working Conditions in General (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 311 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 312 Armin reautschnig Figure 11.1.06 Working Conditions According to Three Levels of Education (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 313 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 314 Armin reautschnig Figure 11.1.07 Working Conditions According to Four Age Groups (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 315 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 316 Armin reautschnig Figure 11.1.08 Working Conditions According to Gender (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 317 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 11.2 Work Satisfaction grAPHiCS 11.2.01 – 11.2.08 VAriABleS +2 A B C D E F +1 -1 -2 A How satisfied with job B Satisfied with balance between time on job and time on other aspects C Find job interesting D Find job stressful E Become unemployed in the next 12 months, how unlikely F Get paid appropriately, considering efforts and achievements grAPHiCS 11.2.01 – 11.2.08 grOuPS: geNDer Age eDuCATiON a b c d e f g h i a Male c 15 – 29y b Female g Not completed primary education d 30 – 44y Primary or first stage of basic e 45 – 60y Lower secondary or second stage of basic f 60y + h Upper secondary Post secondary i First stage of tertiary Second stage of tertiary 320 Armin reautschnig Figure 11.2.01 Work Satisfaction in General (CEE) D PL SK CH A H SLO Static Patterns 321 Figure 11.2.02 Work Satisfaction According to Three Levels of Education (CEE) D PL SK CH A H SLO 322 Armin reautschnig Figure 11.2.03 Work Satisfaction According to Four Age Groups (CEE) D PL SK CH A H SLO Static Patterns 323 Figure 11.2.04 Work Satisfaction According to Gender (CEE) D PL SK CH A H SLO 324 Armin reautschnig Figure 11.2.05 Work Satisfaction in General (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 325 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 326 Armin reautschnig Figure 11.2.06 Work Satisfaction According to Three Levels of Education (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 327 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 328 Armin reautschnig Figure 11.2.07 Work Satisfaction According to Four Age Groups (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 329 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 330 Armin reautschnig Figure 11.2.08 Work Satisfaction According to Gender (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 331 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 11.3 Working Conditions and Work Satisfaction grAPHiCS 11.3.01 – 11.3.08 VAriABleS +2 A B C D E F G H I J K L M +1 -1 -2 A Allowed to be flexible in working hours B Allowed to decide how daily work is organised C Allowed to influence job environment D Allowed to influence decisions about work direction E Allowed to change work tasks F Get similar or better job with another employer G Start own business H How satisfied with job I Satisfied with balance between time on job and time on other aspects J Find job interesting K Find job stressful L Become unemployed in the next 12 months, how unlikely M Get paid appropriately, considering efforts and achievements grAPHiCS 11.3.01 – 11.3.08 grOuPS: geNDer eDuCATiON a b a b e f e f a Male b Female e Not completed primary education Primary or first stage of basic Lower secondary or second stage of basic Age Upper secondary Post secondary c d c d f First stage of tertiary Second stage of tertiary c 15 – 40y d 40y + 334 Armin reautschnig Figure 11.3.01 Working Conditions and Work Satisfaciton in General (CEE) CH D PL A SLO H Static Patterns 335 Figure 11.3.02 Working Conditions and Work Satisfaciton According to Two Levels of Education (CEE) CH D PL A SLO H 336 Armin reautschnig Figure 11.3.03 Working Conditions and Work Satisfaciton According to Two Age Groups (CEE) CH D PL A SLO H Static Patterns 337 Figure 11.3.04 Working Conditions and Work Satisfaciton According to Gender (CEE) CH D PL A SLO H 338 Armin reautschnig Figure 11.3.05 Working Conditions and Work Satisfaciton in General (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 339 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 340 Armin reautschnig Figure 11.3.06 Working Conditions and Work Satisfaciton According to Two Levels of Education (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 341 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 342 Armin reautschnig Figure 11.3.07 Working Conditions and Work Satisfaciton According to Two Age Groups (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 343 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 344 Armin reautschnig Figure 11.3.08 Working Conditions and Work Satisfaciton According to Gender (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 345 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 11.4 Trust in Institutions grAPHiCS 11.4.01 – 1.4.08 VAriABleS +2 A B C D E F +1 -1 -2 A Trust in countryś parliament B Trust in the legal system C Trust in the police D Trust in politicians E Trust in political parties F Trust in the European Parliament grAPHiCS 11.4.01 – 1.4.08 grOuPS: geNDer Age eDuCATiON a b c d e f g h i a Male c 15 – 29y b Female g Not completed primary education d 30 – 44y Primary or first stage of basic e 45 – 60y Lower secondary or second stage of basic f 60y + h Upper secondary Post secondary i First stage of tertiary Second stage of tertiary 348 Armin reautschnig Figure 11.4.01 Trust in Institutions in General (CEE) D PL SK CH A H SLO Static Patterns 349 Figure 11.4.02 Trust in Institutions According to Three Levels of Education (CEE) D PL SK CH A H SLO 350 Armin reautschnig Figure 11.4.03 Trust in Institutions According to Four Age Groups (CEE) D PL SK CH A H SLO Static Patterns 351 Figure 11.4.04 Trust in Institutions According to Gender (CEE) D PL SK CH A H SLO 352 Armin reautschnig Figure 11.4.05 Trust in Institutions in General (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 353 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 354 Armin reautschnig Figure 11.4.06 Trust in Institutions According to Three Levels of Education (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 355 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 356 Armin reautschnig Figure 11.4.07 Trust in Institutions According to Four Age Groups (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 357 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 358 Armin reautschnig Figure 11.4.08 Trust in Institutions According to Gender (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 359 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 11.5 Happiness, Life Satisfaction, Health and Social Capital grAPHiCS 11.5.01 – 11.5.08 VAriABleS +2 A B C D E F G +1 -1 -2 A Most people can be trusted or you can't be too careful B Most people try to take advantage of you, or try to be fair C Most of the time people helpful or mostly looking out for themselves D How satisfied with life as a whole E How happy are you F Take part in social activities compared to others of same age G Subjective general health grAPHiCS 11.5.01 – 1.5.08 grOuPS: geNDer Age eDuCATiON a b c d e f g h i a Male c 15 – 29y b Female g Not completed primary education d 30 – 44y Primary or first stage of basic e 45 – 60y Lower secondary or second stage of basic f 60y + h Upper secondary Post secondary i First stage of tertiary Second stage of tertiary 362 Armin reautschnig Figure 11.5.01 Happiness, Life Satisfaction, Health and Social Capital in General (CEE) D PL CZ CH A I SLO H Static Patterns 363 Figure 11.5.02 Happiness, Life Satisfaction, Health and Social Capital According to Three Levels of Education (CEE) D PL CZ CH A I SLO H 364 Armin reautschnig Figure 11.5.03 Happiness, Life Satisfaction, Health and Social Capital According to Four Age Groups (CEE) D PL CZ CH A I SLO H Static Patterns 365 Figure 11.5.04 Happiness, Life Satisfaction, Health and Social Capital According to Gender (CEE) D PL CZ CH A I SLO H 366 Armin reautschnig Figure 11.5.05 Happiness, Life Satisfaction, Health and Social Capital in General (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 367 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 368 Armin reautschnig Figure 11.5.06 Happiness, Life Satisfaction, Health and Social Capital According to Three Levels of Education (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 369 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 370 Armin reautschnig Figure 11.5.07 Happiness, Life Satisfaction, Health and Social Capital According to Four Age Groups (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 371 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 372 Armin reautschnig Figure 11.5.08 Happiness, Life Satisfaction, Health and Social Capital According to Gender (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Static Patterns 373 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 12 Dynamic Patterns Armin Reautschnig With this graphical prototype, the values of one or more variables for an entire population or multiple subpopulations for several points in time can be represented. Polygons are drawn with parallel horizontal edges, stemming from a vertical axis that forms the origin and corresponds to the processed value of the first point in time. The end points of the parallel edges correspond to the processed values of the second – upper edge – and the last – lower edge – time points in the series. In representations with a time series of t>3 points, additional processed values can be found between the parallel edges in the form of points, which are connected to the previous and successive points in the polygon. Figure 12.01 The Prototype for Dynamic Patterns The vertical spacing between two points corresponds to the smallest unit and its size depends on the number of the units to be illustrated as well as the available space in the graphic. The number of units is thus defined through the number of variables multiplied by the number of subpopulations and then multiplied by a factor equaling the number of points in time minus one and divided by two. The limitation on the represented units results from the 378 Armin reautschnig relationship of the available size and the readability index. Subpopulations are shown accordingly and depend on the type of sub-population like gender, age groups, education levels and the like. Both the subpopulations within the variables and the variable areas themselves are visually separated by blank spaces. The different color designations of the polygons are based on the polygons’ respective positions. The polygonal surfaces are formed by the vertical axis and all the edges that intersect with that axis. If this surface is found to the right of the origin, it is positive; if it is left of the origin, it is negative. The range of the linear scale here is also determined by the quantity of the maximum processed value. The scale gradation here is also based on integers. The conditions of this book and the requirements of representing a time series with t=3 points has led to the limitation of represented units to 28. A distribution of seven variables and four subpopulations has also been applied. The processed values depict the changes in the average for the time points t2 and t3 as they relate to the average of the first time point t1 in the time series. The polygons in the negative area are shaded dark grey, while those in the positive area are shaded light grey. 12.1 Changes of Trust in Institutions grAPHiCS 12.1.01 – 12.1.08 VAriABleS 2002 2004 A 2006 B C D E -2 -1 +1 +2 A Trust in countryś parliament B Trust in the legal system C Trust in the police D Trust in politicians E Trust in the European Parliament grAPHiCS 12.1.01 – 12.1.08 grOuPS: geNDer eDuCATiON a g h b i a Male g Not completed primary education b Female Primary or first stage of basic Lower secondary or second stage of basic Age h Upper secondary Post secondary c i First stage of tertiary d Second stage of tertiary ef c 15 – 29y d 30 – 44y e 45 – 60y f 60y + 380 Armin reautschnig Figure 12.1.01 Changes of Trust in Institutions in General (CEE) CH D PL A SLO H Dynamic Patterns 381 Figure 12.1.02 Changes of Trust in Institutions According to Three Levels of Education (CEE) CH D PL A SLO H 382 Armin reautschnig Figure 12.1.03 Changes of Trust in Institutions According to Four Age Groups (CEE) CH D PL A SLO H Dynamic Patterns 383 Figure 12.1.04 Changes of Trust in Institutions According to Gender (CEE) CH D PL A SLO H 384 Armin reautschnig Figure 12.1.05 Changes of Trust in Institutions in General (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Dynamic Patterns 385 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 386 Armin reautschnig Figure 12.1.06 Changes of Trust in Institutions According to Three Levels of Education (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Dynamic Patterns 387 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 388 Armin reautschnig Figure 12.1.07 Changes of Trust in Institutions According to Four Age Groups (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Dynamic Patterns 389 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 390 Armin reautschnig Figure 12.1.08 Changes of Trust in Institutions According to Gender (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Dynamic Patterns 391 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 12.2 Changes of Happiness, Life Satisfaction, Health and Social Capital grAPHiCS 12.2.01 – 12.2.08 VAriABleS 2002 2004 A 2006 B C D E F G -2 -1 +1 +2 A Most people can be trusted or you can't be too careful B Most people try to take advantage of you, or try to be fair C Most of the time people helpful or mostly looking out for themselves D How satisfied with life as a whole E How happy are you F Take part in social activities compared to others of same age G Subjective general health grAPHiCS 12.2.01 – 12.2.08 grOuPS: geNDer eDuCATiON a g h b i a Male g Not completed primary education b Female Primary or first stage of basic Lower secondary or second stage of basic Age h Upper secondary Post secondary c i First stage of tertiary d Second stage of tertiary ef c 15 – 29y d 30 – 44y e 45 – 60y f 60y + 394 Armin reautschnig Figure 12.2.01 Changes of Happiness, Life Satisfaction, Health and Social Capital in General (CEE) CH D PL A SLO H Dynamic Patterns 395 Figure 12.2.02 Changes of Happiness, Life Satisfaction, Health and Social Capital According to Three Levels of Education (CEE) CH D PL A SLO H 396 Armin reautschnig Figure 12.2.03 Changes of Happiness, Life Satisfaction, Health and Social Capital According to four Age Groups (CEE) CH D PL A SLO H Dynamic Patterns 397 Figure 12.2.04 Changes of Happiness, Life Satisfaction, Health and Social Capital According to Gender (CEE) CH D PL A SLO H 398 Armin reautschnig Figure 12.2.05 Changes of Happiness, Life Satisfaction, Health and Social Capital in General (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Dynamic Patterns 399 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 400 Armin reautschnig Figure 12.2.06 Changes of Happiness, Life Satisfaction, Health and Social Capital According to Three Levels of Education (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Dynamic Patterns 401 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 402 Armin reautschnig Figure 12.2.07 Changes of Happiness, Life Satisfaction, Health and Social Capital According to Four Age Groups (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Dynamic Patterns 403 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 404 Armin reautschnig Figure 12.2.08 Changes of Happiness, Life Satisfaction, Health and Social Capital According to Gender (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Dynamic Patterns 405 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 13 Complex Patterns Armin Reautschnig This special prototype provides an overview of the processed values of a large number of variables for a total population or multiple subpopulations for one point in time. The particular construction of the graphic enables the recognition of connectivity patterns between thematic areas and similarity/ dissimilarity patterns between regional units. This special prototype consists of rectangles that are ordered within a matrix. These rectangles represent the singe variables and are consolidated line-by-line into thematic blocks. The processed values of the variables are portrayed through different shades that the rectangles can take on. The shade corresponds to the rank of the value in the sorted distribution. The averages per variable per country are sorted in ascending order, divided into even parts, usually into the upper third (1), a medium range (2) and the lower third (3), and the corresponding ranks from 1 to 3 are assigned to each of the regional units. Each of these three ranks is then reproduced in a specific color. Figure 13.01 A Prototype for Complex Patterns 410 Armin reautschnig The transmission of information in this graphic is based on the recognition and differentiation of shade. Thus the space used for the representation of the processed values can be decreased, allowing the number of represented units to increase. The medium-dependent readability index for this graphic must be extended to include the factor of color resolution and the correct color reproduction. Though color is also important for the other prototypes, there it is really only a matter of improving readability. In this prototype, however, essentially the entire representation of information depends on color and the differentiation between colors. The impressive human ability to distinguish colors from one another, especially in direct visual comparisons, can, however, be thwarted by restrictions in reproduction possibilities. Increasing the readability index concerning the color factor helps to reduce this problem. With respect to the restrictions imposed by a book in black and white printing, the number of represented units was limited to 48. The matrix consists of six variables in four thematic blocks for two subpopulations. There are three shades of grey used for the representation. Thus the sorted values for the variables are divided into thirds. The lower third is symbolized by dark grey, the middle range by medium grey, and the upper third is shaded in light grey. 13.1 Working Conditions, Work Satisfaction, Trust in Institutions, Happiness, Life Satisfaction, Health and Social Capital grAPHiCS 13.1.01 – 13.1.04 VAriABleS grOuPS: geNDer A B C D E F aa 1 2 3 4 bb 1 Working Conditions a a Male Male A Allowed to be flexible in working hours b b Female Female B Allowed to decide how daily work is organised C Allowed to influence job environment D Allowed to influence decisions about work direction E Allowed to change work tasks F Get a similar or better job with another employer 2 Work Satisfaction A How satisfied with job B Satisfied with balance between time on job and time on other aspects C Find job interesting D Find job stressful E Become unemployed in the next 12 months, how unlikely F Get paid appropriately, considering efforts and achievements 3 Trust in Institutions A Trust in country's parliament B Trust in the legal system C Trust in the police D Trust in politicians E Trust in political parties F Trust in the European Parliament 4 Happiness, Life Satisfaction, Health and Social Capital A Most people can be trusted or you can't be too careful B Most of the time people helpful or mostly looking out for themselves C How satisfied with life as a whole D How happy are you E Take part in social activities compared to others of same age F Subjective general health 412 Armin reautschnig Figure 13.1.01 Working Conditions, Work Satisfaction, Trust in Institutions, Happiness, Life Satisfaction, Health and Social Capital in General (CEE) CH D PL A SLO H Complex Patterns 413 Figure 13.1.02 Working Conditions, Work Satisfaction, Trust in Institutions, Happiness, Life Satisfaction, Health and Social Capital According to Gender (CEE) CH D PL A SLO H 414 Armin reautschnig Figure 13.1.03 Working Conditions, Work Satisfaction, Trust in Institutions, Happiness, Life Satisfaction, Health and Social Capital in General (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Complex Patterns 415 GB N S FIN IRL NL DK PL B D E F CH A H P SLO 416 Armin reautschnig Figure 13.1.04 Working Conditions, Work Satisfaction, Trust in Institutions, Happiness, Life Satisfaction, Health and Social Capital According to Gender (Europe) GB N S FIN IRL NL DK PL B D E F CH A H P SLO Complex Patterns 417 GB N S FIN IRL NL DK PL B D E F CH A H P SLO Part III – The Morphological Road to Comparative Research Part III: The Morphological Road to Comparative Research Karl H. Müller | Niko Toš The two roads for comparative research so far are based on an assumption which is too obvious to be mentioned at all, namely on the assumption of directly comparable micro-data across nations. The data-sets for Part II, for example, used the European Social Survey (ESS) which is by far the most advanced cross-national survey with the highest quality standards for the inter-comparability of data. The seemingly trivial assumption of directly comparable data can be stated, inter alia, in the following way: A comparison between a data-set A and a dataset B, based on two different surveys A* and B*, can be undertaken because the survey designs for A* and B* used functionally identical elements not only for the questionnaire, but also for sampling, interview-modes, encodings of open variables, etc. For the first road to comparative research, the availability of practically identical ways of measurements and of data productions can even be considered as the necessary pre-requirement for any type of statistical comparative research work at all. In sharp contrast, the first article of Part III by Christian Bischof and Karl H. Müller is based on an outright violation of the assumption of directly comparable data sets. In this article, comparisons between a data-set A and data-set B will be constructed although no underlying surveys are available which have been conducted in a functionally identical manner. In fact, the morphological approach starts in a situation where, for example, two surveys from two different regions are available which share a broad topic like housing, health or aging, but which use mostly different components in the data production for this broad common domain. It must be added immediately that the differences between the data productions for these two regions are not relatively mild ones like in the case of different modes for conducting a survey ( e.g. , face to face versus telephone interviews) or in the case of different scales for an identical set of questions. The differences in the data production process for the two regions are rather wild ones, namely different sets of questions altogether. Even at second sight, meaningful comparisons cannot be based on data sets which, though situated in one or more common general domains, use different dimensions for each of these domains respectively. After all, the 422 Karl H. Müller | Niko Toš heterogeneity and non-comparability of national survey data sets has been a vital argument for new waves of European surveys like the European Social Survey (ESS) or the Survey of Health and Aging (SHARE). But exactly at this point, the morphological approach begins with its set of procedures and operations. In the final article of Part III an outline will be provided on emerging forms of a new theoretical background knowledge for empirical social research which should lead, on the one hand, to new methods for comparative research and, on the other hand, to a much deeper understanding of the cognitive intricacies and subtleties inherent in the social action and interaction processes during conventional survey interviews. Figure PArT iii.01 The Morphological Road to Comparative Research 14 The Third Road for Comparative Research: Morphological Designs Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš This article presents the outlines of a new road for comparative research which could become of special relevance for large quantities of unused data outside the current data bases for comparative analyses. These datasets have been produced and analyzed within a single region only and, so far, were of no use for comparative research due to the absence of functionally identical datasets in other regions or countries. Subsequently, these surveys which have been designed and assembled for the purpose of analyzing a single population only will be characterized as atomic survey-datasets.1 Essentially, atomic survey datasets are, due to the missing links to similar survey datasets in other areas, lost for comparative research. At this point, the morphological approach sets in, tries to create bridges between these different atomic survey datasets and attempts to bring them into the arena of comparative research. With a common class of data transformations which can be labelled as formation and as aggregation, the morphological approach generates a set of morphological constructs {MC } i which, despite the heterogeneity of the underlying atomic survey datasets, can be used for comparative analyses. Towards the end of this article, several different designs within the third road to comparative research will be described in greater detail. But the main thrust of this paper lies in a high-risk test of the morphological approach in order to establish its viability for comparative research. 14.1 The Traditional Limits for Comparative Research So far, atomic survey datasets, i.e. , directly incomparable micro-data between regions, were thought to be the strongest argument for the generation of comparable cross-regional survey programs like the European Social Survey (ESS), Eurobarometer or the World Value Survey. Usually, comparative research has reached its insurmountable limits if confronted with different questions from different surveys in different countries. As an example, take the following two questions from Table 14.01. Normally, the outcomes for Q ,(A, C ) cannot be compared with the results k 1 for Q ,(B, C ). Assume, that for Q ,(A, C ) two thirds of the female population, l 2 k 1 in contrast to only one third of the male group, have big difficulties in 1 Consequently, molecular datasets could be introduced as chains composed of identically generated datasets across regions like in the case of international survey programs (SHARE, ISSP, European Value Surveys, etc.). In the subsequent article the emphasis lies on atomic survey datasets. Mutatis mutandis, the same arguments and strategies could be used for panel datasets as well. 426 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš starting one’s own business whereas three quarters of the female group are rather satisfied with their workplace, compared with only 50% of the male respondents. Any inference from Q ,(A, C ) in country C1 with respect to l 1 the potential results for Q ,(B, C ) – and vice versa – would be ill-founded l 2 and, due to the absence of directly comparable data, would constitute a wild speculation without any empirical foundations.2 At this point, the morphological approach begins with its operations in creating links or bridges. But quite obviously, the morphological approach is not a magical procedure which can be used for any kind of heterogeneous datasets irrespective of the degree of data incomparability. After all, data on media consumption in country C will provide no comparative clues on the 1 health status of the population in country C . 2 TABle 14.01 Two Different Questions from Different Surveys in Two Different Countries Question Q from Survey A in country C [Q (A, C )]: Start one’s own business k 1 k 1 Satisfaction with the way things Question Q from Survey B in country C [Q (B, C )]: l 2 l 2 are handled at workplace Thus, the morphological approach rests itself on two assumptions which, a fortiori, are needed for the traditional ways of comparative research, too. The first assumption is one of at least second-best or satisficing practices and requires that the atomic surveys which have been conducted, fulfil the usual quality requirements for surveys in a sufficiently professional manner. In other words, the atomic survey datasets used for morphological analyses should be free of serious deficits, biases and errors with respect to sampling, interviewer accuracy, etc. The second assumption is content-oriented and requires a small number of shared domains between the selected atomic surveys. These shared domains are to be understood in a very general way and are fulfilled in all those instances where a uniform classification scheme CS can be built for each of the selected atomic surveys. What is needed, thus, for the morphological approach, are different atomic surveys on two, three or more shared domains like health, political attitudes or working conditions and the like. Violating even these two conditions simply means abstaining from any type of comparative data-analysis, morphological, traditional or otherwise. 2 The result can be generalized from the Q /Q case to the general Q /Q case where i and j k l i j run through the entire set of survey questions. 14.2 The Basic Steps in the Morphological Approach Figure 14.01 presents an overview for the third road to comparative research. The morphological approach starts with two or more different atomic surveys A, B … in different regional units like cities or countries. Here, the variables, items and scales between the surveys A, B … are usually heterogeneous, except for socio-demographic variables like age, gender or household size. Heterogeneity between surveys means that survey-questions are focused on different dimensions within a broader domain, that questions are phrased in significantly different ways within the same dimension and that the scales used for these questions differ substantially, too. Under normal circumstances, comparative research has come to an end in cases of strong survey heterogeneity.3 Figure 14.01 Morphological Designs However, the morphological approach starts with a set of atomic surveys A, B … and creates a sub-set of shared domains which can be described in a homogeneous fashion by a single classification scheme CS. This classification scheme with a small list of shared domains like aging, housing, quality of life and the like forms the basis for the morphological approach. 3 To repeat, once more, the morphological approach, as any other alternative, must come to an insurmountable barrier, once a survey on eating habits in country C is to be compared 1 with a different survey on professional training in country C .2 428 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš Following Figure 14.01, the essential steps in the morphologiocal approach are called formation and aggregation. Formation means a dual selection process. – The first selection is made with respect to a number of elementary forms which have been assembled in Figure 14.02. – The second selection concerns the number of dimensions used, given that the selction of elementary forms has been accomplished. Figure 14.03 highlights the distribution of forms with four dimension in each of the common domains. Figure 14.02 Elementary Forms of Shared Domains between Atomic Survey Data Sets More specifically, the formation process involves, first, the specification of a classification scheme CS for a group of atomic survey data sets. This classification scheme CS leads to a specific number N of homogeneous themes or domains which are shared by all atomic survey data-sets under consideration. Usually, a CS for atomic data sets includes broad domains ( e.g. , social capital, partnership and family, etc.) which have been covered by a number of different questions or item-batteries in each of the atomic surveys. The specification of a number of shared domains leads to the selection of a corresponding elementary form. Elementary forms F (i = 1, i 2, 3, …) are based on the combination of unit rectangles where each unit The Third road for Comparative research: Morphological Designs 429 rectangle represents a single shared domain between atomic survey data sets. These unit rectangles can be combined to elementary geometric constructs like triangles, squares, pentagons, hexagons, heptagons and the like. Figure 14.02 summarizes these elementary geometric forms F from a single area up i to a configuration of a nonagon with nine broad common domains among a set of atomic surveys. Figure 14.03 Elementary Forms F with i = 1, 2, 3 …, 9 Shared i,4 Domains and Four Dimensions Obviously, the next step consists in the selection of a common number of variables or dimensions F (d = 1, 2, 3, …) which are needed for the i,d subsequent creation of morphological constructs. Normally, several variables or dimensions are available for each of these shared domains. Depending on the selected atomic surveys the usual number of dimensions will be three, four or five and, probably seldom, higher. Numbers beyond seven or eight depend on the availability of suitable survey questions and lower numbers lead to difficulties for the morphological approach itself or to the purpose of comparative research altogether. Figure 14.03 demonstrates the variety of elementary forms for F (I = 1, i,4 2, …, 9) with four dimensions and similar diagrams could be produced for three, five or more dimensions as well. In terms of data operations, formation 430 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš means a data transformation from an atomic survey dataset with different scales to a sub-data set with a specific number of shared domains, a specific number of dimensions for each of these domains, and, most importantly, homogeneous scales across all dimensions and domains. Following Diagram 14.01, the next step in the morphological approach, after formation, is called aggregation which is directed towards a morphological construct MC for each of the atomic data sets. These morphological i constructs become, then, the basis for comparative investigations. Thus, based on a morphological construct for each atomic survey data set under consideration, the comparative analysis can be focused on the structure of the morphological constructs MCA,B… and, due to the availability of an identical set of variables for all atomic surveys, on the socio-demographic profiles of this morphological consruct. In terms of data operations, aggregation is focused, first, on the integration of dimensions and domains and, second, on the creation of a horizontal or vertical stratification of the sub-dataset which, then, forms the basis for comparative analyses. Due to the necessary selection and construction of forms inherent in this approach the terms morphology and morpholical approach have become the main classifiers for the third road to comparative reserach. Subsequently, a test design and a preliminary demonstration of the viability of the third road to comparative research will be laid out in greater detail. 14.3 A Test-Design for the Morphological Approach Within the present context, a test-design will be constructed which should help to demonstrate the usability and the relative robustness of the morphological approach. Following Figure 14.04, a single national survey will be selected and split into two different groups of questions A and B. It must be emphasized that the split is not one in populations as in split ballot designs, but a split in survey questions.4 The rationale for focusing on a single survey only is rather straightforward: By concentrating on a single population, by using two different subsets of questions from one survey and by performing the morphological approach with an identical population one can easily compare the two resulting 4 It should be added that the morphological approach builds on different sets of questions and not, as in the case of MTMM-designs (multi-trait multi-method) on slight vatiations of a single question. On MTMM in general, see Bagozzi/Yi, 1991, Bunting/Adamson, 2000 or Saris, 2003. The Third road for Comparative research: Morphological Designs 431 Figure 14.04 A Test-Design for the Morphological Approach stratifications with respect to their population composition. In this way, two groups of results could be obtained. – If the population is grouped in high numbers into the same categories by the formation and aggregation operations, despite two different sets of survey questions, then the morphological approach seems to be a viable path to follow and the title of Part III – the third road to comparative research – can be justified. – However, if the population is stratified across the two morphological constructs, following the necessary formation and aggregation operations, in a random manner, then the morphological approach should be abandoned altogether. (See, once again, Figure 14.04) 432 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš Turning to the test design in more detail, the starting point lies in the construction of a classification scheme CS which divides the survey into a small set of broad domains or general themes. As usual, a zero hypothesis H can be put forward which asserts the non-0 viability and the erroneous nature of the entire morphological approach. H : The morphological constructs used for comparative analyses cannot be 0 compared with each other in any meaningful way. The morphological data-constructs used for comparisons are, due to the absence of directly comparable data, artificial and spurious and lead to ill-founded results. In order to refute H , the validity and the reliability of the morphological 0 constructs MC must be shown in the following manner. H : The morphological constructs used for comparative analyses can be A compared with each other in many meaningful ways. The morphological data-constructs used for comparisons are, due to the absence of directly comparable data, both valid and reliable and comparisons lead to empirically well-founded results. A perfect support of the H -hypothesis would consist in a series of comparisons 0 between MCA and MCB in which each single categorization is misplaced. This configuration is shown in Figure 14.05 where one and the same population has been categorized in three layers or strata (low, medium, high). Due to the fact, that a single data set has been used, it can be determined how much the two constructs MCA and MCB differ from each other. As Figure 14.05 demonstrates this categorization has been undertaken in an entirely different way for the two morphological constructs MCA and MCB. Actually, not a Figure 14.05 A Perfect Failure for the Morphological Approach The Third road for Comparative research: Morphological Designs 433 single individual in MCA is placed at the same level in MCB. In the worst case of Figure 14.05, zero percent of the population have been categorized in an identical manner through the two morphological constructs MCA and MCB. Contrary, the perfect support for H and for the morphological approach is A depicted in Figure 14.06 where the two morphological constructs MCA and MCB produce an identical categorization for the entire population P. Here, all respondents within the morphological construct MCA can be found at the identical levels in the morphological construct MCB. Figure 14.06 An Optimal Result in Favor of the Morphological Design In the case of identical categorizations for an entire survey sample on the basis of two different sets of questions, one would have produced extremely strong empirical evidence for the construct validity of the morphological approach. In the next section, the numerical demarcations will be specified in greater detail which, quite obviously, must be situated somewhere between 0 (perfect misspecification) to 1 (perfect correspondence). Finally, given that this approach produces valid and reliable categorizations for two different sets of questions within a single dataset, one can safely infer that this approach works also in the case of two different surveys with different sets of questions across different populations. And this inductive inference can be made, due to the present test-design with a single population, free of any ecological fallacies. 14.4 Setting up the Test Using the Austrian data for the ESS as primary test dataset, a classification scheme CS with three broad domains has been specified. For each of these domains four variables or dimensions were available.5 The classification scheme CS was composed of – Social capital (trust in institutions, general trust) – Working conditions – Central actor-resources (income, education, life satisfaction, etc.) TABle 14.02 Sub-Survey A and Sub-Survey B from the European Social Survey (ESS) Survey A Survey B Trust in Institutions Trust in country‘s parliament Trust in the legal system Trust in the police Trust in politicians Most people can be trusted or you can’t be Most people try to take advantage of you too careful Making mind up about political issues Politicians in general care what people like me think Working Conditions Allowed to be flexible in working hours Allowed to decide how daily work is organized Allowed to influence job environment Allowed to influence decisions about work direction Allowed to change work tasks get a similar or better job with another employer Start own business Satisfaction with the way things are handled at workplace General Domain Household’s total net income, all sources Highest level of education Borrow money to make ends meet, difficult Feeling about household’s income or easy nowadays life Satisfaction Happiness Father’s highest level of education Mother’s highest level of education 5 For the test design any multiple thematic survey like, for example, the Austria Social Survey, conducted in 1986, 1993 and 2003, could have been chosen. The reason for selecting the European Social Survey was motivated by the fact that the ESS, due to its availability for twenty or so European countries, allows an identical replication of the morphological approach for the Austrian ESS data for the entire dataset as well. The Third road for Comparative research: Morphological Designs 435 Table 14.02 summarizes the two different groups of questions from the Austrian ESS questionnaire which were used for a differentiation into two different sub-surveys, namely into Survey A and into Survey B. The underlying elementary form F is represented by the right element in the first row of 3,4 Figure 14.03 which consists of three shared domains with four dimensions each. The procedures with a classification scheme CS and different dimensions are neither particularly spectacular nor is it clear how a new road to comparative research can be based on these elementary forms or dimensions, specified so far. But until now, only the first step in the morphological procedure, the formation step, has been performed. What is still missing is the second step which has been qualified as aggregation procedure. Figure 14.07 The Morphological Aggregation Procedure Thus, the crucial step for the morphological approach is the present one and it consists in a common transformation process which is depicted in Figure 14.07. Here, a common aggregation procedure for two elementary forms produces two morphological constructs MCA and MCB which, then, become the basis for comparative analyses between population A and population B. To obtain the morphological constructs MCA and MCB, the following aggregation steps have to be undertaken. First, each of the twelve dimensions in both data sets must be divided into three segments or strata, SU, SM and SL. The criterion for separation is distribution-dependent where SU occupies the upper third of the distribution, SM the middle range and SL the lower third of the distribution. Thus, single 436 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš dimensions like flexibility in working hours or the feeling about the household’s income may exhibit a distribution with a single peak, two peaks or several peaks. Independent of the shape of the distribution, the upper third of the distribution is classified as SU, the middle third as SM and the lower third as SL. Second, a data transformation of the actual values in the dataset into the domain of -1, 0 and +1 must be performed. The transformation rules are straightforward: SU equals 1, SM = 0 and SL becomes -1. Third, at this point each respondent in the data set is characterized by a string of varying numbers of -1, 0 and +1. Consequently, an aggregation procedure is needed which generates a single number for each respondent in the dataset. Figure 14.08 exemplifies the corresponding aggregation procedure for the twelve dimensions. From Figure 14.08 one can see that the total number of SL or the number of domains with a value of -1 are subtracted from the total number of SU or the number of domains with a value of +1. Thus, a respondent with seven SU-segments and two SL-areas receives the overall value of 5 or a respondent with two SU areas and eight SL domains ends up with a value of -6. Figure 14.08 A Transformation of the Data for Three Respondents with Twelve Dimensions into a Vertical Inequality Scale from -12 to +12 SU: White Squares; SM: Gray Squares; SL: Black Squares Fourth, a new scale can be constructed which, due to the twelve dimensions selected for the data sets A and B, ranges from -12 to +12. +12 corresponds to a homogeneous distribution of +1 only and -12 marks the opposite sequence of a string of -1 only. Thus, each respondent, due to the selection of twelve dimensions, is located at one of the discrete points in the scale between -12 and +12. The Third road for Comparative research: Morphological Designs 437 Fifth, the resulting distribution, due to these data transformations, must be a normal distribution with μ = 0 and σ = 1. This distribution is separated, once again, into three layers, namely into GU, GM and GL. GU comprises the upper third of the population, GM the middle layer and GL the lower third of the population distribution. Sixth, the separation of the two populations A and B into three layers GAU, GAM and GAL and GBU, GBM and GBL leads to the morphological constructs MCA and MCB. Seventh, the two morphological constructs MCA and MCB can then be compared with respect to their socio-demographic composition like gender and age distribution or other socio-demographic categories which are usually available in an identical fashion across surveys. In this way, two morphological constructs MCA and MCB have been generated which are based on sub-surveys from the ESS and, due to the overall test design, on the same survey population. 14.5 The Test-Results for the Morphological Approach The final step in establishing the morphological approach as a viable new path for comparative research consists in a comparison between the two morphological constructs MCA and MCB. Since the ESS has been split into two sets of questions with an in identical population for each ESS-sub-class the crucial criterion lies, quite obviously, in the retention rate r for the groups GU, GM and GL. The retention rate r measures the number of identically classified instances in the survey population P and focuses on the similarities or dissimilarities in the composition of GAU, GAM and GAL on the one hand and of GBU, GBM and GBL on the other hand. Three a priori generalizations on the viability of the morphological procedure can be formulated right away. – First, the higher the retention rate between the two group stratifications of the morphological constructs MCA and MCB, the higher the construct validity of MCA and MCB and, thus, of the morphological approach. – Second, the higher the number of successful test instances with split surveys and single populations, the higher the reliability of the morphological procedure. – Third, the higher the construct validity and the reliability, the more the morphological approach can be applied to instances with two, three or more different surveys either across time or across regional units. 438 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš The retention rate r can be formally introduced as (∑n ) r = i (∑n +∑n ) i k where the expression in the numerator sums up the common instances between survey A and B and the denominator is composed of the sum of common elements plus the sum of misclassified respondents.6 The expected value for the retention rate r for a purely random classification method of an initial configuration of three groups [g , g , g ] into a new 1 2 3 schema with three groups [g , g , g ] is, quite obviously r = 1/3. 7 1* 2* 3* A priori, the following assessments can be undertaken with respect to the retention rate r from a value close to 0 up to a value very close to 1. A value of r ≈ 0 like in Figure 14.05 would exhibit an endemic misclassification of the morphological procedure altogether. In this case, the morphological procedure offers an unintended side-effect. Due to the non-random nature of a complete misspecification additional considerations could be devoted to the problem whether a configuration with a highly significant negative result could be transformed into one with highly significant positive results. A value well below r < 0.33 would indicate a misclassification bias of the morphological approach. Since the expected value for a purely random classification procedure lies at r = 0.33, it can be safely assumed that a significantly lower value for the retention rate r implies a systematic bias inherent in the morphological constructions. Values of r > 0.33 but below r ≤ 0.50 would show that the morphological approach performs slightly better than a pure random classification but is of no use for comparative research since the failure rate is sufficiently large and the margin of errors is simply too wide. The most interesting and relevant domain of the retention rate lies in the range 0.50 ≤ r < 1. Here, the morphological approach operates in a highly significant manner well beyond a pure random classification procedure. Moreover, the retention rate is sufficiently high to be qualified as pattern preserving. Pattern preservation can be defined as a data transformation D → D* which keeps general data patterns like a rising or a falling trend, clusters or characteristic relations largely intact so that D* exemplifies these patterns, albeit with some variations, as well. Especially interesting are retention rates near or above 0.67 6 A retention rate can be calculated in two ways, namely first with the population of Survey A as origin and the population of Survey B as successor and, second, with population B as origin and A as successor. 7 Given a configuration of three groups g , g , g , one has a total of nine possibilities with 1 2 3 three rank preserving specifications [g |g ; g |g ; g |g ] and six possible misspecifications. 1 1* 2 2* 3 3* The Third road for Comparative research: Morphological Designs 439 because this would show that the morphological approach produces very powerful classification preserving results. The value of r = 1 or r sufficiently close to 1 would show that both morphological constructs, despite their different underlying dimensions and data-bases, yield the same results. In this case, the morphological approach would operate almost in a purely magical way because it would generate, despite an underlying heterogeneity, (almost) perfect homogeneity. Before going into the issues of construct validity, reliability and retention rates, Table 14.03 presents a summary of the correlations between each pair of the dimensions used for Survey A and Survey B. This overview is useful because it demonstrates that the constitutive pairs of dimensions from Survey A and from Survey B are not correlated in a uniformly high manner. On the one hand, each of the three groups contains at least a pair of rather weakly correlated dimensions. Second, the correlations for the three groups themselves, specified as average correlation for all four pairs, differ significantly as well. The pairwise correlations in the domain of trust in institutions and in people turn out to be the highest (0.54) and they fall considerably in the area of working conditions (0.38). Turning now to the decision problem of accepting H or H , the following 0 A procedures and rules can be established. With respect to the construct validity, two extreme cases can be distinguished. Assume a single dataset divided into two sub-sets A and B with n (n = 1, 2, …, N) common domains and m (m = 1, 2, …, M) indicators or dimensions for each of the domains respectively where n domains x m dimensions A,B A,B are used as the empirical basis for comparisons. Assume, furthermore, a morphological formation and aggregation procedure that leads to two final morphological constructs MCA and MCB which separates the population of A and B in o (o = 2,3,…, O) categories each. Then, H can be rejected if A,B 0 two independent tests on construct validity and reliability are passed. Construct validity can be assumed in the case of 0.50 ≤ r < 1 simply because the retention rate lies within a domain far away from a random classification. Turning to the reliability of the morphological approach, it can be assumed that the reliability increases inductively with the number of successful test applications. In the present context, reliability will be taken into account only in a weak way by applying the morphological test design for ESS data from Austria and for the entire ESS country group as well. Table 14.04 offers a first overview on the results of the morphological procedures and on the overall retention rates. The retention rates have been calculated for the Austrian data set of the ESS as well as for the entire ESS-data set across all countries. In this way, the calculations both for a single 440 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš country as well as for the entire set of ESS-countries should demonstrate that these retention rates are not the result of particularly happy circumstances within a single country, but demonstrate a general feature of the entire ESS-data set. Moreover, using ESS-data for Austria as well as for the entire group of ESS-countries provides at least a weak hint on the reliability of the morphological approach. TABle 14.03 Correlations between the Dimensions in Survey A and in Survey B (ESS)* Survey A/Survey B (ESS) Domain I: Trust in Institutions and in People Trust in country‘s parliament / 0.56 Trust in the legal system Trust in the police / 0.43 Trust in politicians Most people can be trusted or you can’t be too careful / 0.58 Most people try to take advantage of you Making mind up about political issues 0.60 Politicians in general care what people like me think Average Correlation for Domain i 0.54 Domain II: Working Conditions Allowed to be flexible in working hours / 0.56 Allowed to decide how daily work is organized Allowed to influence job environment / 0.62 Allowed to influence decisions about work direction Allowed to change work tasks / 0.24 get a similar or better job with another employer Start own business / 0.09 Satisfaction with the way things are handled at workplace Average Correlation for Domain ii 0.38 Domain III: General Resources Household’s total net income, all sources / 0.17 Highest level of education Borrow money to make ends meet, difficult or easy / 0.41 Feeling about household’s income nowadays life Satisfaction / 0.68 Happiness Average Correlation for Domain iii 0.49 Average Correlation across Domains 0.47 * The correlations have been calculated for the entire eSS-data set in round 1 and not for the Austrian data. The Third road for Comparative research: Morphological Designs 441 The most important result from Table 14.04 lies in the fact that the retention rates lie well in the critical and pattern-preserving range of 0.50 ≤ r < 1. As a rule of thumb, the retention rates come close to the value of r ≈ 0.67 which means, essentially, that approximately two thirds of the population have been classified in a rank-preserving way. A second essential outcome lies in the overall consistency of retention rates for a single broad domain or across the three shared domains. This result is particularly encouraging because it demonstrates the applicability of the morphological approach already for elementary forms of the variety F , 1, 4 F , F , … . 1, 5 1, 6 TABle 14.04 Retention Rates for Survey A and Survey B (ESS) (in %) Retention Rates Austria Europe A/B B/A A/B B/A Retention Rate for Domain I (Social Capital) lower group 57.5 65.3 70.6 66.2 Middle group 39.8 44.1 46.0 46.7 upper group 74.8 59.3 64.5 68.3 Retention Rate for Domain II (Working Conditions) lower group 66.3 71.4 66.6 66.5 Middle group 65.4 55.3 63.0 58.1 upper group 55.0 67.5 57.1 65.6 Retention Rate for Domain III (General Resources) lower group 59.3 42.2 80.2 41.0 Middle group 48.0 47.6 47.8 33.9 upper group 54.2 67.0 33.2 86.0 Overall Retention Rate across Domains lower group 54.3 72.8 76.6 68.0 Middle group 48.3 34.3 53.3 36.3 upper group 69.4 71.7 59.5 87.4 Average Retention Rate across Domains 57.3 59.6 63.1 63.9 Finally, the third result points to the consistency between a single national survey and a large group of national surveys which offers at least a weak support for the reliability of the morphological approach. In sum, with retention rates well around r ≈ 0.67 and, thus, a relatively high construct validity plus an a priori reliability of the morphological procedure due to the similarity between the Austrian and the European retention 442 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš rates, H can be safely rejected and H accepted. These results point to the 0 A overall viability of the morphological approach and for using morphological constructs as units of comparisons in the case of atomic data sets. Moreover, using atomic surveys and splitting them into sub-atomic parts with identical populations should become the indispensable testing ground for exploring and eliminating potential weaknesses in the morphological construction processes. In principle, exploring the test-design with single populations should lead to two forms of potential cognitive gains. On the one hand, this testing ground of split surveys and identical populations for each sub-survey should contribute, in the long run, to the development of positive and negative heuristic rules for the formation and aggregation procedures of the morphological approach. For example, the third common domain for the ESS with its focus on general resources turned out to be the one with the comparatively lowest retention rates. Looking more closely on the selected dimensions and on the construction process, it was rather difficult to split the two education variables into three equal segments. Although the ESS did not offer another pair of dimensions in the shared domain of general resources, it becomes possible to identify the weak components in the formation and aggregation process. Thus, there is a positive learning curve inherent in the morphological approach which has not been unfolded and which stays currently almost in its initial configuration. On the other hand, performing the morphological approach for a single dataset should produce a lot of new explicit and implicit knowledge with respect to superfluous and redundant survey questions. High retention rates for two subsets of questions within a single survey implies that one of the subsets can be replaced by an entirely new domain and by a new set of questions and dimensions altogether. 14.6 Different Trajectories within the Third Road of Comparative Research Having established, in principle, the overall viability of the morphological approach, one can specify a large number of different designs within the morphological framework. Specific research interests and research questions require different morphological constructs as well as different forms of comparative analyses morphological style. At the outset, the morphological approach is capable to cope with a large variety of formal structures in terms of the underlying composition and of spatio-temporal levels. The Third road for Comparative research: Morphological Designs 443 On the one hand, the morphological approach can be applied to different populations, be they composed of persons, households, enterprises, cities, nations and the like.8 In compositional terms, this framework is not restricted to specific elements like persons alone, but can be used for different segments of contemporary societies as well like non-government organizations, research institutes, internet communities, etc. as well. On the other hand, this approach can be used for different spatio-temporal levels as well, ranging, spatially, from neighborhoods to cities, regions, countries, supra-national regions up to the global level and temporarily from daily variations, monthly changes, yearly developments up to the slow transformations of the longue durée. Different Morphological Constructs Morphological constructs can be built in two different ways, namely, like it has been done in the test-deign, in a vertical way or in a horizontal manner. The selection of vertical or horizontal constructs depends, on the one hand, on the overall research question as well as on the goal of a comparative analysis and, on the other hand, on the available classification scheme CS and on the dimensions selected for the morphological construct. In some applications it might turn out that vertical (horizontal) constructs should be specified, but no appropriate CS with a small number of vertical (horizontal)dimensions can be found within the available surveys. Vertical Constructs Morphological constructs as vertical formations reflect, very generally, underlying distributions of socio-economic inequality. For vertical constructs, classification schemes should cover essential domains of societal inequalities and dimensions should be chosen in a way that their lower ends can be interpreted in one of the following ways: as socio-economic risks, as being in a highly under-privileged position, as very disadvantageous, as very marginalized and the like.9 Here, the lower and higher ends of the various dimensions must constitute an inequality relation. For example, the dichotomy of very high socio-economic risks and very high socio-economic life chances can be viewed 8 It must be stressed that the morphological approach has been applied already to different sets of actors like firms, scientific institutes, state agencies and the like. As an example of a morphological analysis for scientific institutes, see Müller et al. , 2002. 9 Likewise, the upper ends should be interpreted in a number of different ways, too, like socio-economic life chances, very over-privileged, very advantageous, very central and the like 444 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš as one of the possible interpretations for the two end points of the scale. In general, the road with vertical morphological constructs should be selected whenever the research interest is directed towards societal inequalities and towards vertical stratifications. Looking at Table 14.02, for instance, many, if not all, of the dimensions bear strong relations with societal inequality. For instance, income, level of education, but also different forms of autonomy at the workplace can be seen as important aspects of societal inequality. Similarly, a high level of trust in institutions or in one’s social environment can be viewed as an individual resource whereas low levels of trust exclude a person from many routines and practices and can be viewed as a lack of social capital. Due to the selection of dimensions in Table 14.02 with clear relations to societal inequality, the road with vertical morphological constructs has been chosen. Figure 14.09 A Data-Transformation for Three Survey Respondents with Twelve Dimensions into the Domain of -1, 0 and +1 LC R POS i i i White Squares: Life Chances; Gray Squares: Medium; Black Squares: risks POS ; Position of respondent i; LC: Life chances of i; R: Socio-economic risks of i i More specifically, the vertical formation procedure consists, first, in a data transformation for each of the selected dimensions from the existing variety of different scales and values to a single scale only. In principle, different options are available like a continuous scale between 0 and 1 or various discrete linear scales. The question of scaling will be discussed in greater detail in section Non-Linear Scales. In the test-design, a particularly useful transformation has been employed which was based on the underlying distribution in each of the dimensions and which consisted in a transformation of the distribution of a single variable The Third road for Comparative research: Morphological Designs 445 or dimension into the domain of -1, 0 and +1. Here, a single dimension like the “health status” on a scale from 0 to 10 was transformed into the [-1, 0, +1]-scale in the following distribution-dependent way. The lower third of the distribution SL received the value -1, the upper third SU the value +1 and the middle segment SM the value 0. Likewise, income, measured in income groups or levels of education, measured in years, were be transformed in the same way. The lower third of the distribution received the value -1, the upper third is encoded as +1 and the medium segment gets, once again, the value 0. Figure 14.9 demonstrates, once again, a possible distribution for three survey respondents, based on a classification scheme CS with a total of three shared domains and four dimensions each. Figure 14.10 The Vertical Morphological Aggregation Procedure Due to the overall aggregation and to the summation of socio-economic life chances minus the number of socio-economic risk position each individual received a single value in the domain of –N to +N where N stands for the total number of dimensions. For example, in the test example, each respondent could be characterized by a single value in the domain from – 12 to +12. Due to the aggregation process, the overall outcome is a normal distribution which can be separated, once again, in three groups, namely into GU, an upper group, GM, a medium group and GL, a lower group. (See also Figure 14.10) 446 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš In terms of comparative analysis with two atomic surveys, the minimal instance for a morphological design with vertical dimensions and scales consists of two morphological constructs MCA and MCB which serve as units of comparative analysis. (See also Figure 14.11) Figure 14.11 Two Vertically Distributed Morphological Constructs A and B Additionally, this type of aggregation offers an interesting way for horizontal stratifications especially around the values of -1, 0 and +1, as can be seen from Figure 14.12. Horizontally, the value 0 can be reached by twelve medium positions or by six risk-positions and six positions of socio-economic life chances. Again, empirical analyses have shown very interesting discrepancies between groups with a high number of medium positions versus groups with a relatively large number of positions of risks and life chances simultaneously.10 10 For more details, see Müller/Link, 1997. The Third road for Comparative research: Morphological Designs 447 Finally, the morphological constructs can serve as units for comparative analysis in a number of ways. First, the morphological constructs MCi can be compared with respect to their socio-economic attributes like gender distribution in each of the groups, their age distribution, their qualification levels, etc. In principle, the entire set of socio-demographic dimensions available across surveys can be used for comparative investigations. Figure 14.12 Searching for Different Horizontal Configurations within a Vertical Morphological Construct +N 0 -N Horizontal Scale (Number of Potential Indifference Positions) Second, a new platform for comparisons is composed by the dimensions of the shared domains. It will become a fascinating research field in itself to determine the conditions and requirements for using a single dimension from Survey A as a predictor for the corresponding distribution in the survey population B. Differences in the socio-demographic composition between population A and B can be used as weights. The most challenging issue lies in the determination of degrees of semantic homogeneity between the 448 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš dimensions which have been selected for the shared classification scheme CS. In the case of high semantic homogeneity like in the sets of questions on aspects of work satisfaction or trust in different institutions it might well be that the empirical distributions for the population A, with appropriate weightings, can be used for population B as well. In this sense, a powerful link can be created between atomic datasets which, so far, lay outside the domain of comparative research. Figure 14.13 A Transformation of Three Data-Sets with Twelve Dimensions into a Horizontal Life-Style Scale from -12 to +12 Horizontal Constructs Morphological constructs can be built as horizontal configurations, too. Horizontal constructs correspond to different life-styles11 or to different forms of life (Ludwig Wittgenstein) which should not be interpreted in terms of societal inequality but as horizontal stratifications. Take, for example, a classification scheme CS which consists of leisure activities, cultural and artistic preferences and of consumption styles. Moreover, the selected dimensions are composed of the frequency of visiting a theater or an exhibition of contemporary art or the importance of contemporary design in the purchase of consumer durables, etc. Here, the lower and the upper end-points of the selected dimensions could be interpreted in a number of different ways like traditional versus innovative, simple versus complex, etc. In contrast to the vertical constructs with a homogeneous interpretation in terms of societal 11 For an overview of life-style approaches, see Dangschat/Blasius 1994, Otte, 2004, Richter, 2005, Rössel, 2005 or Schwenk, 1996. The Third road for Comparative research: Morphological Designs 449 inequality, horizontal constructs and horizontal stratifications are much more heterogeneous and one has to be very explicit with respect to the classification scheme CS and the interpretation for the selected dimensions. Figure 14.13 and Figure 14.14 exhibit the aggregation process using horizontal constructs. Figure 14.14 A Normally Distributed Aggregated Morphological Construct on Socio-Economic Life-Styles Quite obviously, horizontal constructs and horizontal group stratifications become the new units of comparative analysis where major comparative issues lie, once again, in the socio-demographic group composition of MCA and MCB and in the weighting and adjustments procedures for the comparability in the dimensions of the shared domains. Non-Linear Scales So far, the underlying scales were linear and discrete. But Table 14.5 shows that a wide variety of scales could be used for the formation and aggregation procedures. TABle 14.05 Different Types of Scales in the Morphological Approach Linear Non-Linear Discrete Type i Type ii Continuous Type iii Type iV Overall life satisfaction, for example, if measured on a scale between zero and ten exhibits only a small number of respondents with values of five and smaller and, once again, a very small number in the highest category of ten. It is questionable whether the distance between 9 and 10 is the same as between, say, 6 and 7 or 0 and 1. Here, non-linear forms of scaling of Type II could be used as a substitute. In other instances, logarithmic scales could turn out as a useful way for scaling, especially in cases where the dimensions are scaled 450 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš according to large quantities like the amount of household income, square meters for housing, or years spent in of education, etc. Non-linear scales could give rise to wild vertical or horizontal constructs as well where the separations for different groups are not made on the basis of the overall population distribution, but on the scale-values, with very few instances in the upper range, a relatively small amount of cases in the middle range and the great majority in the lower group. In this way, a wide variety of procedures – vertical, horizontal, different types of scales – become available for obtaining the morphological constructs MC .i Designs for Morphological Comparisons Morphological constructs, if produced in a homogeneous and consistent manner for a number of different atomic surveys, become the new units of comparative analysis. Most importantly, a large number of different research designs can be created on the basis of morphological constructs as the subsequent explorations will demonstrate. Comparisons for a Single Unit across Time The first morphological design builds comparisons across time within a single region, using a set of different surveys {S } where t lies in the interval from i, t t , the initial period, to t , the final period.. This design is particularly useful 0 k for the creation of time series on vertical formations like risk groups or on horizontal strata like different life style groups. This design is particularly relevant where only a small number of identical surveys is available within a single region and the available survey replications are many years apart from one another. Here, a number of alternative surveys for each of the missing years could be selected which could fill at least partially the gaps between the missing years. The design for a single unit across time requires the usual steps in the morphological approach, namely the specification of an appropriate set of common or shared domains, the formation of dimensions and their aggregation along vertical or horizontal scales. In this way, time-series on the development of socio-economic risk groups, for example, can be created and can be analyzed in depth with respect to their changing socio-demographic composition, etc. Moreover, these time series can be compared with time series which result from a focus on the available identical surveys only and which use linear extrapolations for the missing years. The Third road for Comparative research: Morphological Designs 451 Comparisons for Multiple Units for a Single Point in Time The second morphological research design creates comparisons with a larger number of spatial units, be it cities, regions or countries, for a single point in time. Here, different surveys with a significant number of shared or common domains are to be selected, first, and a classification scheme CS must be constructed. Then, the different formation and aggregation steps have to be performed, dependent on the overall comparative research interest, on the classification scheme CS and on the interpretation of the selected dimensions. This instance is of particular interest in the case of cities and of sub-national regions which over the last years have accumulated a large number of atomic surveys which have been designed with very little consideration of comparative research interests. Here, the morphological approach offers a valuable instrument to link these atomic surveys and to create new units of comparison, either in vertical or in horizontal forms. Figure 14.15 The N x T Scope for Morphological Comparisons Comparisons for Multiple Units across Multiple Points in Time Finally, the third morphological research design is aimed towards comparisons with an increasing number of spatial units and time points. This approach is the most demanding one and requires a large number of atomic surveys and a careful specification of a common classification scheme CS. However, the 452 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš potential gains of this design, once it has been successfully implemented, are unusually high since it leads to comparisons of, say, socio-economic risk groups across a number of spatial units and across time. Figure 14.15 summarizes the different research designs within the morphological perspective. Figure 14.15 can be easily divided into four quadrants which correspond to different designs laid out so far. The first quadrant in the lower left area is occupied, essentially, by the test design of the morphological approach where one operates with a single unit, with identical populations and with a single point in time. The second quadrant in the lower right side is occupied by the design with single spatial units and multiple points of time. The third quadrant in the upper left side is reserved for the design with multiple units and a single point in time. Finally, the upper right hand quadrant uses multiple units with multiple points in time. 14.7 Bridges from the Morphological Approach to Areas outside the Social Sciences The concluding section will build two broad bridges from the morphological approach to domains outside empirical social research and even outside the social sciences, namely to the cognitive neuro- sciences on the one hand and to the area of bio-medical and health research on the other hand. These two bridges provide additional cognitive support for the morphological approach and should offer new incentives for pursuing and intensifying this type of morphological analysis. The Bridge to the Cognitive Neuro-Sciences The first bridge crosses the great divide between the morphological approach in social research and the fields of the cognitive neuro-sciences. With the first bridge, more cognitive support can be offered for the assumed equivalence of morphological constructs despite their different underlying dimensions. The first bridge is built with the help of a well-known principle in the cognitive neuro-sciences which has been labeled as the principle of undifferentiated encodings (PUE). PUE has been postulated already by Johannes P. Müller in 1826 and has been put forward in recent decades again and again by Heinz von Foerster [ e.g. , Foerster, 2002, Foerster/Glasersfeld, 1999] or Ernst von Glasersfeld [Glasersfeld, 1997]. In order to describe PUE at greater length, it can be divided into three parts. The Third road for Comparative research: Morphological Designs 453 Undifferentiated Signaling: The first part concerns the sensory border between actors and their environments and asserts that only undifferentiated quantities at different levels of intensities, but not qualitatively differentiated signals cross the border between the external and the internal worlds of actors. Thus, sensory inputs enter as undifferentiated quantities and not in differentiated qualities. The response of a nerve cell does not encode the physical nature of the agents that caused its response. Encoded is only ‘how much’ …, but not ‘what’. [von Foerster, 2002:215] Discrete Signaling: The second PUE-part stresses the fact that the quantities are not encoded in a continuous fashion, but in a discrete manner. Signal Integration: Finally, the third PUE-element emphasizes the necessity for an overall integration of sensory inputs into an overall pattern or result. Figure 14.16 The Elements of the Principle of Undifferentiated Encoding (PUE) At first sight there are no family resemblances in sight which could lead from the principle of undifferentiated encoding within the cognitive neuro-sciences to the morphological approach. Upon second thought however, the following methodological links can be identified which help to cross the distance between PUE in the cognitive neuro-science domains on the one side and the morphological approach in empirical social research on the other side. 454 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš Due to PUE and the emphasis on undifferentiated signaling a strong support can be offered for the functional equivalence of different dimensions within an identical broader domain. Surveys, by necessity, are capable of selecting only a small fraction of the daily routines and of their impact on the preferences and evaluations of individuals. Likewise, surveys are capable of identifying individual coping capacities in a highly selective manner, too. Thus, a single survey offers a highly selective minimal model of the overall stream of routines and of the cognitive evaluations of actors. At this point, it is important to note that the two procedures in the morphological approach, namely formation and aggregation, correspond with PUE in all its three aspects. Undifferentiated encoding is used by treating different dimensions within a common shared domain as equivalent. Discrete signaling is accounted for by building a discrete quantitative scale for qualitatively different inputs. Signal integration occurs in the morphological approach by using a special algorithm for all the aggregation of selected survey dimensions. Thus, a single morphological construct MCA can be interpreted as a minimal model which fulfills the PUE-requirements. Due to the practical infinity of potential dimensions within a set of common domains, a morphological construct MCB, using different dimensions, leads to a minimal PUE-compliant model as well. Due to PUE, both morphological constructs can be considered as equivalent in terms of their overall neuro-cognitive repercussions and implications. Both morphological constructs offer a functionally equivalent account of the overall neuro-cognitive pattern of undifferentiated encodings, discrete signaling and the signal integration between the cognitive-neural organization of actors and their socio-economic environments. The Bridge to Bio-Medical Research The second bridge leads from the vertically oriented side of the morphological approach to bio-medical and health research. Here, one can find a common deep-level language through which the domains of vertical socio-economic dimensions, inequalities as well as health can be expressed in a homogeneous vocabulary of stressors and of neuro-immunological processes. Initially, it is useful to start with a taxonomy of different types of stressors which can be found within the relevant body of literature.12 Here, one is confronted with a heterogeneous set, comprised of sensory stressors (strong light, noise, sensory deprivation, etc.), block-stressors (preventing essential routines like eating, sleeping, social contacts, etc.), achievement stressors 12 See on this point see Cooper, 1996, Horwitz/Scheid, 1999, or Sarafino, 2002. The Third road for Comparative research: Morphological Designs 455 (tests, examinations, work-tasks, but also monotony at work, etc.), social stressors (large crowd of people, loneliness, isolation, etc.), environmental stressors (noise, pollution, toxic materials, etc.), decision-based stressors (goal conflicts, quick decisions, but also lack of decision-making, etc.) or future-based stressors (fear, anxiety of the future, etc.) Seemingly, the heterogeneity of stressors is accompanied by a heterogeneity of stress reactions which vary in time (minutes, hours, days, weeks …), in intensity or in emotions, associated with each stress reaction. Nevertheless, common to all these stress reactions is an attempt to reduce the discrepancy between the effects of stressors and internal target values. Moreover, all stress reactions involve the activation of the hypothalamus-pituitary-adrenal axis and produce comparatively high quantities of endocrine hormones, particularly corticosteroids, with cortisol as the most important one, and catecholamines. Likewise, all physiological reactions to stress manifest themselves in a broad range of measurable changes like a higher production of stress hormones, higher degrees of blood pressure, heart rate, respiration rate, galvanic skin responses or in larger amounts of free fat acids. The general pattern of stress responses possesses at least two main connections to the domain of sickness and ailments, namely through their direct effects on the cardiovascular system on the one hand and through their immediate impact on the immune system on the other hand. With the short background on stress-research, it appears plausible to create a bridge from societal inequalities to special classes of stressors like social, environmental, future-based or decision-based stressors. Here, the following subset-relation can be put forward: Lower Segments SL of Dimensions of Societal Inequalities ⊂ Stressors It is quite obvious that this subset-relationship needs a very detailed justification which cannot be provided within the framework of the present article. Five main arguments can be given, however, which should offer some plausibility for a sub-set relation between SL, the lowest one third in the different dimensions of societal inequalities and stressors. – First, SL-positions, which can be specified in a wide array of living and working conditions, are characterized, inter alia, by their relative permanence. Thus, many of the SL–parts of socio-economic inequality dimensions like low, insufficient or deteriorating incomes or low degrees of qualifications are to be classified as long-lasting or, like in the case of low qualifications, as (nearly) permanent. Thus, being positioned in the SL–parts normally acts as a continuous stressor and not as a single, rare or isolated occurrence. 456 Christian Bischof | Vlado Miheljak | Karl H. Müller | Niko Toš – Second, there exists a remarkable asymmetry between the language of societal inequality, in particular the distinction between socio-economic risks and life chances on the one hand and the physiological stress language on the other hand. While socio-economic risks and life chances have been introduced symmetrically, no symmetry can be identified for the stress domains. Feeling unsafe in the public sphere (socio-economic risk/stressor) does not have a life chance corollary in terms of stressors. Feeling safe in the public domain does not constitute an alternative source for stressors. Likewise, a noisy environment at the workplace or at home implies an essential socio-economic risk and at the same time an environmental stressor whereas a quiet atmosphere at work or at home cannot be associated with a different group of stressors. Thus, socioeconomic risks can be linked to stressors, socio-economic life-chances imply, by and large, the absence of stressors. – Third, the distribution-dependent specification for thresholds for the SL–parts provides additional support for the subset relationship between the SL–areas of dimensions of societal inequality and stressors. Since the majority of the population is, by definitional necessity, above the SL-threshold, individual actors, falling in a specific SL–part, perceive themselves usually relatively deprived. Thus, the available literature on the importance of relative deprivation13 can be added as further evidence for the proposed SL–part-stress linkages. – Fourth, while stress reactions vary in length, intensity and emotional involvement, the basic physiological reaction patterns are unspecific with respect to the sources of stress. In other words, one does not find a “bad boss-stress reaction”, confined to a specific region in the neuro-immune system in contrast to a “loud noise-stress reaction”, affecting other parts of the neuro-immune system. Thus, a multi-dimensional array of essential living conditions across the contexts or settings of actors and across their cognitive-emotional organization can be interpreted as a summary of all relevant potential stressors whose scope and degree of completeness is limited by the restrictions inherent in conventional survey research only. – Fifth, stressors and stress reaction are clearly not invariant to the actual number of stressors since stress reactions are functionally related, probably in a complex and non-linear manner, to the overall number of stressors. This, in turn, provides additional support why the morphological aggregation procedure should be interpretable in terms of a net value for the overall number of socio-economic stressors. 13 On this literature, see more recently Walker/Pettigrew 1984 or Olson/Hafer 1996. The Third road for Comparative research: Morphological Designs 457 – In this way, a second bridge has been built which leads from the morphological approach to biomedical and health research and back. And this, in turn, leads to the question whether these two bridges from empirical social research to the life science, broadly conceived, can be viewed as a happy coincidence or whether one can identify more general reasons for the family resemblance between a new road for comparative research and the cognitive neuro-sciences or biomedical research. This question, inter alia, stands in the center of the next article. 15 Towards New Frontiers in Comparative Survey Research Karl H. Müller | Niko Toš The final chapter will turn its focus on the widening horizons and on the emerging cognitive landscapes for new ways and directions in comparative survey research.1 Moreover, these new territories and frontiers for comparative survey research will add more cognitive weight to the procedures and rules of the morphological approach which was presented in the previous chapter. This article will start with some general remarks on the increasing pressures for comparative survey research2 and will continue with a basic phase transition in the cognitive organization of science, namely the transition from Science I to Science II. The subsequent parts of this article deal with the repercussions and consequences of this phase transition for the social sciences in general and for comparative survey research in particular. Thus, the reader’s attention will be drawn to a new organization and to new structures in the theoretical background knowledge for comparative survey research which should exert a considerable influence in re-shaping survey methodologies and comparative survey research altogether. Adapting and accommodating to this new background knowledge, comparative survey research, in my assessment, should and will create new survey designs compatible with the emerging background knowledge and will leave behind its established core routines and its standard procedures as special cases, very much as Newtonian physics has become a special terrain within contemporary physics. In my judgement, the next years and decades will experience a fundamental change in the core approaches to identify attitudes,3 to determine subjective as well as objective living conditions or to explore individual life styles. The paths for comparative survey research of the 20th century and its trajectories in the 21st century will be situated in significantly different cognitive territories. 1 The term comparative survey research comprises three main components, namely data production through surveys, analysis of a single survey data set and comparative analysis of survey data. In the subsequent article, the term survey research is related to the production process of survey data, survey analysis is used for the statistical operations with single survey data and comparative survey analysis refers to the analytical operations with survey data for at least two different populations. 2 On comparative research in general, see, for example, Hantrais/Mangen, 1996, Harkness/ Vijver/Mohler, 2003, Hoffmeyer/Wolf, 2003 or Perry/Robertson, 2002. 3 On classical approaches to attitudes, see, e.g. , Ajzen, 1989, Ajzen/Fishbein, 1980, Eagly/ Chaiken, 1993 or Fishbein/Ajzen, 1975. 15.1 The Rise and Decline of Traditional Comparative Survey Research At the outset, a few general reflections will be undertaken on the ambiguous status of comparative survey research both in science and in society. On the one hand, survey based research has become the most frequently used publication mode across the social sciences. The following table, compiled by Willem E. Saris and Irmtraud N. Gallhofer4, shows a remarkable increase of survey-based research in a wide variety of the social sciences, including, surprisingly, social psychology and economics as well. TABle 15.01 The Rise of Survey Research economics Sociology Political Social Public Science Psychology Opinion (39.4%) (59.6%) (28.9%) (48.7%) (95.0%) 1949/50 5.7% 24,1% 2.6% 22.0% 43.0% 1964/65 32.9% 54.8% 19.4% 14.6% 55.7% 1979/80 28.7% 55.8% 35.4% 21.0% 90.6% 1994/95 42.3% 69.7% 41.9% 49.9% 90.3% On the other hand, comparative survey research has been questioned in recent decades on cognitive, epistemological and social grounds by at least three important components inherent in the processes and operations of comparative survey research itself, namely by social researchers, by the respondents of surveys and, at least in an indirect manner, by the underlying information and communication technologies. Currently, these three elements, in combination, gain in cognitive impact and could significantly reduce the capacities and the potentials for comparative survey research in its conventional form in the future. It might well be that comparative survey research has reached its maximum diffusion level and, despite its high level of achievements and sophistication, is already entering a phase of relative decline.5 4 See the summaries by Saris/Gallhofer, 2007:2p. 5 It belongs to the well-known results of innovation research that old technologies are usually replaced at the height of their efficiency and their relative strength. Thus, analog cameras have been substituted by digital ones at the height of their performance-levels. On this point, see especially Utterbeck 1989, 1996 or von Foerster 2003:284. Towards New Frontiers in Comparative Survey research 463 New Alternatives for Data Production and the Issue of Survey Measurements Within the field of empirical approaches towards the social worlds, comparative survey research has been pressured by two developments within the social sciences themselves, the first being the diffusion of new forms of so-called qualitative research and the second focusing on the still unresolved question of survey measurements. Turning to the first process, the last decades have witnessed a wave of new approaches for exploring the social worlds under the heading of qualitative social research6. Leaving aside the problem of the relative merits of quantitative or qualitative designs and methods, the new domains of qualitative social research offer also a new way of interacting with individuals or social groups. By and large, these new qualitative approaches for data generation and data analysis view their respondents in an entirely different way to the procedures of standardized surveys. Qualitative interactions and methods try to maintain an open and flexible access to respondents, stress their autonomy and want to bring to surface not only specific episodes and perspectives by individuals or social groups, but more general biographical structures or styles of narration as well. In a significant number of instances, qualitative and quantitative research produce incoherent or even contradictory results, which, in turn, call for more complex ways of combination and integration like the so-called method of triangulation. Aside from a new wave of qualitative research, the quantitative data themselves which are generated in traditional survey research have been a constant focus of discussion and interpretation. From the 1920s and 1930s onwards, attitudes and questionnaire-designs for identifying attitudes gradually occupied the center of the analytical stage, cul-minating in Gordon W. Allport’s “The concept of attitude is probably the most distinctive and indispensable concept in contemporary American social psychology” [Allport 1937:3] in 1937. Accompanying the successful diffusion of comparative survey research, one finds, however, an alternative perspective which sees itself diametrically opposed to the dominant survey designs and to the hegemony of standardized empirical research. In the formative years of survey and attitude research Richard T. La Piere has shown beyond reasonable doubt, but also with no lasting success, that attitudes and actions are separated by an unbridgeable gap. His assessment from 1934 seems as valid and controversial now as it was then. 6 On the widening and deepening field of qualitative research, see Flick/Kardorff/Steinke, 2005, Helferich, 2005 or Lamnek, 2005. 464 Karl H. Müller | Niko Toš The questionnaire is cheap, easy and mechanical. The study of human behavior is time-consuming, intellectually fatiguing and depends for its success on the ability of the investigator … Yet it would seem far more worthwhile to make a shrewd guess regarding that which is essential than to accurately measure that which is likely to prove quite irrelevant. [La Piere 1934:237] Following this counter-tradition to the dominant forms of comparative survey research Peter Converse’s article on the nature of belief systems in mass publics in 1964 marks another hallmark in approaching survey measurements and survey data in a radically different way. His article on non-attitudes can be seen as another important reference point for an alternative view of the survey measurement process. According to Converse, attitudes measured in normal survey research qualify as non-attitudes only and as such they are highly volatile and subject to frequent changes. Subsequently, the issue of non-attitudes provoked a new response on the part of conventional survey research which emphasized the “real” measurement of “real” attitudes, but allowed for varying degrees of measurement errors. Thus, non-attitudes quickly changed into true attitudes again, albeit in a slightly blurred and fuzzy version. Thus, it seems worthwhile to go deeper into the issue of measurement processes and of measurement types not only in survey research7, but across different scientific domains. Table 15.02 presents an elementary division of measurement types for measurements across the natural and the social worlds. Here, two measurement dimensions are used, the first one on the repeatability of measurements (exhaustive/repetitive) and the second one on the degree of observer dependency. – Exhaustive measurements lose, due to the measurement process, the possibility of a renewed measurement whereas repetitive measurements can be performed over and over again. – Similarly, strong observer dependency means that the measurement process itself produces or generates the quantity to be measured whereas in weak observer-dependent contexts the quantities to be measured could be measured, in principle, before or after the actual measurement process as well. Normally, the first measurement type in Table 15.02 is linked to the realms of quantum physics, where measurements are both exhaustive and strongly observer-dependent [ e.g. , Zeilinger, 2005], and the fourth type to the macro worlds across nature and society. Following the conventional wisdom 7 On measurements in survey Andrews, 1984, Blalock, 1968/1990, Edwards/Bagozzi, 2000, Esposito/Rothgeb, 1997, Hox, 1997, Krosnick/Abelson, 1991, Lass/Saris/Kaase, 1997, Lord/Novick, 1968, Miethe, 1985 or Sniderman/Therbiault, 2004. Towards New Frontiers in Comparative Survey research 465 [ e.g. , Hand, 2004 or Henshaw, 2006], measurements and observations in survey domains, if properly designed and conducted, can be treated like measurements of velocities, length, temperature or distances in the macro-natural arenas. In short, measurements in survey research, if adequately adjusted for measurement errors, correspond to type IV [Saris/Gallhofer, 2007].8 TABle 15.02 Types of Measurement Strong Observer-Dependency Weak Observer-Dependency exhaustive Measurement Type i Measurement Type ii repetitive Measurement Type iii Measurement Type iV In sharp contrast to an established consensus on survey measurements, it will be argued subsequently that measurements in survey contexts, as performed according to the rules and guidelines of empirical social research,9 fall under the first measurement type and not under the fourth type. With respect to the dimension exhaustive/repetitive, asking a survey question once destroys the possibility for asking it again immediately afterwards. Asking the same item two, three or more times in a row creates a new context for respondents. Likewise, asking the same survey questions in a repeated manner meets definite barriers and constraints on the side of respondents.10 Thus, along the first dimension measurements in survey research are in no way similar to consecutive and repeated measurements for physical macro-objects and their properties like velocity or temperature. For the dimension of weak/strong observer dependencies two broad alternatives are feasible in survey interactions. According to the conventional 8 Measurement type II is reserved for those cases where the measurement process destroys the conditions of the possibilities for renewed measurements. For example, measuring the breaking point for materials makes a renewed measurement impossible. Measurement type III applies whenever the measurement process creates a measurable quantity via the measurement process itself. Like in the case of a roulette, a croupier as a strong observer produces a sequence of numbers from 0 to 36 in a just in time-manner. 9 See, for example, the handbook of survey methodology by de Leeuw/Hox/Dillman, 2008. 10 Respondents could be asked, however, if they are willing to participate in a weekly or even in a daily survey. But such a demand must be stated clearly in advance and must be fully accepted by respondents. Again, in a daily questionnaire it will become exceedingly difficult to ask the same question twice. Additionally, daily surveys will be accepted by respondents only íf they deal with daily changing processes. This condition is usually fulfilled in the field of consumption, media utilization within the last 24 hours or in the area of social contacts. But it should become exceedingly difficult to ask respondents on their trust in institutions in daily intervals. 466 Karl H. Müller | Niko Toš wisdom in survey research, responses are based on stable assessments which are well embedded in the cognitive-behavioral repertoire and inside the neuro-cognitive organization of respondents. Thus, Jon A. Krosnick, Charles M. Judd and Bernd Wittenbrink adhere to the storehouse or file drawer image of attitudes and see a great theoretical and practical value … to hypothesize that a single attitude exists in a person’s mind: the net evaluation associated with the object. [Krosnick, Judd/Wittenbrink 2006:26] In this traditional view, survey measurements are founded on respondents’ introspective reports of their stable long-term attitudes and beliefs which are well-stored in the long-term memory of respondents. On this account, survey measurements can be subject to measurement errors which, however, can be corrected and adjusted. In the alternative perspective, survey responses are created just in time within the context of a survey itself, without prior fixed quantities or specific values in the cognitive repertoire and organization of respondents. Rather, due to the fixed menu of admissible survey responses, respondents can be assumed to match this unusual format with their ordinary language routines and, albeit in a spontaneous manner, with some of their past experiences. In this perspective, survey responses are creative reactions on unusual requests which in most instances are produced and delivered in a spontaneous manner. By necessity, the second alternative in conjunction with the exhaustive character of survey measurements, leads to the first measurement type in Table 15.02. It will become the main task in the next sections to build up additional support from the cognitive sciences and from elements in the background knowledge of comparative survey research to justify the assumption that survey responses are exhaustive in nature and are the results of strong observer dependencies. The Increasingly Strange Social Form of Survey Questioning Aside from the scientific domain itself, survey research has come under implicit attack by the respondents of surveys, too, on two major fronts. First, in the course of societal differentiations and of migration processes, it becomes more and more difficult to include special groups in a representative survey sample. For example, migrant persons with severe language problems in their new environments, younger and highly mobile persons, very old persons living in retirement homes, these are just three societal groups which are usually excluded from a survey data set. Due to the increasing mobility and due to individualization processes it can be expected that representative surveys become more and more difficult to implement. Towards New Frontiers in Comparative Survey research 467 Second, respondents turn out to be less and less willing to follow the ordinary procedures of a survey interview. Response rates for telephone or face to face surveys are declining continuously and respondents tend to defect especially in the case of telephone interviews.11 It seems worthwhile to investigate this issue of a growing survey apathy or, alternatively, of survey irrelevance in closer detail. The method for this in-depth analysis can be qualified as a variant of an ordinary language investigation and seeks to establish implicit or explicit rules, structures and constraints in the language game called survey questioning.12 In general, survey interactions, whether by telephone or face to face,13 belong to a large class of question and answer-interactions which have found their way into different societal domains or systems and play their indispensable and non-substitutable roles. A very broad division separates question and answer interactions into a continuum of symmetric and asymmetric instances. – Symmetric interactions come close to Jürgen Habermas’ ideal speech situation14 where each participant has an equal chance of asking and of answering and where role reversals are a necessary part of the symmetry arrangement. – In asymmetric interactions the roles of the questioning part is usually fixed and no role reversals are allowed. Additionally, one observes a clear power asymmetry between the actors responsible for the questioning from the persons undertaking the answering parts.15 Here, all important strategic moves are reserved for the questioning part and the responding part has to adapt to these requirements. Symmetric questioning and answer routines have been part of the conversations in civil society contexts and have become deeply embedded in virtual worlds where forums, chats and other forms of on-line discussions use this open and equally accessible format. Moreover, new forms of opinion expression on a wide variety of issues and topics like blogs, facebook groups, twitter and the like diffuse rapidly within the cyber-sphere. Here as well, 11 In terms of quality control, telephone interviews are considered to have the lowest values for reliability or validity. See, for example, Saris/Gallhofer, 2007. 12 On the process or on the language game of questioning, see, for example, Groenendijk/ Strokhof, 1997. 13 In the subsequent discussion, self-administered surveys are not being considered since they do not involve other actors outside the respondent her- or himself. 14 See, for example, Habermas, 1981 and 1984. 15 With respect to power, Max Weber’s definition of power can be used as a useful reference point. See Weber 1980. 468 Karl H. Müller | Niko Toš symmetric forms of expressing one’s opinion and reacting to it have become the dominant mode. Moreover, blogs can be linked to each other, leading to new virtual communities and the like. Typical contexts for asymmetric and power loaded question and answer-interactions can be found, inter alia, in national legal systems, in national systems of security, in national education systems or in the national systems of tertiary education and of professional learning.16 The recurrent use of the word national is not used accidentally because these question-answer interactions emerged in the course of the state building processes in the domain of an observing and of a caring state [Abraham de Swaan]. In Western societies these asymmetric interaction forms were built on earlier models of control in military or in religious contexts. Turning to these asymmetric question and answer interactions more closely, it becomes noteworthy that aside from their power asymmetry they usually operate under special constraints, especially under a constraint of subjective truth or truthfulness. With respect to the question and answer interactions in surveys it goes (almost) without saying that they belong to the asymmetric cluster. Survey questioning and answering allows for no role reversals and operates, as its differentia specifica, with a very restricted number of options for answering. Thus, survey interactions are characterized by a practical necessity to select a specific pre-defined answer from a menu of admissible answers.17 Moreover, the small number of admissible answers is rather uncommon to most respondents since they are required to use numbers and scales or different degrees of agreement or disagreement. Survey interactions operate on a weak constraint of truth in the sense of honesty, truthfulness or accuracy since these features simply are expected to be fulfilled within survey interactions. Surveys usually are free of legal or criminal sanctions except for exceptional instances of fraud on part of the interviewers.18 In recent years, the ways of expressing opinions has increased substantially in the Western world, especially following the revolution in information and communication technologies (ICT). Currently, the virtual web-communities 16 National systems of broad domains like education, security, innovation or health comprise, from a cross-national perspective, a heterogeneous set of ensembles with a multiplicity of different actors and with characteristic forms of bindings or bondings between actors. For a theoretical summary, see Hollingsworth/Müller/Hollingsworth, 2002. 17 In this respect, surveys are similar to multiple choice tests which operate on the basis of a pre-given set of opinions plus the necessity to choose one of them. 18 A national census can be an exception to this condition which, at least in some countries, have criminal sanctions both for interviewers and respondents. Towards New Frontiers in Comparative Survey research 469 with unrestricted forms of expressions and the potential for freely formed self-expressions abound. Thus, from the cognitive horizons of respondents, survey interactions become more and more outdated and, from the repertoire of everyday expressions, rather marginalized. Economic Pressures and Technological Advances A third element which currently weakens comparative survey research comes directly from the fields of the already mentioned ICT-technologies and their economic repercussions. In particular, the rapid diffusion of a new wave of information and communication technologies (ICT)19 brought about a dual effect. On the one hand, the new ICT-technologies lowered the entrance barriers for new firms in the field of market and opinion research. On the other hand, the new ICT-technologies do not only offer new modes for the production of survey data, they also provide solutions for alternative ways of organizing surveys. The first point is easy to describe. Through advances in ICT-technologies, the costs for setting up small and medium-sized telephone labs has decreased dramatically. More specifically, rapidly decreasing hardware costs for laptops or PCs, computer aided software for surveys plus statistical software packages and, finally, decreasing telephone costs or the emergence of call-centers have led to a situation where new firms with small amounts of funds can easily enter into the field of market and opinion research. Moreover, due to the increased competition, the prices for conducting surveys are declining rapidly as well. This, in turn, led to a considerable widening of surveys outside the classical fields of market research or the social sciences. Currently, both print and audiovisual media use surveys for elections and for opinions on widely debated political issues extensively. The characteristic constraints and consequences of these technological advances and economic developments are not widely discussed. Usually, survey research operates under increasingly strong economic pressures and limitations. Within these new configurations, the critical element in surveys becomes time. Thus, surveys normally are generated under critical time constraints since interviewers are not full-time employed and their income usually depends on the number of successful interviews. Thus, the survey interactions become more and more reduced and individuals who cannot be contacted immediately are replaced by the next available person in the sample. In this way, the quality of survey data decreases substantially and it comes as 19 More systematically, see Müller 2008a. 470 Karl H. Müller | Niko Toš no surprise that the data quality of telephone interviews is considered as the lowest among all possible modes.20 Turning to the second point, these new information and communication technologies enable and support a new mode for survey research, namely online or e-surveys.21 Here, the degree of autonomy of respondents is increased substantially because there are usually no time-constraints and, dependent on the technical implementation, respondents can go back to review their previous replies. These new technological advances allow, moreover, better and better simulations of face to face questionnaires, with voices talking to respondents or with virtual cards which mimic more and more components of face to face interactions. This new technological potential increases the development of new types of online-surveys which, for example, follow the path of deliberative surveys.22 Here, respondents receive to each question a substantial amount of background information which can be used prior to answering a single survey item. To sum up, the interaction processes in surveys are undertaken currently under heavy economic pressures and that means under strong time constraints. Additionally, the measurement processes through telephone interviews are organized in a way that leaves less and less room for reflective answering and for more symmetric forms of interactions, reducing, thus, the data qualities of surveys considerably. Additionally, new modes23 of online-surveys widen the possibilities of survey research well beyond the conventional designs and open up a new and rapidly expanding trajectory of non-standard survey research. 15.2 A Phase Transition from Science I to Science II The most dramatic change for comparative survey research in the years ahead, however, is cognitive in nature. Within the last decades, the science system as a whole has entered a phase of radical transformation from an old regime, called Science I, to a new regime under the name of Science 20 For a summary on the issue of modes and data quality, see Saris/Gallhofer, 2007. 21 For a summary, see, e,g., Couper, 2000 or Dillman, 2000. 22 On deliberative surveys, see, e.g. , Luskin, R.C., J.S. Fishkin, R. Jowell, 2002 or, for a more general background, Carson/Hartz-Karp, 2005 or Booth, 2006. 23 On the issue of survey modes in general, see e.g. , Pruchno/Hayden, 2000. Towards New Frontiers in Comparative Survey research 471 II.24 Table 15.03 captures some of the important shifts from the traditional cognitive organizations of science to its new theory structures. TABle 15.03 Main Distinctions between Science I and Science II Dimensions Science I Science II evolutionary Biology and leading Fields of Science Classical Physics the Sciences of Complexity Pattern Formation and Theoretical goals General, Universal Laws Pattern recognition Nested and embedded Theoretical Perspectives Axiomatic, Reductives Processes leading Metaphors Clocks Clouds ludwig Wittgenstein Core Philosophers rene Descartes (Cogito) (Cogitamus) epistemology Observer excluded Observer included Dualism (res cogitans/res Monism, Self-Organization Ontology extensa) Capacities generative Mechanisms Trivial Mechanisms Non-Trivial Mechanisms Forecasting Capabilities High low Complexity low High Non-Linear, far from Perspectives on Change Linear, Equilibrium equilibrium emphasis on Mild emphasis on Wild Distributions Distributions Distributions Potential for inter-Disciplinary low High research Cognitive Distances be-tween the Social Sciences and the High Medium/Small leading Field of Science Before presenting some of the relevant changes in the theoretical background knowledge in greater detail, a short parable will be offered which captures these current changes in a metaphorical way. The subsequent parable takes place in the context of airplanes and airplane cockpits. According to this parable, each scientific discipline operates in a plane of its own and, as a 24 Science I was the dominant form of science from the beginning of modern science I the 16th century up to 1900/1950. Science II, consequently, emerged over the last decades and will turn out to be the new hegemonial regime, although Science II will not replace Science I completely. In a variety of domains and applications Science I-models and methods will be still being used. On this distinction between Science I and Science II, see especially Hollingsworth/Müller, 2008 and on a wider discussion of this separation see Boyer, 2008, Mayntz, 2008, Nowotny, 2008 or Sornette, 2008. 472 Karl H. Müller | Niko Toš consequence, one finds two big operating fleets, one for the natural sciences and one for the social sciences.25 In the traditional airplane cockpits, a social or a natural scientist as data pilot or data navigator had a large number of instruments at her or his disposal which, if constructed properly, provided accurate measurements of the states of their planes. The instruments provided a direct access to either natural or societal processes and the measurements produced reliable results. The main task of the data pilots or data navigators was to record and analyze their data and to send their findings to a control center which was assumed to be in control of the steering processes both for nature and for society. Best practices in social research, for example, consisted, then, in an intelligent reading and in a fruitful synthesis of these various instruments which produced, in the end, a reliable and a stable picture of the overall internal flight dynamics. The cockpit designs and the rules to operate with them worked exceedingly well for several hundred years. For a short period it seemed that especially the airplanes under the flag of nature could be navigated and controlled nearly perfectly. The larger and smaller planes from the social sciences operated, though less successfully, with the same standards and procedures as the airplanes under the flag of nature. Nevertheless, several attempts were undertaken to develop alternative instruments and rules particularly suited to the needs of the societal fleet. Quite suddenly and unexpectedly however, the core parts of the cockpit designs and instruments for the large fleet under the flag of nature were overthrown nearly completely and substituted with new designs, new instruments and new operation techniques. These new environments, although only partially recognized and acknowledged at that time, differed radically from the old settings. The new instruments turned out to be ambiguous to read and highly difficult to interpret. Moreover, the old rules, linking input data to output data no longer worked sufficiently, reducing, thus, the forecasting capacities significantly. Additionally, new types of data on the internal states of airplane components or of environmental elements became more and more important. As a consequence, the data pilots began to develop more complex models for data integration. 25 It should be added that a large number of airplanes were available at that time, both for the natural and for the social sciences. It can be assumed that each scientific discipline, even the smallest ones, had its own carrier. Thus, the airplanes of the social sciences or of the natural sciences consisted of a very large fleet of airplanes, respectively. Towards New Frontiers in Comparative Survey research 473 Rather rapidly, these new instruments and designs found their ways in more and more airplanes. But the replacement process was undertaken in a slightly confusing manner because the traditional cockpits and instruments were kept in place and the new instruments and devices which replaced the old instruments and operation techniques in most of their important aspects looked strikingly similar to the older versions. In fact, these new settings created enormous challenges for the data pilots because they were required to interact with their flight objects and their environments, be it passengers, animals, plants or clouds in order to be able to continue with their operations. Thus, in the new types of planes data navigators were confronted with a collection of complex instruments which appeared to operate at least partially at random. Moreover, the airplanes turned out to be self-organizing, too, and exhibited all characteristics of highly complex self-organizing ensembles which led to a complete re-thinking of control and steering issues under the heading of “illusion of control”.26 For a long time the data navigators, being trained in traditional cockpit designs and cockpit techniques, were quite unaware that they became embedded more and more in complex settings. Especially the airplanes which constituted the fleet called society experienced a period of severe crises. Using their traditional techniques and restricting data navigators to their traditional cockpits and instruments led to massive failures and highly misleading estimations. Most notably, even dramatic changes in the direction of the planes escaped the attention of societal data navigators who stuck to their old ways of analyses and data-handling. Only very slowly and in small jumps the societal data navigators shifted to the new modes of operation and to the new forms of interactions with their passengers or their non-trivial instruments. And only step by step they were able to adapt to the modes of operation of a self-organizing plane. Especially for societal data navigators, it turned out to be a long and an extremely challenging process to learn the characteristic modes of operation in and with these new complex environments. This short parable makes one aware that the current foundations of empirical social research have begun a transformation process which also affects the theoretical and epistemological background knowledge for the social sciences and, in particular, for comparative social research.27 26 On the topic of “illusion of control”, see Piatelli-Palmarini, 1994 or Pohl, 2004. 27 On the background knowledge relevant for Science I see Hollingsworth/Müller, 2008 or Hollingsworth/Müller/Hollingsworth/Gear, 2008. 15.3 Significant Changes in the Background Knowledge for Comparative Survey Research in the Age of Science II Before entering into content issues, the concept of background knowledge (BK) for a scientific discipline Di must be introduced in a clear and accessible manner. In a general way, the theoretical, epistemological and methodological background knowledge for a scientific discipline Di is directly related to the leading field or to the core scientific discipline (LD ) at a given time t, to the t theory structures of LD , to its dominant epistemologies and to its general t methodology. In the days of Science I with theoretical physics as leading scientific field for example, the search for universal laws or a reductionist view of theory structures with physics as its basis were typical elements of the theoretical background knowledge for the social sciences in general or for comparative survey research in particular. Likewise, epistemological or methodological rules, associated with theoretical physics like objectivity, scientific realism or the covering law model of explanations became characteristic building blocks for the social sciences, too.28 Currently, we live amidst the diffusion of Science II, with the life sciences as leading field and with its corresponding epistemologies and ontologies. In the context of Science II the theoretical, epistemological and methodological background knowledge for scientific disciplines outside the leading field of the life sciences undergoes significant changes. In a non-metaphorical way, at least seven building blocks can be identified which are characteristic for a science of living systems by living systems and which, therefore, become characteristic for the emerging theoretical, epistemological and methodological background knowledge in Science II. Clearly, these new building blocks will exert a considerable cognitive pressure on the theory and research organization from the era of Science I and should lead to new theory structures and research designs for comparative survey research as well. Table 15.04 as well as Figure 15.01 summarize those changes in background knowledge that will be of particular relevance for comparative survey research. As can be seen both from Figure 15.01 and from Table 15.04, the main 28 It should be added that throughout the 19th and the 20th century one finds a very broad literature and a constant discussion that the standards and theory structures of theoretical physics cannot and should not be used within the social sciences. See, above all, the separations and distinctions between so-called nomothetic and idiographic sciences [Acham, 1974] or, as a particularly relevant point of debate, the so-called ‘Positivismusstreit’ [Adorno et al. 1978] Towards New Frontiers in Comparative Survey research 475 differences between the old and the new background knowledge cover the entire domain of analyses, namely the subjects of investigation, the objects of analysis and, finally, the interaction modes between subject and object of analysis. All three domains differ strongly between Science I and Science II. In short, Science II has become a science of living systems for living systems in which the subjects of analysis, being living systems themselves, are an indispensable and inclusive part of an investigation. The objects of analysis, namely living systems, turn out to be far more complex than the physical objects within Science I. Finally, the interactions between subjects and objects are organized, as will be shown later, in a closed triadic as well as recursive manner. TABle 15.04 Changes in the Theoretical, Epistemological and Methodological Background Knowledge of Science I and Science II Domains of Backgroung Science I [Theoretical Science II [Life Sciences Knowledge Physics als Leading as LD] Discipline (LD)] Objects of investigation Trivial Actors Non Trivial Actors Trivial Action embedded Cognition Schemes Cognitive isolationism Cognitive Holism Single Account requisite Variety Sufficient Necessary Subjects of investigation Observer exclusion Observer-inclusion interactions (between Sequential, Linear Recursive, Non-Linear Subjects and Objects) equilibrium eigenforms Dyadic, Asymmetric Forms Triadic, Symmetric Configurations The methodological and theoretical elements of the new background knowledge emerge from the leading field of Science II, namely form the cognitive life sciences, broadly conceived whereas the new epistemological components come from a diverse group of frameworks which are particularly focused on the specificities of living systems like the approaches by Robert Rosen [2005] and Walter M. Elsasser [1998], radical constructivism or, as specially 476 Karl H. Müller | Niko Toš relevant subsets of radical constructivism,29 second-order cybernetics30 and the autopoietic approach.31 These and similar perspectives are especially relevant for shaping the core epistemologies of Science II-research.32 From both sides, the theoretical-methodological and the epistemological one, the conventional wisdom of comparative social research is not only questioned in its core aspects and in its central designs,33 but comparative social research is also very much encouraged to change its traditional perspectives in order to become compatible with the new Science II landscapes.34 Following Figure 15.01 and Table 15.04, the first distinctive feature between the background knowledge of Science I and Science II lies in the units of analysis and has been captured by the dichotomy of objects (Science I) and living systems (Science II). As it turns out, living systems as the main actors on the stage of Science II are structured and organized in a significantly different way than the physical objects in the phase of Science I. Living systems can be characterized by attributes like autonomy, internal state-determination, multilevel organization, learning and the like. The most important differences to the objects under Science I lie, however, in the relations between a researcher and her or his domain of investigation. Under Science I, objects and researchers were situated in different ontological domains whereas under Science II the researcher her- or himself is a living system, too, and can be described with categories like autonomy, internal state-determination, multi-level organization, learning and the like. This point has far-reaching implications, some of which will be dealt with under the notion of observer inclusion (See, once again, Figure 15.01). Aside from an analytical distinction in terms of living actors, a second new element in the theoretical background knowledge is related to the description of the environments of living systems which should be conceptualized in a significantly different way, too. In Science I, physical objects could be studied as embedded in a physical environment and with directly observable 29 On radical constructivism in general, see, as summaries Watzlawick, 1981, Watzlawick/ Krieg, 1991, Schmidt, 1987 or Glasersfeld, 1997. 30 For second-order cybernetics, see especially von Foerster, 2003. 31 On the autopoietic approach, see, for example, Maturana(1985) or Maturana/Varela, 1987. 32 Second-order cybernetics has been developed explicitly by Heinz von Foerster as a science of living systems for living systems. On Heinz von Foerster and his work at the Biological Computer Laboratory, see especially Foerster, 2003, Müller/Müller, 2007 and Müller, 2008. 33 For interesting overviews and approaches, see Palombo, 1999, Ryckman, 2000. 34 The morphological approach in the previous chapter can be seen as a typical example for a new design which corresponds rather well with the new theoretical background knowledge of Science II. Towards New Frontiers in Comparative Survey research 477 interactions between objects and their surroundings. Under Science II, cognitive models and methodologies will follow more and more along the pathways of embedded or situated cognition.35 This road towards embedded or situated cognition is so important because it puts special emphasis on the distributed nature of cognition, on thinking environments and why, following a well-known reversion by Humberto R. Maturana, the mind is not the head [Maturana, 1985]. Figure 15.01 The Emerging Epistemological, Methodological and Theoretical Science II-Background for Comparative Social Research   In situated cognition, a brain needs not only a senso-motoric active organism, but also a responsive environment for its own internal cognitive operations. In this view, the environment becomes the necessary co-evolving counterpart for individual actions and interactions. Table 15.05 exhibits some of the characteristic differences between the two forms of environments in Science I and in Science II. 35 On embedded cognition see especially Adams/Aizawa, 2008, Bodenhausen/Lambert, 2003, Gibbs, 2005, Noe, 2009 or Robbins, 2008. 478 Karl H. Müller | Niko Toš Following the distinctions between objects and living systems and the notion of embedded cognition, the third point of separation between the old and the new background knowledge is focused on the complex nature of cognitive processes in living systems. The study of cognition under Science II departs more and more from the pathways of conventional folk-psychology or from typical Science-I traditions like behaviorism. One of the challenging general heuristic rules for the study of cognitive processes is called cognitive holism and has been propagated strongly already in 1969 by Heinz von Foerster [Foerster, 2003]. TABle 15.05 The Changing Roles of Environments between Science I and Science II Environments in Science I Environments in Science II Weak Boundaries Strong Boundaries Direct Crossings indirect Crossings Cause – effect Triadic relations Direct environmental effects Boundary Transformations, Indirect Effects external Dynamics internal Dynamics Adaptation internal Complexity Drifts According to this rule, it is possible, in principle, to isolate conceptually certain components in the stream of cognitive processes, for instance the faculty to perceive, the faculty to remember,36 the faculty to infer or many other senso-cognitive as well as cognitive-motoric faculties.37 However, one should not expect that these conceptual separations between perception, memory or inference finds a 1:1 correspondence in the neural organization of actors. In its generalized form this principle can be formulated in the following way: 36 In principle, it would be possible to differentiate between a large number of these faculties like the faculty to infer, the faculty to learn, the faculty to evaluate, the faculty to communicate or the faculty to move, to name just a few more additional faculties. 37 On current summaries of the neuro-cognitive architectures of these different faculties, see Gazzaniga/Bizzi/Black, 2004 or Calvert/Spence/Stein, 2004. Within the cognitive neuro-science arena, one finds meanwhile numerous sub-fields and disciplinary niches specializing on a particular senso-motoric, emotional or cognitive faculty. For a diverse set of literature, see Calvin, 1996, Calvin/Bickerton, 2000, Campbell, 1984, Damasio, 1994/2003, Deacon, 1997, Edelman, 1987/1990/1992/2007, Hofstadter, 1982, Hofstadter/Dennett, 1982, Hofstadter 1985/1995/1997, Holland, 1995, Lakoff/Nunez, 2000, Minsky, 1990, Norretanders, 1997, Pinker, 1997, Plotkin, 1997, Pollock, 1989, Ratey, 2001, Roth, 1999 or Sternberg/Wagner, 1994. Towards New Frontiers in Comparative Survey research 479 If one wishes to isolate these faculties functionally or locally, one is doomed to fail. Consequently, if the mechanisms that are responsible for any of these faculties are to be discovered, then the totality of cognitive processes must be considered … [von Foerster/Müller 2003:29f.]38 The inseparability of these faculties can be shown by a reduction ad absurdum. It can be demonstrated that the assumption of independent faculties in isolation leads to absurd consequences. In particular, it can be demonstrated that, if one of these seven faculties mentioned above is omitted, the entire system is devoid of cognition. In greater detail this proof runs as follows: (i) Omit perception: the system is incapable of representing internally environmental regularities. (ii) Omit memory: the system has only throughput. (iii) Omit prediction, i.e. the faculty of drawing inferences: perception generates to sensation or memory to recording. (iv) Omit learning: the system is bound to a fixed modus operandi. (v) Omit evaluation: the system is incapable of selecting. (vi) Omit communication: ……………………….. (vii) Omit movement: the system is incapable of acting. The fourth point of departure between the old and the new background knowledge moves away from the descriptive requirements for living systems and their environments or from the theoretical requirement of a holistic organization of cognitive theories. Instead, the fourth point emphasizes the importance of a requisite descriptive and data variety. Under Science II, the focus shifts away from descriptions and measurements at a single level to more general configurations and, above all, to different levels of descriptive and data variety. The study of living systems requires a multiplicity of data sources which cover the entire range of measurements, ranging from the cellular and neural level, to high-level brain measurements, up to the levels of internal or external verbal accounts or the observations of acting and interacting persons in the case of human societies. All these different levels, measurement types and data formats are needed in order to reach a fuller understanding of living systems in their contexts or environments. 38 Aside from the functional and the local theses, one could put forward two additional theses, one on genetic holism and one on epistemological holism: Genetic Thesis: If one wishes to isolate these faculties genetically, one is doomed to fail. Consequently, if the genetic mechanisms that are responsible for any of these faculties are to be discovered, then the totality of genetic processes must be considered … Epistemological Thesis: If one wishes to describe these faculties in an external, observer-independent stance, one is doomed to fail. Consequently, if the mechanisms that are responsible for any of these faculties are to be discovered, the mode of description has to change into an internal, observer-including stance. 480 Karl H. Müller | Niko Toš The most important challenge with respect to the requisite variety of measurements and observations will be to bridge the currently deep gap between behavioral observations and the level of the brain scans and neural measurements. But Science II, in contrast to Science I, will generate a rich flow of neural and brain data not only on different types of thought processes, but on daily routines and practices as well. For comparative survey research, this new stage of a requisite descriptive and data variety will bring about a dense stream of neural patterns and data for the different stages in survey interactions which should allow to tackle classical issues and controversies like the one on non-attitudes in a fresh manner. Fifth, aside from the requisite descriptive variety, another epistemological point of considerable relevance for Science II-research in general lies in the inclusion of observers or, alternatively, of researchers in their research. Focused on social research in particular, research designs will change more and more from an exclusive into an inclusive mode, with social researchers as an indispensable element in it. In the world of Science I, mass, space and energy were the fundamental building blocks for a science of objects. Warren McCulloch was probably the first to note the peculiarity that breakthroughs in physics require the invention of surprising regularities or theorems of great abstraction which, however, are not included in the conceptual machinery for physical objects. Let us now compel our physicist to account for himself as a part of the physical world. In all fairness, he must stick to his own rules and show in terms of mass, energy, space and time how it comes about that he creates theoretical physics. [McCulloch 1988:73] Thus, in Science II one is, by necessity, confronted with a more inclusive task. Following, once again, McCulloch, a physicist must then become a neurophysiologist …, but in so doing he will be compelled to answer whether theoretical physics is something which he can discuss in terms of neurophysiology ... To answer ‘no’ is to remain a physicist undefiled. To answer ‘yes’ is to become a metaphysician. [ Ibid. ] In other words, Science II is the era of living systems, being researched by living systems. This new configuration brings the observer or the researcher as a necessary component back into her or his research domain. Thus, Science II becomes to a remarkable extent a self-referential and self-inclusive form of science. Sixth, another element of the new background knowledge for comparative survey research lies in the closed organization of relations into which living systems, the observing scientist included, can and should enter. In Science II, the new minimal configuration for the study of living or learning systems by living systems is not a dyadic relation between subject and object which Towards New Frontiers in Comparative Survey research 481 has been characteristic of the conventional scientific method, but is built in a triadic fashion, with the observing scientific researcher as one node, the domains under observations as another node and with a final node linking and closing these two nodes to a triadic ensemble. In short, Science II-research designs are to be built in their minimal form not with one, not with two but with three components.39 Science II, in contrast to Science I, will be characterized more and more by designs and by operations in configurations of a closed triadic nature which includes the observing researcher R, the domain under observation which usually is composed of observing living systems or, alternatively, of participant observers PO. Additionally, R and PO are closed by an inter-mediate element like rule systems or theories about the neural organization of PO and R which acts as a generative mechanism GM between R and PO. Characteristically, the types of relations in triadic configurations change from causal relations to generative relations. Figure 15.02 as well as Table 15.06 highlight the significant differences between the causal forms of Science I and the generative configurations of Science II. Figure 15.02 Triadic and Generatively Closed Research Designs Seventh, Table 15.06 included already the notion of necessary Eigenforms, inherent in triadic and generative configurations. Thus, the seventh ingredient of the new background knowledge emphasizes the production of Eigenforms. This implies that the recursive interactions between observers qua researchers and their living fields of investigation, if properly organized in a triadic 39 See also the paper from 1976 by Francisco J. Varela where he starts from the usual dualistic suspects like observer/observed, subject/object, describer/described, operator/operand and the like and continues: It is very obvious, however, that these poles are not effectively opposed but rather moments of a larger whole which sits in a meta-level with respect to both terms. [Varela 1976:65] 482 Karl H. Müller | Niko Toš fashion, lead to new stabilities or, to use an expression by Heinz von Foerster, to Eigenforms. Eigenforms become the central goals of research processes within Science II. TABle 15.06 Causal and Generative Relations Causal (A → B) Generative [PO(A, B), R, GM]* Asymmetrical in time Symmetrical in time Separation into cause and effect No causes and effects Cause is necessary, sufficient or both Mutual dependence Observers excluded Observers included Non-recursive recursive Openness Closure generalizations Necessary eigenforms * A, B: Events, O: Observer, GM: Generative Mechanism, PO: Living Systems as Field of Research More specifically, recursively closed interactions between researchers as living systems and living systems as the research domain, produce, by necessity, Eigenforms. This result, that there emerge Eigenforms, rests upon a theorem which runs under the name of Closure Theorem.40 Simply put, in every operationally closed system there arise Eigenforms (Eigenvalues, Eigen-behaviors, etc.).41 Recursively closed interactions and the emergence of Eigenforms exhibit a series of general and highly interesting properties which can be listed in a short summary. Eigenforms in numerical domains are discrete even if the domain of the primary argument is continuous. Moreover, Eigenforms represent equilibria, and, depending upon the chosen domain of the primary argument, these equilibria may be fixed points, functional equilibria, operational equilibria, behavioural equilibria, structural equilibria, etc. Furthermore, Eigenforms, emerge in a generative relationship, the one implying the other, and vice versa. These Eigenforms, because of their self-generating nature, imply topological closure or circularity and topological closure produces Eigenforms. Additionally, in an atomical social context each observer’s experience of her or his own senso-motoric coordination can be referred to by a symbol of this experience like “object” or “thing” which, at the same time, “may be taken 40 Among the many variants and paraphrases of this astonishing theorem, see, for example, Francisco Varela and Joseph Goguen’s version 1979. 41 On this point, see also von Foerster 2003:316. Towards New Frontiers in Comparative Survey research 483 as a token for the externality of communal space”. [Foerster 2003:267] The recursive language of Eigenforms can be extended easily to a molecular social context and to a configuration with two observing systems. The topology of closure can be applied to the entire spectrum of observable operations, from an atomic context with a single actor, from a two-actor configuration to many actor-ensembles as well as to self-referential configurations. Finally, Eigenforms are not dependent on infinite recursions, as has been suggested by Heinz von Foerster. The construction of Eigenforms ... can be done without the idealized excursion to infinity … via a device invented by Church and Curry in the 1930s that is commonly called the ‘lambda calculus’. [Kauffman 2005:131] Turning to the social sciences specifically, research designs will experience a long-term drift toward recursive designs. Table 15.07 presents a short list of current designs which already exhibit this new feature of operational closure and Eigenforms. TABle 15.07 Recursive Designs in Social Research Methods Applications Similarity Relations recursive Operations eigen Values Circular Social or Cognitive Recursive, towards Stable Social or Questioning42 Perspectives a Homogenization of Cognitive group Perspectives View* Delphi-Methods43 Scenarios, Cognitive Recursive, towards group-Consensus* Assessments Consensus Formation generative Social rule-based Dynamics Recursive, towards Equilibrium, Limit Sciences44 Stable Configurations Cycles, Strange Attractors, etc. Meta-Analysis45 results of empirical Recursive, towards robust results of research robust Knowledge empirical Tests* Triangulation46 utilization of Different Recursive, towards Stable results* research Methods robust Knowledge * eigenform for a Particular Design 42 On circular questioning see, for example, Pfeffer, 2001. 43 Delphi-Methods have been introduced already in the 1950s. See, for example, Rescher, 1998. 44 Generative social science has become a generic term for rule and actor based designs. For a summary, see Epstein, 2006. 45 Meta-Analysis has become a common procedure in the eighties and nineties of the 20th century in areas like clinical research and psychology. For a summary, see, for example, Hunter and Schmidt, 2004. 46 Triangulation has become popular quite recently as a design to integrate a heterogeneous set of research methods across the quantitative and the qualitative spectrum. See, for example, Punch, 1998:242–246. 15.4 The Fourfold Roots of a Fundamental Incompleteness of Survey Research in the Age of Science II In sum, these new elements in the theoretical background knowledge for comparative survey research should lead, by and by, to the recognition of a series of four significant deficiencies which, in combination, lead to the verdict of a fundamental incompleteness of traditional survey designs. The first incompleteness stems from the restricted code in which survey items are to be presented. Currently a single survey item is to be composed of an introduction, a motivation part, information regarding the content, instruction of the respondent, interviewer instruction, requests for answers and of answers with categories or response scales [Saris/Gallhofer 2007:121]. As has been shown in the first section of this article, the mode of fixed responses as well as asymmetric question and answer interactions have become more and more marginalized in contemporary life worlds. From this perspective, fixed categories and response scales should be at least accompanied by additional responses which are open and narrative in nature. These narratives, in turn, can be coded and produce, thus, an important supplementary body of data from a single respondent. In the final section another way towards supplementary data is shown with the help of visual productions which are generated by the respondents themselves. Again, these visual forms exhibit a large number of implicit measures that can be brought to surface. The second incompleteness is fundamental in nature and comes from the reliance of internal assessments and internal descriptions only. Here, the terms external description on the one hand and internal description on the other hand are to be used in the following way. External descriptions comprise any description by a competent observer of the overt manifestations and of the results of cognitive operations by an observable actor. In contrast, internal descriptions are tied to the self-description and self-evaluation of a competent actor alone. In a more detailed manner these two modes can be specified in the following way. The internal behavior description IN-BD of an actor A is based on a complex set of relations Ω between the internal states S , the inputs I , the overall A A context C and the previous history H : A A IN-BDA : Ω [S , I , C , H ] A A A A A The external behavior description EX-BD for an actor A by an observer O turns out to be far more complicated because it is based on a set of complex relations Ω between the internal states S of the observer, the inputs I O A,O Towards New Frontiers in Comparative Survey research 485 of the actor (as perceived by the observer) and of the observer, the overall context C for the actor (as perceived by the observer) and of the observer A,O and, finally, the previous history H both of the actor (as perceived by the A,O observer) and of the observer: EX-BDA,O : Ω [S , I , C , H ] A,O O A,O A,O A,O Turning more specifically to the context of survey research, an internal description comprises a set of responses by an actor A to the questions and items of given survey S . i IN-S DA : Ω [S , I , C , H ] i A A A A A where S , I , C , H constitute the relevant components which enter into the A A A A overall evaluation. Likewise, an external survey-based description S DA,O comes about through a i more complex configuration, namely via EX-S DA,O : Ω [S , I , C , H ] i A,O O A,O A,O A,O In comparative survey research under the flag of Science I it was necessary and sufficient to have a single internal descriptive account IN-BD of a respondent as the basis for subsequent analyses. This single account was considered as necessary and sufficient for two different domains, namely for the internal preferences, goals, attitudes, evaluations, etc. of an actor as reported by an actor and for the actions and interactions of an actor as a manifestation of these underlying preferences, goals, attitudes, evaluations, etc. This focus on individual respondents was supported by the view of individuals as carriers of stable internal preferences, goals, attitudes, evaluations which, due to their inscriptions in long-term memory, can be measured directly, albeit with a certain amount of measurement errors. In conventional survey research with the background knowledge of Science I one simply is justified to assume that, for example, the significant differences in working conditions between Scandinavia and Eastern Europe which have been shown in Part II of this book reflect different work settings which are viewed rather favorably in Scandinavia and quite unpleasantly in Eastern Europe. External observations can be supplemented in a wide variety of ways, for example through the inclusion of relevant macro-data or event data. A particularly interesting way of adding external data lies in social experiments and in supplementing survey questions with external behavioral data.47 The necessity for external observations can be strengthened by the fact that even more complex survey inputs like vignettes do not help to bridge the gap 47 For interesting examples in the field of environmental consciousness, see, e.g., Diekmann/ Preisendörfer, 2001 or Diekmann, 2007. 486 Karl H. Müller | Niko Toš between the traditional survey contexts and actual behaviors.48 Another way towards a requisite data variety comes through the inclusion of medical tests in the context of health surveys. A different path which has not been explored yet lies in the inclusion of another respondent R who, in a mode of circular 2 questioning, could be asked about the most likely answers of R .1 The third fundamental incompleteness is linked with the absence of survey researchers from survey research. This does by no means mean that survey researchers should report on their profile of responses in the survey they have constructed. Rather, the emphasis lies on a detailed documentation of a survey in terms of its targets as seen by a single researcher or a research group, its actual composition, its selection procedures that led to the final version, its relations, similarities and dissimilarities to existing surveys and, above all, the intended novelty and the “cash value” [Wilfried Sellars] of producing a new survey data set. One could easily think of a special survey for surveys which contains both open and closed questions and which gives an in-depth view of all the domains mentioned above.49 Moreover, such a survey of surveys could produce a whole new group of relevant meta-data which are currently missing.50 The important point here is that the inclusion of survey researchers, their expectations and their intentions, can be used for an empirically guided self-reflexive research on survey research. The fourth and final incompleteness leads to the interactions between respondents and researchers which happen currently in a highly restricted and almost random manner. In terms of triadic configurations, no generative mechanism links currently the side of survey researchers with the respondents, no recursive interactions occur and, thus, no drift towards Eigen-values can emerge. Figure 15.03 summarizes, once again, the emerging building blocks of a new background knowledge in its impact on comparative survey research in its traditional designs. 48 It should be added that vignette surveys try to provide a more complex input pattern. Nevertheless, also vignette surveys exhibit striking discrepancies to observations of actual behaviors. See, for example, Groß/Börensen, 2009. 49 In fact, such a survey of surveys could be organized as an online-survey with a large amount of additional information and links. 50 For this form of documentation, one can refer to the European Social Survey which for the time being qualifies as a best practice example and which has produced a large amount of materials on survey construction, on methodology or on general specifications. However, there is still a long way from the current list of documents to the final step of an in-depth meta-data documentation on the side of survey researchers. Towards New Frontiers in Comparative Survey research 487 Figure 15.03 The New Background Knowledge and the Fundamental Incompleteness of Comparative Survey Research  15.5 Towards New Designs for Comparative Survey Research in the Age of Science II Given the epistemological, methodological and theoretical sketch on significant components of Science II background knowledge for social comparative research in the previous chapters, it can be expected that research designs for comparative research will undergo significant changes as well in order to correspond to these new background components. The most important shift with respect to research designs can be captured with a well-known distinction within radical constructivism, namely the separation between trivial and non trivial configurations, be they machines, mechanisms, actors or research designs. The distinction between trivial research designs (Science I) and non-trivial designs (Science II) affects a large number of research dimensions 488 Karl H. Müller | Niko Toš simultaneously.51 Non-trivial designs and their main distinctions to trivial machine or design configurations can be summarized via Table 15.08. TABle 15.08 Trivial and Non-Trivial Research Designs52 Trivial Research-Designs Non-Trivial Research-Designs („Hard Sciences“) („Soft Sciences”)53 Models input/Function/Output Functors (Operators) independent Variable/Function Operating on Functions Dependent Variable State-Determined Systems Cause/law/effect with Non-linear Dynamics Stimulus/Nervous System/response (Discontinuous, Qualitative goal/System/Action Changes, Chaotic environment/Organism/Behavior Behavior, etc.) Motivation/Character/Actions, etc. Model Characteristics Predictable unpredictable independent of Pre-History History-Dependent Synthetically determined Synthetically un-determined (Functions Identifiable) (Functions not identifiable) Analytically computable Analytically not computable (Value of functions effectively computable) (Value of functions not effectively computable) reductionist Relational, systemic Methodology First-order Designs Second-order Designs Observer-independent Observer-dependent Hetero-referential Self-referential Organization Single researchers Teams, Groups it-Based Bit-Based 51 It should be added that the trivial and the non-trivial mode of description are universal modes respectively which can be used for living systems and objects alike. Thus, one is confronted with two types of errors, namely using a trivial description mode for non-trivial ensembles (trivialization error) or a non-trivial mode for trivial configurations (spiritualization error). 52 The subsequent differentiations in Table 15.08 have been put forward essentially by Heinz von Foerster, like in Foerster 1984:8ff. 53 Hard sciences and soft sciences can be differentiated in the following way. Hard sciences deal with relatively easy or soft problems, whereas soft sciences are concerned with hard problems. Towards New Frontiers in Comparative Survey research 489 In the old days of Science I, the natural or the social sciences operated in a cognitive setting which could be characterized as a trivial research environment, relying on trivial models, mechanisms and methods and on trivial research designs.54 Figure 15.04 The Opening Cognitive Landscapes for New Types of Surveys      The final sections in this article will present three groups of new perspectives and designs in the domain of comparative survey research. The first group will consist of suitable cognitive models, the second one of new survey research designs and the third one of novel methods of comparative data analysis. All three groups correspond to the new theoretical, methodological and epistemological background knowledge with its focus on living systems, embedded complex cognitive models, observer-inclusion and the like. Figure 15.05 presents an overview of the structure of the final sections in this article. 54 For reasons of simplicity, the entire science landscape is divided into two major areas, called the natural and the social sciences. The latter comprised the social sciences in the conventional sense and the humanities. 490 Karl H. Müller | Niko Toš Figure 15.05 New Cognitive Models, Novel Survey Designs and Innovative Methods for Complex Social Micro-Research 15.6 Towards New Cognitive Models: Building a Logic of Over-Learned and Under-Learned Responses What follows in this section is a short overview of suitable cognitive models for survey research in accordance with the emerging background knowledge of Science II. In recent decades the life sciences in general and the diffusion of the cognitive sciences in particular have produced a broad range of cognitive models and experimental settings that provide a suitable theory and modeling background for a deeper understanding of the cognitive interactions and of the cognitive performances in survey research. A Phenomenology of Under-learned and Over-learned Responses Initially, a distinction will be introduced which should prove useful for the interpretation of the data generated under the auspices of traditional survey designs. This basic distinction is frequently used in memory research55 and, at least partially, in the research on embedded cognition.56 Here, a separation 55 On over-learned and under-learned facts in memory research, see, e.g. , Bower, 1977 or Taylor, 2004. 56 See, for example, Underwood, 1996. Towards New Frontiers in Comparative Survey research 491 can be made between over-learned and under-learned contexts. In memory research, the terms over-learned and under-learned refer essentially to the duration and to the frequency of learning processes or to the all or none character of learning.57 In embedded cognition, an over-learned or an under-learned context is characterized by the stability or the instability of action sequences. In over-learned contexts one finds either identical, sequentially fixed or constant actions irrespective of varying contexts or different pre-histories. In over-learned configurations, a single fixed operation or a fixed series of operations is required, expected and, in the case of violations, sometimes even sanctioned. In under-learned contexts, one usually finds varying, un-stable, context-dependent and fluid actions which are genuinely innovative, are created on the spot and are highly volatile, depending on the degree of openness in a particular situation. Turning to a special instance of over-learned and under-learned contexts, namely to responses, the following basic distinctions can be made. An over-learned response is triggered by a question, a command and the like and is characterized by a fixed word, a number, a sentence, a sequence of numbers or a short narrative which essentially remain unchanged in the short run and which change only under exceptional circumstances in the long run.58 Typical examples for over-learned responses include the name of a person, her or his date of birth, a person’s education level, or the employment status. Over-learned responses can be differentiated between general and specific responses. The former contain numerous instances which belong to the cognitive repertoire of minimally competent persons like one’s first name, the latter are dependent on the knowledge and performance levels of a concrete individual.59 In contrast, under-learned responses are highly volatile even in the very short run, they are subject to frequent changes in short time intervals and they are sensitive both to contexts and to pre-histories. They, too, like the specific over-learned responses, depend on the cognitive organization and repertoire of individuals. Another rather obvious feature of under-learned responses is that they are easily forgotten within a very short period of time. Table 15.09 offers several basic distinctions between under-learned and over-learned responses. 57 See, for example, Glass/Lian, 2008. 58 It should be mentioned that the production of over-learned responses has been characteristic of traditional education systems. 59 A specialist on World War II and an expert on maritime research, for example, will share very large sets of general as well as specific over-learned responses. 492 Karl H. Müller | Niko Toš TABle 15.09 An Overview of Over-Learned and Under-Learned Responses Over-Learned Responses Under-Learned Responses Single Solution Multiple Solutions Possible Stable, Fixed Highly Volatile, Unstable Trivial Non-Trivial repetitive Creative Key input Only Highly Selective of inputs encoded in long-Term Memory Short-Term Memory only Constant reproduction across Time Just in Time Productions Context-independent Context-Sensitive Path-independent Path-Dependent State-independent State-Dependent global Consistency global inconsistency Observation Errors Identifiable Only Trivial Observational errors Possible A final important distinction between over-learned and under-learned responses refers to the notion of observation and measurement errors. In over-learned responses measurement errors can be identified and, equally important, corrected. This correction can be accomplished due to the possibility for repeated measurements and due to the stable and fixed solutions in over-learned contexts. In fact, over-learned responses correspond to the fourth measurement type in Table 15.02. In sharp contrast, under-learned responses cannot be subject to measurement errors since an under-learned response misses an essential component, namely the reference values or the so-called true values. In under-learned responses one is confronted with a series of creative state-, input-, context- and history-dependent just in time responses where each of these responses, in the absence of neuro-physiological data from the cognitive neuro-sciences, must be treated as a “true value” under a set of specific, but highly varying circumstances. The distinction between over-learned and under-learned responses has been introduced, quite obviously, also with respect to surveys and survey questioning. From the dichotomy in Table 15.09 a survey is usually a mixture between over-learned and under-learned responses. The former are highly concentrated in the socio-demographic section of a survey with questions on age, gender, occupation and the like whereas the latter are distributed over most of the remaining parts of surveys, especially over all the so-called Towards New Frontiers in Comparative Survey research 493 attitudinal or evaluational segments. Nevertheless, despite the under-learned nature of most survey responses, survey measurement theory operates on the basis of randomly distorted over-learned responses and on the dichotomy between true values and measurement errors. GA-Systems as Cognitive Models of Survey Interactions In this section a brief sketch of new groups of cognitive models will be outlined which should help to grasp the intricacies of the social interactions and of the cognitive competencies needed in survey contexts. These new models belong, by and large, to the domain of evolutionary computation60 and they provide the necessary ingredients for modeling the cognitive competencies and the social interactions inherent in the questioning and answering of surveys. Using genetic algorithms (GA)61 as a specific framework, it will be assumed that both an interviewer and a respondent in a survey are organized as GA-ensembles. In the context of survey interactions, a GA-system is situated in an environment that produces a flow of verbal inputs for the GA-system which enter into the domain of internal processing. In turn, a GA-system generates verbal outputs for its environment which, once again, lead to a new round of verbal inputs for the GA-system. The basic ingredients of a GA-system have been captured in Figure 15.06. Internally, a GA-system consists of an internal message list, a set of encoded classifiers {C , C , …, C }as if → then rules and an output interface which 1 2 n generates a flow of verbal responses for the environment. John Holland, one of the inventors of GA-systems, provides the following short summary of the GA’s processing cycle. The basic execution cycle of this system proceeds as follows: 1) Place all messages from the input interface on the current message list. 2) Compare all messages on the current message list to all conditions of all classifiers and record all matches. 3) For each set of matches satisfying the condition part of some classifier, post the message specified by its action part to a new message list. 4) Replace the current message list with the new message list. 60 Evolutionary computation is an emerging field with different areas like genetic programming, genetic algorithms, evolutionary strategies or evolutionary programming. For an overview, see de Jong, 2006. 61 On generic algorithms, see, for example, Goldberg, 1989, Holland, 1986, Holland et al. , 1989, Holland, 1989, Koza, 1992, Michalewicz, 1992, Mitchell, 1996 or Rawlins, 1991. 494 Karl H. Müller | Niko Toš 5) Process the message list through the output interface to produce the system’s current output. 6) Return to step 1. [Holland et al. , 1989:106] A GA-system is basically a rule system and is equipped with three types of rules. Empirical rules are composed of different sets like categorical rules (If species S has property P , then also P ), associative rules (If species S has 1 1 2 1 property P , then activate category C ), predictive rules (If species S meets 1 1 1 species S , S will produce Action A ) or diachronic rules (If Event E occurs, 2 1 1 1 then react with Action ). 1 The second class of rules consists of inferential rules which are based on inductive generalization procedures like specialization rules, unusualness rules, law of large number heuristics or regulation schemes. The primary function of inferential rules is to produce better empirical rules. [Holland et al., 1989:43] Finally, a set of unchangeable and in-built operative rules can be seen as the innate hardware of any GA-system: Operating principles are neither learnable nor teachable. They are innate system manipulation procedures ... These include the procedures for calling up the relevant empiri cal rules for representing the environment; the bidding system by which such rules compete to construct the current representation of reality ... Other operating principles invoke some of the procedures of knowledge alteration … [ Ibid:46] Figure 15.06 A Scheme for a GA-Classifier System Towards New Frontiers in Comparative Survey research 495 The cognitive processing within a GA is based on a bidding process in which one of the GA-rules gets activated and executed. … In basic classifier systems the mechanism for activating rules is very simple: all matched rules are activated. For classifier systems with learning mechanisms it is desirable to make activation dependent on additional parameters. These include the strength of a classifier, which is a measure of its past success, and support, which is a measure of likely relevance to the current situation. [Holland et al. 1989:105f.] Following the above quotation the core of the classifier system lies in a bidding process where, given an internally encoded input from the environment, a suitable or viable GA-response is selected among a group of different available alternatives. This selection process is dependent on the strength of each of the classifiers which, in turn, is a function of their previous strength and their specificity. Aside from the bidding process, a classifier system develops, in evolutionary time, a higher degree of internal complexity by a cross-over process and by the production of new classifiers as a recombination of previously successful ones. The cognitive organization of a GA-system exhibits several remarkable properties which are necessary for their overall flexibility and for their absorption of environmental complexities. – An essential feature of evolving classifier systems lies in the multiplicity of available rules for any given input. Since fixed responses or defaults are to be considered as the exceptional cases only, a GA-system possesses a large potential repertoire of rules for any given input. – Associated with the heterogeneity of the GA-rule-composition, a classifier system is blatantly inconsistent from a logical point of view since contradictory rules are maintained within the rule corpus and conflicting rules become part of a bidding process. – The strong GA-inconsistency allows, however, for a high degree of flexibility and scope. Inconsistent rules may co-evolve in separate niches of a GA-system’s interaction with its environment. – GA-systems, due to the recombination of rules, are characterized by a drift towards increasing their internal complexity. In other words, a GA-system is building up ever higher internal complexity levels in order to interact with their environments. The environments, in turn, are subject to the same complexity drift which becomes, then the basis for a permanent struggle for higher complexity. – Another characteristic feature of a classifier system lies in the volatility of its outputs which is due internally to the probabilistic bidding processes and eternally to varying contexts. 496 Karl H. Müller | Niko Toš – Fixed and stable outputs, e.g. , over-learned actions in general or over-learned responses in particular, can emerge in GA-systems, but they should be considered as special cases which become defaults for the GA. – The distinction in over-learned and under-learned responses can be given a more precise meaning within a GA-context. Over-learned responses are executed regularly independent of contexts or pre-histories and result in a specific, pre-given response. Under-learned responses, in turn, are reserved for those situations where the inputs and the available internal rule repertoire do not match and a creative GA-response is required. – Additionally, an interesting inversion can be observed between the generality of rules and their strengths. Unlike a system of axioms where theorems can be deduced from a small set of general strong principles, general rules in the GA-architecture are usually weak and remain inactive in specific contexts. Usually, narrow, but context specific rules usually dominate over general rules.62 In general, it appears that individuating information ... has substantial power to override default assumptions based on category membership. [Holland et al. 1989:219] This short overview of GA-systems should be sufficient to use GA-architectures for the interactions in survey contexts. Towards New Cognitive Foundations for Survey Research In this section it will be assumed hypothetically that the survey interactions between respondents and interviewers take place as a dialogue between two GA-systems. Here, the part of the interviewer will not be analyzed in greater length although this side could produce interesting new results, too. The center of the investigation is occupied with the GA-respondents, their cognitive architectures and the characteristic features of the interaction processes.63 Using, once again, the distinction between over-learned and under-learned responses, it is fair to assume that for GA-organized respondents many of the survey questions, especially the ones related to attitudes or evaluations, 62 This result is of particular relevance for issues like the role of values or emotions. See, for example, Cornelius, 1996, Davidov/Schmidt/Schwartz, 2007 or Schwarz/Sudman, 1996. 63 It goes without saying that a GA-architecture in its current form is under-critical and under-complex in view of the complex cognitive tasks inherent in survey responses. Thus, the present outline should be seen as a counter-factual sketch which is focused on GA-systems and which brings to light several characteristic features of the cognitive organization of respondents where the underlying GA-architecture can offer heuristic guidelines and a weak explanatory support. Towards New Frontiers in Comparative Survey research 497 will belong to the under-learned category. The available inputs in terms of questions, the restricted options for answers, the quantitative scales, etc. are not matched directly by the internal rule repertoire which could correspond to these specific inputs directly. Thus, answers to survey questions like trust in various institutions, life satisfaction in its various aspects, to name some prominent examples, require a creative response by a GA-respondent and qualify, thus, as under-learned. Turning to the GA-respondent side only, the following characteristic features of the interaction process can be specified. One of the most important elements of the GA-based interactions lies in the sheer multiplicity of available responses by GA-respondents. The GA-organization allows a permanent recombination of new rules and, equally important, the co-existence of older and newer rules. The GA-organization acts rather gracefully and very seldom removes older rules from its rule-set. This special feature of a multiplicity of answers has been noted by survey researchers as well. As pointed out especially by John R. Zaller [1992], respondents in surveys have a much richer repertoire of different responses at their disposal. Consequently, Zaller’s response axiom states that Individuals answer survey questions by averaging across the considerations that are immediately salient or accessible to them. [Zaller 1992:49] What becomes of particular relevance here is that the high number of available responses is directly related to the under-learned situation and to the unusual requests for answering which require a creative response. In GA-language, due to the under-learned situation of a specific survey-question, a multiplicity of rules become activated since none of the available rule matches the input of a survey question. The multiplicity of responses can be shown whenever a survey question does not require a selection of a single option, but asks for assessments of each of the options sequentially. Take, for example, the question of different images of society which has been used in the Austrian Social Survey in 1993. Here, respondents were asked to which of the following four general views or images of society they could agree: to a meritocratic-conservative (a), to a “Marxist” (b), to a corporatist (c) or to a social relations-oriented view (d). As it turned out, even the contradictory pair of images, namely (a) and (b), was clearly treated in a non-contradictory manner. Only 54% of the respondents opted for one of the consistent options (a+/b-, a-/b+) and 46% agreed to inconsistent options. With respect to all four images of society, the largest single group agreed to all four images, followed by an agreement to three different images. 498 Karl H. Müller | Niko Toš Second, this multiplicity of alternatives covers only the bright side of the coin. There is a dark side to this coin, too, because this multiplicity of alternatives contradicts an implicit assumption of survey designs and especially of survey designers. Usually, survey questionnaires are developed on the tacit assumption that respondents possess a consistent attitude-system which can be captured through the items and dimensions of a multi-thematic survey. Wilson and Hodges [1992] describe this hidden assumption as the mental file view where respondents possess a well-ordered mental drawer, consisting of mental files on issues like legalized abortion, migration or trust in the police. Whenever a survey question is asked, they look for the appropriate file and report its content. However, one of the most obvious characteristics of a GA-system lies in the global as well as in the local inconsistency of its rules which differ only in their relative strength. The GA-architecture allows all rules to be treated as hypotheses, more or less confirmed, thereby stepping around difficult global consistency requirements. [Holland 1989:488] Surveys very seldom are designed to exhibit underlying inconsistencies in the attitude and belief system of respondents. The GA-architecture is structured in a way that the usual consistency relations do not apply. 80% of the respondents may be optimistic about their long-term future or about the future of the society as a whole. Any interpretation which would indicate that 100 – 80 = 20% of the population are quite concerned about the future would be extremely misleading. Normally, roughly 80% of the respondents will reply as well that they were very much concerned and worried with respect to the state of the environment and the sustainability of the mode of economic production and distribution. Global and local inconsistency is an essential element of a GA-architectures and this feature is reflected strongly in survey responses, too. Apart from the global inconsistency of the GA-architecture, the third general characteristic deals with the logic of under-learned responses which does not comply with classical logic. Assume, in line with two valued logic, the availability of conceptual pairs like true/false, confirmed/rejected, allow/ forbid, etc. As Hippler and Schwarz [1986], for example, demonstrated the conceptual pairs allow/forbid and not forbid/not allow are treated in surveys not as equivalent, but quite distinctively since allow and forbid are consistently seen as stronger statements than not allowing and not forbidding. Moreover, Hippler and Schwarz provide an un-intended support for the distinction between under-learned and over-learned responses because this observed asymmetry only holds for respondents with weakly developed attitudes Towards New Frontiers in Comparative Survey research 499 (under-learned), not for persons with very strong attitudes (over-learned). Another feature of under-learned responses is that the usual transitivity relations a > b, b > c → a> c do not hold.64 Furthermore, a tacit assumption especially in value-related survey research lies in the importance of general values as a determinant for preferences or specific routines. However, the GA-architecture reveals an interesting inversion between rule strength and generality. In brief, the most general rules turn out to be the weakest ones, the most specific rules, due to their context specificity, usually become the strongest ones. Fourth, an underlying GA-architecture points to the important role of path dependencies, context effects and to the sensitivity to small input variations in survey-interactions. This special part has been studied in survey research extensively under labels like response effects, question order effects and the like.65 The important point to be emphasized from a GA-perspective, however, lies in the simultaneity of a large variety of contexts, of a sensitivity to small input variations and of different pre-histories which cannot be isolated or de-composed in an un-ambiguous matter. Contexts may vary with interviewers and with specific events during a survey interaction, the wording of questions and, more importantly, the subsequent interpretation may vary with the pre-history of survey respondents prior to a survey interaction, etc. The simultaneity of these variations cannot be controlled sequentially which, in turn, raises insurmountable problems for any comprehensive theory of measurement errors. Aside from a multiplicity of potential replies, the overall inconsistency of these multiple alternatives, the non-classical logic underlying survey responses and context or history effects, volatility becomes a fifth essential feature of GA-based survey interactions. From a GA-based perspective, this volatility is composed of four different components. Initially, the volatility is partly due to the probabilistic bidding process which constitutes a necessary component of variation. Another part of the volatility comes from the necessity of producing creative responses which, by itself, must be considered as a non-trivial and inherently instable process. Additionally, the bidding process is, due to varying contexts of survey interactions, highly complex which, once again, adds to the volatility of responses. Finally, a fourth important aspect with respect to the volatility of under-learned responses lies in the scales 64 For a wonderful article on the topology of nervous nets and the non-transitivity of values, see McCulloch 1980. 65 See, for example, Bradburn/Mason, 1964, Cronbach, 1946, Krosnick/Alwin, 1987, Martin, 1964 or Schwarz/Hippler, 1991. 500 Karl H. Müller | Niko Toš which are available in many survey questions. When confronted with a scale between 0 and ten for example, respondents in under-learned situations usually are indifferent with respect to a broad range of values which adds another element in the overall volatility of responses.66 Numerous examples have been generated which point to the instability and the variation in responses. John R. Zaller gives an illuminating example in terms of changes in wordings. A record instance of the effect of changes in question wording may be a New York Times poll in 1983 which found that public support for a ‘freeze’ on nuclear weapon production varied between 18 and 83 percent, depending on how the issue was framed. [Zaller 1992:29] A particularly striking example comes from the German Welfare Survey 1984 in which one of the most central questions, namely overall life satisfaction was, by mistake, asked twice in an identical fashion. The correlation between both responses was only 0.60. Sixth, a highly fascinating feature in the creative nature of under-learned survey responses reveals itself by focusing on GA-architectures. Usually, the input side in survey interactions consists of a series of verbal items which, in conjunction, should be taken into account by a GA-system. But in a GA-architecture, it cannot be taken for granted that the entire input has been used in the process of producing an answer. Thus, one is suddenly confronted with the possibility that other forms of understanding outside the intended domain of survey researchers were operative in generating a specific answer. For obvious reasons, the term “unintended consequences” of a survey question points to the possibility that respondents did not reply to an intended question, but to a different one which was composed of selective elements of the original one. Take, for example, a seemingly straightforward question like a self-assessment of one’s overall position in society and a measurement method, using a scale from 1 to 10.67 More than 80% of the unskilled workers positioned themselves above the societal average (6 and higher), in contrast to roughly 57% of the skilled workers. In GA-language it seems very likely that many respondents in the unskilled group produced an answer with respect to their subjective overall position from 1 to 10, and not to the intended societal positions from 1 to 10. 66 While this indifference range may vary between respondents, it can be assumed that each respondent is indifferent with respect to at least two values on such a scale. 67 This example comes from the Austrian Social Survey 1993 and is discussed at length in Müller, 1998. Towards New Frontiers in Comparative Survey research 501 Another feature of unintended effects comes into play whenever under-learned items like work satisfaction and a scale from 0 to 10 are asked in various, seemingly different dimensions. As has been shown in Part II of this book, the answers to different dimensions of work satisfaction tended to be quite similar across Europe, despite very heterogeneous working and living conditions. In GA-language, the seemingly different dimension of work satisfaction became subject to a default operation which, among other results, would have produced very similar outcomes, had the list of dimensions been longer than the existing one. The feature of defaults is of critical importance for the morphological approach because these defaults offer empirical support for the similarity relations, essential for the morphological approach, between seemingly different dimensions. The seventh feature of under-learned survey responses leads outside the GA-domain proper and to the domain of long-term and short-term memory. Under-learned responses, due to their under-learned nature, do not enter into long-term memory and are, thus, quickly erased from the memory screen in a very short period of time. Within survey settings, twenty to thirty minutes are sufficient for having completely forgotten a specific under-learned response to a survey question.68 These seven GA-based features conclude the section on a new cognitive background theory for survey interactions. As a final remark, a short quotation will be presented which, once again, points to the intricacies and the complexities in producing survey responses. In a meanwhile classical description Daniel C. Dennett provides us with an interesting sketch of the bidding process inherent in spoken language. Dennett uses a pandemonium of demons which starts to operate internally and, finally, produces a spoken phrase. Here, Dennett’s description will be slightly adapted to survey responses. The pandemonium of demons will, after a person has been asked a survey question, start to operate internally and will result in a winner of the bidding process which, then, gets executed. Let’s consider a survey question like “How satisfied are you with your life?” and ... a pandemonium of word-demons ... First we go into vocal noise-making mode Beeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeep … 68 Following Saris/Gallhofer, 2007:220, twenty minutes within a conventional survey interview are sufficient for practically forgetting an under-learned response to a survey question. 502 Karl H. Müller | Niko Toš … The internal ‘noise’ excites various demons in us who begin trying to modulate the horn in all sorts of random ways by interfering with its stream. The result is gibberish, but at least it’s English gibberish (in English speakers) Yabba-dabba-doo-fiddledy-dee-tiddly-pom-fi-fi-fo-fum … But before any of this embarrassing stuff actually hits the outside world, further demons, sensitive to patterns in the chaos, start shaping it up into words, phrases, clichés … And so, how about that? baseball, don’t you know, in point of fact, strawberries, happenstance, okay? That’s the ticket. Well, then ... which incites demons to make further serendipitous discoveries, augmented by opportunistic shaping, yielding longer bits of more acceptable verbage, until finally a whole sentence emerges: I’m going to knock your teeth down your throat! Fortunately, however, this gets aside, unspoken, since at the same time (in parallel) other candidates have been brewing and are now in the offering, including a few obvious losers, such as You big meany! and Read any good books lately? and a winner by default, which gets spoken: [I don’t understand this question!] ... We can suppose that all of this happens in swift generations of ‘wasteful’ parallel processing, with hordes of anonymous demons and their hopeful constructions never seeing the light of the day ... [Dennett 1991:237p.]69 15.7 Towards New Forms of Survey Designs and Data Analysis Towards the end of this article a few general remarks should open up new directions and new frontiers for survey designs as well as for novel forms in the analyses of survey data. Introducing Visual Surveys In recent years, new survey designs were proposed under the name of deliberative or interactive polling70 which, by adding additional information and background materials, attempt to move the under-learned responses 69 The sentence in brackets has been implemented by us, in the original version the final sentence is: ‘Your feet are too big!’ 70 On deliberative polling, see, e.g. , Sturgis/Roberts/Allum, 2005. Towards New Frontiers in Comparative Survey research 503 towards the over-learned group. These designs organize surveys as a learning process but run into an interesting trade-off between high representativeness/ low deliberation and low representativeness/high deliberation. With respect to new survey designs, one of the core differences between Science I and Science II was the dichotomy between general and universal laws on the one hand and pattern recognition and pattern formation on the other hand. A new survey design which is based on this distinction moves the verbal responses from categories and numbers to the domain of pattern recognition and pattern production. In essence, these surveys are composed of measurement and scaling designs which produce two types of outputs. The first output group consists of a series of patterns where the task of respondents lies in the selection of a single, most appropriate pattern. The second output group is composed of visual tasks which end up in a specific pattern, generated by the respondents themselves. (For an overview, see Müller, 2004) Two examples should be sufficient to demonstrate this type of visual survey. The first example deals with a pattern recognition task and is exemplified by Figure 15.07. Here, one sees a collection of visual patterns which represent one’s life course so far. The task of respondents is twofold. Either they select one of the pre-given patterns or they produce a new one. Figure 15.07 Measurement Method as Pattern Recognition Figure 15.08, in turn, shifts to an elementary form of pattern production where respondents are asked to evaluate five different domains and rate their overall importance for them. In sharp contrast to the sequential mode of questioning in traditional surveys, this pattern producing task generates a configuration where each new alternative A , A , A and A is placed within 2 3 4 5 a self-specified context of an already existing reference point (A ). A self-1 generated configuration like the one in 15.08, while under-learned, too, provides an abundance of structural and also of quantitative information on the relative strength of each of the five alternatives. 504 Karl H. Müller | Niko Toš According to our experience with a small number of visual surveys so far, shifting from verbal responses to pattern formation and pattern recognition could and should become an interesting new way for survey designs. In our assessment, visual or pattern-based surveys could capture the complexities of the cognitive architectures of respondents in a much better way than the Science I-based measurement methods with their restricted scales and pre-arranged verbal responses. Figure 15.08 Measurement Method as Pattern Formation A = Subjective Rank Ordering of Items A for Domain (Domain : Values, Preferences, Goals, etc.) i i j j Towards New Forms of Survey Data Analysis New cognitive models like the ones on over-learned and under-learned responses, GA-based architectures and new survey designs, for example the visual designs introduced above, should lead, finally, to new ways of data analysis, too. Referring back to the start of Chapter 14, one of the challenges for comparative research is to bring in datasets which, so far, remained outside the domain of comparative analysis. For example, directly incomparable data between regions or countries were thought to be the strongest argument Towards New Frontiers in Comparative Survey research 505 for the generation of comparable cross-regional or cross-country data-sets. Thus, what is currently urgently needed are the outlines for a new road to comparative research which could become of special relevance for large quantities of unused data outside the current data bases for comparative analyses. These datasets ..... (round and round ... and round ......) Bibliography Acham, K. (1974), Analytische Geschichtsphilosophie. Eine kritische Einführung. Freiburg: Alber. Adam, B., Beck, U., Loon, J.v. (2000) (eds.), The Risk Society and beyond. Critical Issues for Social Theory. London: Sage Publications. Adams, F., Aizawa, K. (2008), The Bounds of Cognition. Oxford:Wiley Blackwell. Adelman, P.K., Antonucci, T.C., Cohan, S.E., Coleman, L.M. (1989), Empty nest, cohort, and employment in the well-being of midlife women. Sex Roles Vol. 20, 173–189. Adorno, T.W. (1978) (ed.), Der Positivismusstreit in der deutschen Soziologie. 6th Edition. Darmstadt: Luchterhand. Ajzen, I. (1989), Attitude Structure and Behavior, in: A.R. Pratkanis, S.J. Breckler, A.G. Greenwald (eds.), Attitude Structure and Function. Hillsdale:Lawrence Erlbaum, 241–247. Ajzen, I., Fishbein, M. (1980), Understanding Attitudes and Predicting Social Behavior. The Expectancy-Value Model, in: Actes du Congrès de l’AFM. Poitiers:w.p., 681– 695. Almond, A.G., Verba, S. (1963), The Civic Culture. Political Attitudes and Democracy in Five Nations. Princeton: Princeton University Press. Andrews, F.M. (1984), Construct Validity and Error Components of Survey Measures: A Structural Equation Approach, in: Public Opinion Quarterly 48, 409–442. Apparala, M.L., Reifman, A., Munsch, J. (2003), Cross-National Comparision of Attitudes Toward Father’s and Mothers’ Participation in Household Tasks and Childcare, Sex Roles 48(5/6). Atkinson, J., Meager, N. (1986), Changing Working Patterns: How Companies Achieve Flexibility to Meet New Needs. London: National Economic Development Office. Averill, J.R., More, T.A. (2000), Happiness, in: M. Lewis, J.M. Haviland-Jones (eds.), Handbook of Emotions, 663–676, New York: The Guilford Press. Bagozzi, R.P., Yi, Y. (1991), Multitrait-Multimethod Matrices in Consumer Research, in: Journal of Consumer Research 17, 426–439. Bahry, D., Silver, B.D. (1990), Soviet Citizen Participation on the Eve of democratization. American Political Science Review Vol. 84, No. 3, 821–847. Barnes, S.H., Simon, J. (1998), The Postcommunist Citizen. Budapest: Erasmus Foundation. Barša, P. (2003), Politické teorie multikulturalismu. Brno: Centrum pro studium demokracie a kultury (2. vyd.). Baršová, A. (2005), Integrace přistěhovalců v Evropě: od občanské integrace k multikulturalismu a zpět? Příspěvek na konferenci Soudobé spory o multikulturalismus a politiku identity (24. 1. 2005). Dostupné na internetu: [http://www.migraceonline.cz/studie_f.shtml?x=223608] cit. 20. 9. 2005. 508 Bibliography Bastelaer, A. v., Lemaitre, G., et al. (1997), The Definition of Part-time Work for the Purposes of International Comparison. Paris: OECD. Bauer, T.K., Lofstrom, M., Zimmerman, K.F. (2000), Immigration Policy, Assimilation of Immigrants and Native’s sentiments towards Immigrants: Evidence from 12 OECD-Countries. IZA Discussion Paper No. 187, Bonn: IZA. Beck, U. (1992), Risk Society. Towards a New Modernity. London: Sage. Beck, U. (1997), The Reinvention of Politics. Cambridge: Polity Press. Beck, U. (2000), The Brave New World of Work. Oxford: Polity Press. Beck, U., Beck-Gernsheim, E. (2002), Individualization. Institutionalized Individualism and Its Social and Political Consequences. London:Sage Publications. Ben-Shlomo, Y., Davey Smith, G., Shipley, M., Marmot, M.G. (1993), Magnitude and causes of mortality differences between married and unmarried men. Journal of Epidemiology and Community Health 47, 200–205. Berger, P.L., Kellner, H. (1979), Marriage and Social Contruction of Reality, in: P.L. Berger, Facing up to Modernity, 27–17. Harmondsworth: Penguin Books. Bernard, J. (1975), Comments on Glenn’s Paper. Journal of Marriage and the Family 37, 601–602. Better Health Channel: www.betterhealth.vic.gov.au. Blalock, H.M. Jr. (1968), The Measurement Problem: A Gap between Languages of Theory and Research, in: H.M. Blalock, A.B. Blalock (eds.), Methodology in the Social Sciences. London: Sage Publications, 5–27. Blalock, H.M. Jr. (1990), Auxiliary Measurement Theories Revisited, in: J.J. Hox, J. de Jong-Gierveld (eds.), Operationalization and Research Strategy. Amsterdam: Swets and Zeitlinger, 33–49. Blatný, M., Osecká, L., Macek, P. (1998), Osobnostní předpoklady životní spokojenosti a sebehodnocení: Temperament a zaměřenost, in: Svoboda, M., Blatný, M., Sociální procesy a osobnost, 13–18, Brno: Psychologický ústav AV ČR. Blatný, M., Osecká, L. (1998), Zdroje sebehodnocení a životní spokojenosti: osobnost a strategie zvládání. Československá psychologie 42, 385–394. Bliese, P.D. (2006a), Multilevel Modeling in R (2.2). (R package version 0.5). Bliese, P.D. (2006b), Multilevel: Multilevel Functions. (R package version 2.2). Bliese, P.D., Hanges, P.J. (2004), Being Both Too Liberal and Too Conservative: The Perils of Treating Grouped Data as Though They Were Independent. Organizational Research Methods, 7(4). Bodenhausen, G.V., Lambert, A.J. (2003) (eds.), Foundations of Social Cognition. A Festschrift in Honor of Robert S. Wyer. New York: Lawrence Erlbaum. Boje, T., Strandh, M. (2003), Sweden, in: C. Wal ace, Country Contextual Reports. Vienna: Institute for Advanced Studies, HWF Project. Booth, M. (2006), Public Engagement and Practical Wisdom, in: S. Paulin (ed.), Community Voices: Creating Sustainable Spaces. Perth: University of Western Australia Press, 12–26. Bower, G.H. (1977), Psychology of Learning and Motivation, Vol. 11. Advances in Research and Theory. New York:Academic Press. Bibliography 509 Boyer, R. (2008), The Quest for Theoretical Foundations of Socio-Economics: Epistemology, Methodology or Ontology?, in: Socio-Economic Review 6, 733– 746. Bradburn, N.M., Mason, W.M. (1964), The Effect of Question Order on Responses, in: Journal of Marketing Research 1, 57–61. Bradley, H., Erikson, M., et al. (2000), Myths at Work. Oxford: Polity Press. Brown, S.L. (2000), The Effect of Union Type on Psychological Well-being: Depression Among Cohabitors Versus Marrieds. Journal of Health and Social Behavior 41, 241–255. Bryk, A.S., Raudenbush, S.W. (1992), Hierarchical Linear Models: Applications and Data Analysis Methods. Newbury Park, CA: SAGE Publications. Bunting, B., Adamson, G. (2000), Assessing elaibility and Validity in the Context of Planned Incomplete Data Structures for Multitrait-Multimethod Models, in: A. Ferligoj, A. Mrvar (eds.), Developments in Survey Methodology 15. Ljubljana: FDV, 37–53. Burjanek, A. (2001), Xenophobia among the Czech Population in the Context of Post-Communist Countries and Western Europe. Czech Sociological Review 9 (1), 53–67. Calvert, G.A., Spence, C., Stein, B.E. (2004) (eds.), The Handbook of Multisensory Processes. Cambridge: MIT Press. Calvin, W.H. (1996), The Cerebral Code. Cambridge: The MIT Press. Calvin, W.H., Bickerton, D. (2001), Lingua Ex Machina: Reconciling Darwin and Chomsky with the Human Brain. Cambridge: The MIT Press. Campbell, J. (1984), Grammatical Man. Information, Entropy, Language, and Life. Harmondsworth: Penguin. Carson, L., Hartz-Karp, J. (2005), Adapting and Combining Deliberative Designs: Juries, Polls, and Forums, in: J. Gastil, P. Levine (eds), The Deliberative Democracy Handbook: Strategies for Effective Civic Engagement in the Twenty-First Century. San Francisco: Jossey-Bass, 120–138. Castles, S. (2005), Hierarchical Citizenship in a World of Unequal Nation-States. Political Science and Politics 38(4) (October), 689–693. Cazes, S., Nesporova, A. (2003), Employment Protection Legislation (EPL) and its Effects on Labour Market Performance. Geneva: ILO. Central Intelligence Agency (CIA) (2002), The World Factbook. http://www.cia.gov/ cia/download2002.htm. Central Intelligence Agency (CIA) (2006), The World Factbook. Central Intelligence Agency (CIA). Cheung, Y.B. (1997), Can marital selection explain the differences in health between married and divorced people? From a longitudinal study of a British Birth Cohort. Public Health 112, 113–117. Clark J.A., Legge, J.S. jr. (1997), Economics, Racism and Attitudes toward Immigration in the New Germany. Political Research Quarterly 50 (4), 901–917. Clark, A. (2008), Supersizing the Mind: Embodiment, Action, and Cognition. New York: Oxford University Press. 510 Bibliography Cohen, L.J. (1999), Changing paradigms of Citizenship and the Exclusiveness of the Demos. International Sociology, 14(3), 245–268. Conditions, E. F. f. t. I. o. L. a. W. (2002), Working conditions in candidate countries and the European Union. Resume. Luxembourg: Office for Official Publications of the European Communities. Converse, P. E. (1964), The Nature of Belief Systems in Mass Publics, in: D.E. Apter (ed.), Ideology and Discontent. New York: Free Press, 206–261. Coombs, R.H. (1991), Marital Status and Personal Well-Being: A literature Review. Family Relations 40, 97–102. Cooper, C.L. (1996) (ed.), Handbook of Stress, Medicine, and Health. Boca Raton: CRC Press. Cornelius, R.R. (1996), The Sciences of Emotion. Research and Tradition in the Psychology of Emotions. Englewood Cliffs: Prentice Hal . Corrin, C. (1992) (ed.), Superwomen and the Double Burden. London: Scarlet Press. Council of Europe (2004), Recent Demographic Developments in Europe. Council of Europe Publishing. Couper, M.P. (2000), Web-Surveys. A Review of Issues and Approaches, in: Public Opinion Quarterly 64, 464–494. Cousins, C., Tang, N. (2003), The UK Country Contextual Reports, in: C. Wal ace, HWF Research Report 2: Country Contextual Reports 7–42 Vienna: Institute for Advanced Studies. Creswell, J.W., Plano Clark, V.L. (207), Designing and Conducting Mixed Methods Research. Thousand Oaks: Sage Publications. Cronbach, L.J. (1946), Response Sets and Test Validity, in: Educational and Psychological Measurement 6, 475–494. Dahl, R. (1995), Demokracie a její kritici. Praha: Victoria publishing. Dalton, R.J. (2000), Citizen Attitudes and Political Behavior. C omparative Political Studies Vol. 33, No. 6/7, 812–40. Dalton, R.J., Kuechler, M. (1990), Challenging the Political Order: New Social and Political Movements in Western Democracies. New York: New York University Press. Damasio, R.A. (1994), DescartesÉrror. Emotion, Reason and the Human Brain. New York:Grosset/Putnam Book. Damasio, R.A. (2003), Looking for Spinoza. Joy, Sorrow, and the Feeling Brain. London: William Heinemann. Dangschat, J., Blasius, J. (1994) (eds.), Lebensstile in den Städten. Konzepte und Methoden. Opladen: Leske+Budrich. Davidov, E., Schmidt, P., Schwartz, S. (2007), Bringing Values back in. The Adequacy of the European Social Survey to Measure Values in 20 Countries, in: Public Opinion Quarterly 3, 420–445. Deacon, T.W. (1997), The Symbolic Species. The Co-evolution of Language and Brain. New York: W.W.Norton. Bibliography 511 de Leeuw, E.D., Hox, J.J. (2003), The Use of Meta-Analysis in Cross National Studies, in: J.A. Harkness, F.J.R. van de Vijver, P.P. Mohler (eds.), Cross-Cultural Survey Methods. New York: John Wiley&Sons, Inc. Dennett, D.C. (1991), Consciousness Explained. Boston: Little, Brown and Company. Dex, S. (1997), Flexible Employment. London: Palgrave/Macmillans. Diekmann, A., Preisendörfer, P. (2001), Umweltsoziologie. Eine Einführung. Reinbek: Rowohlt. Diekmann, A. (2007), Empirische Sozialforschung. Grundlagen, Methoden, Anwendungen. Reinbek: Rowohlt. Diener, E., Lucas, R.E. (2000), Subjective Emotional Well-Being, in: M. Lewis, J.M. Haviland-Jones (eds), Handbook of emotions, 325–337. New York: The Guilford Press. Dijkstra, A.G., Plantega, J. (2001) (eds.), Gender and Economics: A European perspective. London/New York: Routledge. Dillman, D.A. (2000), Mail and Internet Surveys. The Tailored Design Method. New York: Wiley. Doeringer, P.B., Piore, M.J. (1971), Internal Labour Markets and Manpower Analysis. Lexington: USA, D.C. Heath. Downs, A. (1957), An Economic Theory of Democracy. New York: Harper. Drbohlav, D. (2001a), Mezinárodní migrace obyvatelstva – pohyb i pobyt (Alenky v kraji divů), 17–30, in: T. Šišková, (ed.), Menšiny a migranti v České republice: My a oni v multikulturní společnosti 21. století. Praha: Portál. Drbohlav, D. (2001b), Evropa a proces mezinárodní migrace, Integrace 1 (4), 5–10. Eagly, A.H., Chaiken, S. (1993), The Psychology of Attitudes. New York: Harcourt Brace Jovanovich. Earle, J.R., Smith, M.H., Harris, C.T., Longino, C.F. (1998), Women, Marital Status, and Symptoms of Depression in a Midlife National Sample. Journal of Women and Aging 10, 41–7. Easton, D. (1975), A Re-Assessment of the Concept of Political Support. British Journal of Political Science 5, 435–457. Edelman, G.M. (1987), Neural Darwinism. New York: Basic Books. Edelman, G.M. (1990), The Remembered Present. A Biological Theory of Consciousness. New York: Basic Books. Edelman, G.M. (1992), Bright Air, Brilliant Fire. On the Matter of the Mind. New York: Basic Books. Edelman, G.M. (2007), Second Nature. Brain Science and Human Knowledge. Yale: Yale University Press. Edwards, J.R., Bagozzi, R.P. (2000), On the Nature and the Direction of Relationships between Constructs and Methods, in: Psychological Methods 5, 155–174. Elsasser, W.M. (1998), Reflections on a Theory of Organisms. Holism in Biology. Baltimore: The Johns Hopkins University Press (orig. 1987). Epstein, J.M. (2006), Generative Social Sciences. Studies in Agent-Based Computational Modeling. Princeton: Princeton University Press. 512 Bibliography Espenshade, T.J. (1987), Population Replacement and Immigrant Adaptation: New Issues Facing the West. Family Planning Perspectives, 19(3) 115–118. Esposito, J.P., Rothgeb, M. (1997), Evaluating Survey Data: Making the Transition from Pretesting to Quality Assessment, in: P. Lyberg, P. Biemer, L. Col ins, E. de Leeuw, C. Dippo, N. Schwarz, D. Trewin (eds.), Survey Measurement and Process Quality. New York: Wiley, 541–571. European Commission (2005), Europe in Figures. Eurostat Yearbook 2005. Luxembourg Office for Official Publications of the European Communities. European Social Survey. http://www.europeansocialsurvey.com Eurostat. http://epp.eurostat.cec.eu.int Fabrigar, L.R., Krosnick, J.A., MacDougal , B.L. (2006), Attitude measurement: Techniques for Measuring the Unobservable, in: M. C. Green, S. Shavitt, T.C. Brock (eds.), Persuasion: Psychological Insights and Perspectives. Thousand Oaks: Sage Publications. Falk, R. (1994), The Making of Global Citizenship, in; B. van Steenbergen (ed.), The Condition of Citizenship. London: Thousand Oaks, New Delhi: Sage, 127–140. Faulks, K. (1999), Political Sociology. A Critical Introduction. Edinburgh: Edinburgh University Press. Felstead, A., Jewson, N. (1999) (eds.), Global Trends in Flexible Labour. Critical Perspectives on Work and Organisations. London: Hants, Macmillans. Fishbein, M., Ajzen, I. (1975), Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research. Reading: Addison Wesley. Flick, U., Kardorff, E.v., Steinke, I. (2005), Qualitative Methoden. Ein Handbuch. Reinbek:Rowohlt. Foerster, H.v. (1984), Principles of Self-Organization in a Socio-Managerial Context, in: H. Ulrich, G.J.B. Probst (eds.), Self-Organization and Management of Social Systems. Berlin: Springer, 7–24. Foerster, H.v., Glasersfeld, E.v. (1999), Wie wir uns erfinden. Eine Autobiographie des radikalen Konstruktivismus. Heidelberg: Carl Auer-Systeme Verlag. Foerster, H.v. (2003), Understanding Understanding. Essays on Cybernetics and Cognition. New York: Springer. Foerster, H.v., Müller, K.H. (2003), Action without Utility. An Immodest Proposal for the Cognitive Foundation of Behavior, in: Cybernetics and Human Knowing 3–4 (Vol. 10), 27–50. Freedom House (2003), Combined Average ratings. Independent Countries. (URL: http://freedomhouse.org/template.cfm?page=236&year=2003) (access: 8 June 2006). Frey, S.B. (2003), Flexible citizenship for a global society. Politics, Philosophy & Economics, 2(1), 93–114. Gang I.N., Rivera-Batiz, F.L., Yun, M.S. (2002), Economic Strain, Ethnic Concentration and Attitudes Toward Foreigners in the European Union, IZA Discussion Paper No. 578, Bonn: IZA. Ganßman, H. (2000), Labour market flexibility, social protection and unemployment. European Societies 2(3), 243–269. Bibliography 513 Gazzaniga, M.S., Bizzi, E., Black, I.B. (2004) (eds.), The Cognitive Neurosciences III. Cambridge: The MIT Press. Gibbs, R.W. (2005), Embodiment and Cognitive Science. Cambridge: Cambridge University Press. Glasersfeld, E.v. (1997), Radikaler Konstruktivismus. Ideen, Ergebnisse, Probleme. Frankfurt: Suhrkamp. Glass, A.L., Lian, A. (2008), Evidence of Al -or None Learning from a Repetition Detection Task, in: M.A. Gluck, J.R. Anderson, S.M. Kosslyn (eds.), Memory and Mind. A Festschrift for Gordon H. Bower. New York: Lawrence Erlbaum Associates, 123–140. Goguen,J., Varela, F. (1979), Systems and Distinctions: Duality and Complementarity, in: International Journal of General Systems 5, 31–43. Goldberg, D. (1989), Genetic Algorithms in Search, Optimization, and Machine Learning. Bonn: Addison-Wesley. Goldman, N., Korenman, S., Weinstein, R. (1995), Marital status and health among the elderly. Social Science and Medicine, 40, 1717–1730. Gove, W.R., Hughes, M., Briggs-Style, C. (1983), Does Marriage Have Positive Effects on the Psychological Well-Being of the Individual?. Journal of Health and Social Behavior 24, 122–131. Grant, J., Hoorens, S., Sivadasan, S., van het Loo, M., DaVanzo, J., Hale, L., et al. (2004), Low Fertility and Population Ageing. Causes, Consequences, and Policy Options. RAND Corporation. (Prepared for the European Commission). Groenendijk, J., Strokhof, M. (1997), Questions, in: J. van Benthem, A. ter Meulen (eds.), Handbook of Logic and Language. Amsterdam: Elsevier, 1055–1124. Groß, J., Börensen, C. (2009), Wie valide sind Verhaltensmessungen mittels Vignetten? Ein methodischer Vergleich von faktoriellem Survey und Verhaltensbeobachtung, in: P. Kriwy, C. Gross (eds.), Klein aber fein!: Quantitative empirische Sozialforschung mit kleinen Fallzahlen. Wiesbaden: VS Verlag, 149–178. Günther, G. (1962), Cybernetic Ontology and Transjunctional Operations, in: M.C. Yovits et al., op. cit. , 313–392. Haas, B. (2003), The reconciliation of work and family obligations: a comparison between Austria, the Netherlands and Sweden, in: C. Wal ace, HWF Survey Comparative Report Volume 2: Thematic Papers. Vienna: Institute for Advanced Studies, HWF Project 2. Habermas, J. (1981), Theorie des kommunikativen Handelns (Bd. 1: Handlungsrationalität und gesellschaftliche Rationalisierung; Bd. 2: Zur Kritik der funktionalistischen Vernunft). Frankfurt: Suhrkamp. Habermas, J. (1984), Vorstudien und Ergänzungen zur Theorie des kommunikativen Handelns. Frankfurt: Suhrkamp. Hamplová, D. (2004), Životní spokojenost: rodina, práce a další faktory, Sociologické studie/ Sociological studies 04:06. Praha: SOÚ AV ČR. Hamplová, D. (2006), Životní spokojenost, štěstí a rodinný stav v 21 evropských zemích. Sociologický časopis, Czech Sociological Review 42, 35–55. Hand, D.J. (2004), Measurement. Theory and Practice. London: Arnold. 514 Bibliography Hantrais, L., Mangen, S. (1996) (eds.), Cross-National Research Methods in the Social Sciences. London: Pinter. Harkness, J.A., Van de Vijver, E.J.R., Mohler, P.P. (2003) (eds.), Cross-Cultural Survey Methods. Hoboken: Wiley. Heater, D. (1990), Citizenship: The Civic Ideal in World History, Politics and Education. London & New York: Longman. Helferich, C. (2005), Die Qualität qualitativer Daten: Manual für die Durchführung qualitativer Interviews. Wiesbaden: VS Verlag. Henshaw, J.M. (2006), Does Measurement Measure up? How Numbers Reveal and Conceal the Truth. Baltimore: The Johns Hopkins University Press. Hippler, H.J., Schwarz, N. (1986), Not Forbidding Isn’t Allowing: The Cognitive Basis oft he Forbid-Allow Asymmetry, in: Public Opinion Quarterly 1, 87–96. Hirschman, C. (1994), Why Fertility Changes. Annual Review of Sociology, Vol. 20, 203–233. Hochgerner, J. (1998), Flexibilisierung durch Telearbeit, in: H.-G. Zilian, J. Flecker (ed.), Flexibilisierung – Problem oder Lösung? . Berlin: Sigma, 175–192. Hoffmeyer, J.H.P., Wolf, C. (2003)(eds.), Advances in Cross-National Comparison. A European Working Book for Demographic and Socio-Economic Variables. New York: Kluwer. Hofstadter, D.R (41982), Gödel-Escher-Bach. An Eternal Golden Braid. Harmondsworth: Penguin. Hofstadter, D.R., Dennett, D.C. (1982) (eds.), The Mindś I. Phantasies and Reflections on Self and Soul. Harmondsworth: Penguin. Hofstadter, D.R. (1985a), Metamagical Themas. Questing for the Essence of Mind and Matter. New York: Basic Books. Hofstadter, D.R. (1985b), Gödel, Escher, Bach – Ein Endloses Geflochtenes Band. Stuttgart: Klett-Cotta. Hofstadter, D.R., Fluid Analogies Research Group (1995), Fluid Concepts and Creative Analogies. Computer Models of the Fundamental Mechanisms of Thought. New York: Basic Books. Hofstatder, D.R. (1997), Le Ton beau des Marot. In Praise of the Music of Language. New York: Basic Books. Holden, K.C., Kuo, H.D. (1996), Complex Marital Histories and Economic Well-being: The Continuing Legacy of Divorce and Widowhood as the HRS Cohort Approaches Retirement. The Gerontologist 36, 383–390. Hol and, J.H., Holyoak, K.J., Nisbett, R.E., Thagard, P.R. (1989), Induction: Processes of Inference, Learning, and Discovery. Cambridge: MIT Press. Hol and, J.H. (1989), Using Classifier Systems to Study Adaptive Nonlinear Networks, in: D.L. Stein (ed.), Lectures in the Sciences of Complexity. Redwood City: Addison-Wesley, 463–499. Hol and, J.H. (1995), Hidden Order. How Adaptation Builds Complexity. Reading: Addison-Wesley. Bibliography 515 Hol ingsworth, J.R., Müller, K.H., Hol ingsworth, E.J. (2002) (eds.), Advancing Socio-Economics. An Institutionalist Perspective. Lanham: Rowman & Littlefield Publishers, Inc. Hol ingsworth, R.J., Müller, K.H., Hol ingsworth, E.J., Gear, D.M. (2008), Socioeconomics and a New Scientific Paradigm, in: H. Flam, M. Carson (eds.), Rule Systems Theory. Applications and Explorations. Frankfurt: Peter Lang Verlag, 39– 56. Hol ingsworth, J.R., Müller, K.H. (2008), Transforming Socio-Economics with a New Epistemology, in: Socio-Economic Review 3 (6), 395–426. Horwitz, A.V., White, H.R., Howell-White, S. (1996), Becoming Married and Mental Health: A Longitudinal Study of a Cohort of Young Adults. Journal of Marriage and the Family 58, 895–907. Horwitz, A.V., Scheid, T.L. (1999) (eds.), A Handbook for the Study of Mental Health. Social Contexts, Theories, and Systems. Cambridge:Cambridge University Press. Hox, J.J. (1997), From Theoretical Concept to Survey Questions, in: P. Lyberg, P. Biemer, L. Col ins, E. de Leeuw, C. Dippo, N. Schwarz, D. Trewin (eds.), Survey Measurement and Process Quality. New York: Wiley, 47–70. Hox, J. (2002), Multilevel Analysis. Techniques and Applications. Mahwah NJ: Lawrence Erlbaum Associates. http://www.oecd.org/document/36/0,2340, en_2825_494553_2515108_1_1_1_1,00.html http://www.statistics.gr/eng_tables/S1100_SAP_5_euro03.htm Huntington, S.P. (1991), The Third Wave: Democratization in the Late Twentieth Century. London: University of Oklahoma Press. Hunter, J.E., Schmidt, F.L. (2004), Methods of Meta-Analysis: Correcting Error and Bias in Research Findings. London: Sage Publications. Huws, U., Podro, S., Gunnarsson, E., Weijers, T., Arvanitaki, K., Trova, V. (1996), Teleworking and Gender, Sussex: IES Report. Inglehart, R. (1997), Modernization and Postmodernization, New Jersey: Princeton University Press. Inglehart, R., Basanez, M., Moreno, A. (1998), Human Values and Beliefs: A Cross Cultural Source-book. World Value Survey 1990–1993, Ann Arbor, MI: The University of Michigan Press. Inglehart, R., Catterberg, G. (2002), Trends in Political Action: The Developmental Trends and the Post-Honeymoon Decline. International Journal of Comparative Sociology Vol. 43, No.3–5, 300–316. Jaccard, J., Turrisi, R., Wan, C.K. (1990), Interaction effects in Multiple Regression. London: Sage. Jager, A. (2003) (ed.), Country Contextual Report: The Netherlands, in: C. Wal ace (ed.), HWF Research Report #2, 43–92, HWF Series of Project Research Reports. Vienna: Institute for Advanced Studies. Jones-Corre, M.A., Leal, D.L. (2001), Political Participation: Does Religion Matter? Political Research Q uaterly, Vol. 54, No. 4, 751–770. 516 Bibliography Jong, de K.A. (2006), Evolutionary Computation: A Unified Approach. Cambridge: The MIT Press. Joung, I.M.A., Stronks, K., Van De Mheen, H., Van Poppel, F.W.A., van Der Meer, J.B.W., Machenbach, J.P. (1997), The contribution of intermediary factors to marital status differences in self-reported health, Journal of Marriage and the Family 59, 476–490. Jowel, R., Central Co-ordinating Team (2003), European Social Survey 2002/2003. Technical Report, Centre for Comparative Social Surveys, London: City University. Kaase, M., Newton, K. (1995), Beliefs in Government. Oxford: Oxford University Press. Kaase, M., Newton, K., Toš, N. (1995), Zaupanje v vlado in Zaupanje Slovencev v demokratični sistem, Ljubljana: Liberalna akademija in Znanstvena knjižnica FDV. Kalmijn, M., Poortman, A. (2006), His or her divorce? The gendered nature of divorce and its determinants, European Sociological Review, 22, 201–214. Kamens, D.H. (1988), Education and Democracy:A Comparatice Institutional Analysis. Sociology of Education, Vol. 61, No. 2, 114–127. Kauffman, L.H. (2005), Eigen-Form, in: Kybernetes: Heinz von Foerster in memoriam, Vol. 1/2 (Guest editor: Alexander Riegler), 129–150. Keller, B., Seifert, H. (1995) (eds.), Atypische Beschäftigung: verbieten oder gestalten? Köln: Böhlau. Kepinska, E. (2004), Recent Trends in International Migration. The 2004 SOPEMI Report for Poland. Viz. http://www.iss.uw.edu.pl/osrodki/cmr/wpapers/ pdf/056.pdf Koza, J.R. (1992), Genetic Programming. On the Programming of Computers by Means of Natural Selection. Cambridge: The MIT Press. Krosnick, J.A., Alwin, D.F. (1987), An Evaluation of a Cognitive Theory of Response-Order Effects in Survey Measurement, in: Public Opinion Quarterly 2, 201–219. Krosnick, J.A., Abelson, R.P. (1991), The Case for Measuring Attitude Strength in Surveys, in: J.M. Tanur (ed.), Questions about Questions. Inquiries into the Cognitive Bases of Surveys. New York: Russell Sage Foundation, 177–203. Krosnick, J.A., Judd, C.M., Wittenbrink, B. (2005), The Measurement of Attitudes, in: D. Albarracin, B.T. Johnson, M.P. Zanna (eds.), The Hand book of Attitudes. Mahwah: Lawrence Erlbaum Associates, 21–76. Kučera, M. (1994), Populace České republiky 1918–1991. Praha: SoÚ, Česká demografická společnost. Kveinen, E. (2000), Citizenship in Post-Westphalian Community: Beyond External Exclusion. Paper presented at European Consortium for Political Research. Copenhagen 14–19 April 2000. Lamnek, S. (2005), Qualitative Sozialforschng: Lehrbuch, vierte Aufl. Weinheim:Beltz-Verlag. Lakoff, G., Nunez, R.E. (2000), Where Mathematics Comes From: How the Embodied Mind Brings Mathematics into Being. New York: Basic Books. Bibliography 517 Lass, J., Saris, W.E., Kaase, M. (1997), Sizes of the Different Effects: Coverage, Mode and Non-Response, in: W.E. Saris, M. Kaase (eds.), Eurobarometer. Measurment Instruments for Opinions in Europe. Mannheim: ZUMA-Nachrichten Special Edition, 73–86. Leeuw, de E.D., Hox, J.J., Dillman, D.A. (2008) (eds.), International Handbook of Survey Methodology. New York: Lawrence Erlbaum Associates. Letki, N. (2004), Socialization for Participation? Trust, Membership, and Democratization in East-Central Europe. Political Research Quaterly Vol. 54, No. 4, 665–679. Lewis, J. (2002), Gender and welfare state change, European Societies 4(4), 331–358. Lijphart, A., (1999), Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries. London: Yale University Press. Lillard, L.A., Waite, L.J. (1995), Till Death Do Us Part: Marital Disruption and Mortality, The American Journal of Sociology 100, 1131–1156. Lipset, S.M., Schneider, W. (1987), The Confidence Gap: Business, Labor and Government in the Public Mind. Baltimore: Johns Hopkins University Press. Lodovici, M.S. (1999), The Dynamics of Labour Market Reform in European Countries, in: G. Esping-Andersen, M. Regini, Why Deregulate Labour Markets? Oxford: Oxford University Press. Lorant, V., Deliege, D., Eaton, W., Robert, A., Philippot, P., Ansseau, M. (2003), Socioeconomic Inequalities in Depression: A Meta-Analysis, American Journal of Epidemiology, 157(2). Lord, F., Novick, M.R. (1968), Statistical Theories of Mental Test Scores. Reading:Addison Wesley. Lucas, A. (2006), amap: Another Multidimensional Analysis Package. (R package version 0.7–1). Lumley, T. (2006), rmeta: Meta-analysis. (R package version 2.14). Luo, Y., Waite, L.J. (2005), The Impact of Childhood and Adult SES on Physical, Mental, and Cognitive Well-Being in Later Life, The Journal of Gerontology 60B, 93–101. Luskin, R.C., Fishkin, J.S., Jowell, R. (2002), Considered Opinions: Deliberative Pol ing in Britain, in: British Journal of Political Science 32, 455–487. Lykken, D., Tellegen, A. (1996), Happiness is a stochastic phenomenon, Psychological Science 7, 186–189. Mackay, D.M. (1962), Self-Organization in the Time Domain, in: M.C. Yovits et al., op.cit. , 37–48. Marks, N.F. (1996), Flying Solo at Midlife: Gender, Marital Status, and Psychological Well-Being, Journal of Marriage and the Family 58, 917–932. Marsh, A. (1977), Protest and Political Consciousness. London: Sage. Marsh, A., Kaase, M. (1979), Background of Political Action, in: S.H. Barnes, M. Kaase (eds.), Political Action: Mass Participation in Five Western Democracies. Beverly Hills: Sage, 56–71. Martin, J. (1964), Acquiescence – Measurement and Theory, in: British Journal of Social and Clinical Psychology 3, 216–225. 518 Bibliography Martin P., Widgren, J. (2002), International Migration: Facing the Challenge, Population bulletin, 57(1), 3–40. Mastekaasa, A. (1992), Marriage and psychological well-being: Some evidence on selection into marriage, Journal of Marriage and the Family 54, 901–911. Mastekaasa, A. (1994), Marital Status, Distress, and Well-being: An international comparison, Journal of Comparative Family Studies 25, 183. Maturana, H.R. (1985), Erkennen: Die Organisation und Verkörperung von Wirklichkeit. Ausgewählte Arbeiten zur biologischen Epistemologie, second edition. Braunschweig: Friedr. Vieweg&Sohn. Maturana, H.R. (1985), The Mind Is Not the Head, in: Journal of Social and Biological Structures 4 , 308–310. Maturana, H.R., Varela, F.J. (1987), Der Baum der Erkenntnis. Die biologischen Wurzeln des menschlichen Erkennens. Bern: Scherz. Mayda, A.M. (2004), Who is Against Immigration? A Cross-Country Investigation of Individual Attitudes toward Immigrants, IZA Discussion Paper, No.1115, Bonn: IZA. Mayntz, R. (2008), Networks and Self-Organization: Dissecting the Model of ‘Complex Networks’, in: Socio-Economic Review 6, 750–754. McAdam, D., McCarthy, J.D., Zaid, M.N. (1996), Comparative Perspectives on Social Movements. Cambridge: Cambridge University Press. McCulloch, W.S., Arbib, M.A., Cowan, J.D. (1962), Neurological Models and Integrative Processes, in: M.C. Yovits et al., op. cit. , 49–59. McCulloch, W.S. (1980), A Heterarchy of Values Determined by the Topology of Nervous Nets in: W.S. McCulloch, Embodiments of Mind. Cambridge: The MIT Press, 40–45. McDonough, P., Shin, D.C., Moises, J.A.(1998), Democratization and Participation: Comparing Spain, Brazil, and Korea, Journal of Politics, Vol. 60, No. 4, 919–953. Medgyesi, M. (2002), Hungary, in: C. Wal ace (ed.), Critical Review of Literature and Discourses about Flexibility, HWF Research Report No. 1, 137–154, Vienna: Institute for Advanced Studies. Michalewicz, Z. (1998), Genetic Algorithms + Data Structures = Evolution Programs. Berlin: Springer. Miech, R.A., Shanahan, M.J. (2000), Socioeconomic Status and Depression over the Life Course, Journal of Health and Social Behavior 41, 162–176. Miethe, T.D. (1985), The Validity and Reliability of Value Measurements, in: Journal of Psychology 119, 441–453. Milbrath, L.W. (1965), Political Participation. Chicago: Rand McNally & Company. Mills, J.A. (2000), Control: A History of Behavioral Psychology. New York: New York University Press. Minsky, M. (1990), Mentopolis. Stuttgart: Klett-Cotta. Mitchell, M. (1996), An Introduction to Genetic Algorithms. Cambridge: The MIT Press. Müller, A., Müller, K.H. (2007) (eds.), An Unfinished Revolution? Heinz von Foerster and the Biological Computer Laboratory, 1958–1976. Vienna: edition echoraum. Bibliography 519 Müller, K.H. (1998), Sozioökonomische Modelle und gesellschaftliche Komplexität. Vermittlung und Designs. Marburg: Metropolis-Verlag. Müller, K.H. et al. (2002), Wissenschaftliche Institute in Österreich: Schlüsselfaktoren für Innovationsgrade und Forschungsperformanzen. Final Report for the BMBWK. Wien: WISDOM. Müller, K.H. (2004), Symbolische Formen als Vehikel trans-kultureller Kommunikation. Das Beispiel der inetrnationalen Survey-Forschung, in: Signs, Texts, Cultures. Conviviality from a Semiotic Point of View / Zeichen, Texte, Kulturen. Konvivialität aus semiotischer Perspektive. Internet-Journal 15. (http://www.inst. at/trans/15Nr/01_2 /01_2 inhalt_part1 _15.htm). Müller, K.H. (2008), The New Science of Cybernetics. The Evolution of Living Research Designs, Vol. I: Methodology. Vienna: edition echoraum. Müller, K.H. (2008a), A Farewell to Long Waves. Substituting Cyclical Approaches in Innovation and Technology Research with a RISC-Framework , in: RISC-Research Paper, Nr.1. Ljubljana: RISCSeries. Muetzefeldt, M., Smith, G. (2002), Civil Society and Global Governance. The Possibilities for Global Citizenship. Citizenship Studies, 6(1), 55–75. Nelson, M.K., Smith, J. (1999), Working Hard and Making Do. Surviving in Small Town America. Berkely, Los Angeles, London: University of California Press. Nishiyama, S. (2002), A nalyzing an Aging Population – A Dynamic General Equilibrium Approach, CIRJE, Faculty of Economics, University of Tokyo. Noe, A. (2009), Out of Our Heads. Why You Are Not Your Brain, and Other Lessons from the Biology of Consciousness. Hill and Wang. Nørretranders, T. (1998), The User Illusion: Cutting Consciousness down to Size. Harmondsworth:Penguin Books. Norris, P. (1999), Critical Citizens, Oxford: Oxford University Press. Norwegian Social Science Data Services (NSD), European Social Survey – data set ESS1–2002, ed. 6.0 (URL: http://ess.nsd.uib.no/webview/index.jsp). Nowotny, H. (2008), Bargaining, not Borrowing: On Problem Choice and Problem Space, in: Socio-Economic Review 6 , 754–759. OECD Employment Outlook (1999), Chapter 2: Employment Protection and Labour Market Performance. Offe, C. (1994), Der Tunnel am Ende des Lichts, Frankfurt: Campus Verlag. Olson, J.M., Hafer, C.L. (1996), Affect, Motivation and Cognition in Relative Deprivation Research, in: R.M. Sorrentino, E.T. Higgins (eds.), Handbook of Motivation and Cognition: The Interpersonal Context, Vol. 3, New York: Guilford, 85–117. Olson, M. (2000), The Logic of Collective Action. Cambridge: Harvard University Press. O’Reilly, J., Fagan, C. (1998), Part-time Prospects. Oxford: Routledge. Otte, G. (2004), Sozialstrukturanalyse mit Lebensstilen. Wiesbaden: VS Verlag. Painter, J. (2002), Multi level citizenship, identity and regions in contemporary Europe, in: J. Anderson (ed.), Transnational Democracy. Political spaces and border crossing. London: Routledge, 93–110. 520 Bibliography Palombo, S.R. (1999), The Emergent Ego: Complexity and Coevolution in the Psychoanalytic Process. Madison: International Universities Press. Pask, G. (1962), Interaction between a Group of Subjects and an Adaptive Automation to Produce a Self-Organizing System for Decision Making, in: M.C. Yovits et al., op. cit. , 283–311. Perrons, D. (1998), Flexible Working and the Reconciliation of Work and Family Life – A New Form of Precariousness, Brussels: European Commssion, Directorate General Employment and Social Affairs. Perry, R.L., Robertson, J.D. (2002), Comparative Analysis of Nations. Quantitative Approaches. Cambridge: Westview Press. Pfau-Effinger, B. (2003), Transformation of Social Policies in the Socio-cultural Context of European Societies, European Societies 7, 321–347. Pfeffer, T. (2001), Das “zirkuläre Fragen” als Forschungsmethode zur Luhmannschen Systemtheorie. Heidelberg: Carl Auer. Piatelli-Palmarini, M. (1994), Inevitable Illusions. How Mistakes of Reason Rule Our Minds. New York: John Wiley&Sons. Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D. (2006), NLME: Linear and nonlinear mixed effects models. R package version 3.1–77. Pinker, S. (1997), How the Mind Works. Harmondsworth: Penguin. Plasser, F., Ulram, P. (1996), Measuring Political Culture, in: F. Plasser, A. Pribersky (eds.), Political Culture in Eastern Central Europe, Aldershot: Avebury. Plotkin, H. (1997), Evolution in Mind. An Introduction to Evolutionary Psychology. Harmondsworth: Penguin. Pohl, R.F. (2004)(ed.), Cognitive Illusions. A Handbook on Fallacies and Biases in Thinking, Jusgement and Memory. Hove: Psychology Press. Pollert, A. (1988), The “Flexible Firm”: Fixation or Fact, Work, Employment and Society 2(3), 281–316. Pollock, J.S. (1989), How to Build a Person: A Prolegomenon. Cambridge: The MIT Press. Pruchno, R.A., Hayden, J.M. (2000), Interview Modality: Effects on Costs and Data Quality in a Sample of Older Women, in: Journal of Aging and Health 12 , 3–24. Punch, K.F. (1998), Introduction to Social Research. Quantitative and Qualitative Approaches. London: Sage. Putnam, R.D. (1993), Making Democracy Work: Civic Traditions In Modern Italy, Princeton: Princeton University Press. R Development Core Team (2006), R: A Language and Environment for Statistical Computing. Vienna, Austria. Rabušic L., Burjanek, A. (2003), Imigrace a imigrační politika jako prvek řešení české demografické situace? Brno: Výzkumný ústav práce a sociálních věcí. Rabušic, L. (2000), Koho Češi nechtějí? O symbolické sociální exkluzi v české společnosti, Sociální studia 5, 67–65. Ratey, J.R. (2001), A User’s Guide to the Brain: Perception, Attention, and the Four Theaters of the Brain. New York: Pantheon Books. Bibliography 521 Rawlins, G.J.E. (1991), Foundations of Genetic Algorithms. San Matteo: Morgan Kaufman Publishers. Regini, M. (2000), The dilemmas of labour market regulation, in: G. Esping-Andersen, M. Regini (eds.), Why Deregulate Labour Market? , Oxford: Oxford University Press. Renzetti C.M., Curran D.J. (2003), Women, men, and society, Boston: Allyn and Bacon. Rescher, N. (1998), Predicting the Future. Albany: State University of New York Press. Riboud, M., Silva-Jauregui, C., et al. (2001), Does Eurosclerosis Matter? Institutional Reform and Labour Market Performance in Central and Eastern European Countries in the 1990s. Washington: World Bank Working Paper. Richter, R. (2005), Die Lebensstilgesellschaft. Wiesbaden: VS Verlag. Rifkin, J. (1995), The end of work, New York: G.P. Putnam’s Sons. Rizopoulos, D. (2006), ltm: Latent Trait Models under IRT. (R package version 0.6–1). Robbins, P., Aydede, M. (2008) (eds.), The Cambridge Handbook of Situated Cognition. Cambridge: Cambridge University Press. Rogers, R.G. (1995), Marriage, sex, and mortality, Journal of Marriage and the Family 57, 515–526. Rose, R. (1984), Understanding Big Government. London: Sage. Rose, R., Haerpfer, Ch. (1966), New Democracies, Barometer IV: A10-Nation Survey, Studies in Public Policy, Glasgow: University of Strathclyde. Rosen, R. (2005), Life Itself: A Comprehensive Inquiry into the Nature, Origin, and Fabrication of Life. New York: Columbia University Press. Rosenblatt, F. (1962), A Comparison of Several Perception Models, in: M.C. Yovits et al., op. cit. , 463–484. Rössel, J. (2005), Plurale Sozialstrukturanalyse. Eine handlungstheoretische Rekonstruktion der Grundbegriffe der Sozialstrukturanalyse. Wiesbaden: VS Verlag. Ross, C.E., Mirowsky, J., Goldsteen, K. (1990), The Impact of the Family on Health: The Decade in Review, Journal of Marriage and the Family 52, 1059–1078. Roth, G. (2002), Das Gehirn und seine Wirklichkeit: Kognitive Neurobiologie und ihre philosophischen Konsequenzen. Frankfurt:Suhrkamp. Rus, V., Toš, N. (2005), Vrednote Slovencev in Evropejcev, Analiza vrednotnih orientacij Slovencev ob koncu stoletja, Univerza v Ljubljani, Fakulteta za družbene vede, IDV, CJMMK, Dokumenti SJM 13, Ljubljana, 333–369. Ryan, J., Hughes, M., Hawdon, J. (1998), Marital Status, General-life Satisfaction and the Welfare State, International Journal of Comparative Sociology 39, 224–236. Ryckman, R.M. (2000), Theories of Personality. Belmont: Wadsworth Publishers. Salt, J. (2001), Current Trends in International Migration in Europe. Dostupné na internetu: http://www.coe.int/T/E/Social_Cohesion/Migration/ Documentation/Publications_and_reports. Sarafino, E.P. (42002), Health Psychology. Biopsychosocial Interactions. New York: John Wiley&Sons. Saris, W.E. (2003), Multitrait-Multimethod Studies, in: J.A. Harkness, E.J.R. Van de Vijver, P.P. Mohler (2003) (eds.), Cross-Cultural Survey Methods. Hoboken: Wiley, 265–274. 522 Bibliography Saris, W.E., Sniderman, P.M. (2004) (eds.), Studies in Public Opinion: Attitudes, Nonattitudes, Measurment Error and Change. Princeton: Princeton University Press. Saris, W.E., Gal hofer, I.N. (2007), Design, Evaluation, and Analysis of Questionnaires for Survey Research. Hoboken: Wiley Interscience. Scheler, M. (1994), Ressentiment. Milwaukee, WI: Marquette University Press. Schmidt, J.S. (1987) (ed.), Der Diskurs des Radikalen Konstruktivismus. Frankfurt: Suhrkamp. Schumpeter, J. (2004), Kapitalismus, socialismus a demokracie. Brno: CDK. Schützenberger, M.P. (1962), On Probabilistic Push-Down Storages, in: M.C. Yovits et al., op. cit. , 205–213. Schwarzer, G. (2005), meta: Meta-Analysis. (R package version 0.8). Schwarz, N., Hippler, H.J. (1991), Response Alternatives: The Impact of Their Choice and Presentation Order, in: P. Biemer, R.M. Groves, L.E. Lyberg, N.H. Mathiowetz, S. Sudman (eds.), Measurement Errors in Surveys. New York: John Wiley&Sons, 41–56. Schwarz, N., Sudman, S. (1996) (eds.), Answering Questions: Methodology for Determining Cognitive and Communicative Processes in Survey Research. San Francisco: Jossey-Bass. Schwenk, O.G. (1996) (ed.), Lebensstil zwischen Sozialstrukturanalyse und Kulturwissen-schaft. Opladen: Leske+Budrich. Sennet, R. (1998), The Corrosion of Character: the personal consequences of work in the new capitalism, London: W.W.Norton Shore, C. (2004), Whither European Citizenship? Eros and Civilization Revisited, European Journal of Social Theory 7(1), 27–44. Sicherl, P., Stanovnik, T., et al. (2003), Slovenia, in: C. Wal ace, Country Contextual Reports. Vienna: Institute for Advanced Studies, HWF Project. Siemieńska, R., (1990), Płeć, zawód, polityka. Udział kobiet w życiu publicznym w Polsce, Warszawa: Instytut Socjologii Uniwersytetu Warszawskiego. Siemienska R., (2006), Zarządzanie czasem – budżet czasu w Krajowy system monitorowania równego traktowania kobiet i mężczyzn, Wydawnictwo Naukowe Scholar. Sik, E. (1993), From the Second Economy to the Informal Economy, Journal of Public Policy, 12(2), 153–175. Sik, E. (2003), Hungary, in: C. Wal ace, Country Contextual Reports. Vienna: Institute for Advanced Studies, HWF Project. Smith, T. (1997), Factors Relating to Misanthropy in Contemporary American Society, Social Science Research Vol. 26, No. 1, 170–196. Sniderman, P.M., Therbiault, S. (2004), The Structure of Political Argument and the Logic of Issue Framing, in: W.E. Saris, P.M. Sniderman (eds.), Studies in Public Opinion: Attitudes, Nonattitudes, Measurment Error and Change. Princeton: Princeton University Press, 133–166. Solomonoff, R.J. (1962), Training Sequences for Mechanized Induction, in: M.C. Yovits et al., op. cit. , 425–434. Bibliography 523 Sornette, D. (2008), Interdisciplinarity in Socio-Economics, Mathematical Analysis and the Predictability of Complex Systems, in: Socio-Economic Review 6 , 759–770. Stack, S. (1990), New micro-level data on the impact of divorce on suicide, 1959–1980: A test of two theories, Journal of Marriage and the Family 54, 327–334. Stack, S., Eshleman, J.R. (1998), Marital Status and Happiness: A 17-Nations Study, Journal of Marriage and the Family 60(2), 527–537. Stanculescu, M., Berevoescu, S., Berevoescu, I. (2003), Romania. Country Contextual Reports. Vienna: Institute for Advanced Studies, HWF Project. Standing, G. (1999), Global Labour Flexibility, London: Macmillans. Sternberg, R.J., Wagner, R.K. (1994) (eds.), Mind in Context. Interactionist Perspectives on Human Intelligence. Cambridge: Cambridge University Press. Sturgis, P., Roberts, C., Al um, N. (2005), A Different Take on the Deliberative Pol : Information, Deliberation, and Attitude Constraint, in: Public Opinions Quarterly 1, 30–65. Taylor, K.M. (2004), The Effects of Overlearning on Long-Term Retention. Thesis:Dept. of Psychology [College of Arts and Sciences. University of South Florida]. Tobiasen, M. (2000), Global Citizenship – An Analytical Conceptualisation. Paper presented at European Consortium for Political Research. Copenhagen 14–19 April 2000. Topf, R. (1995), Participation Beyond Voting, in: H.D. Klingemann, D. Fuchs (eds.), Citizens and the State, Oxford: Oxford University Press, 52–91. Toš, N., Mohler, P.P., Malnar, B. (1999)(eds.), Modern Society and Values. A Comparative Analysis Based on ISSP Project. Ljubljana/Mannheim: FSS/ZUMA. Toš, N. (1999, 2004), Vrednote v prehodu II. in III, Slovensko javno mnenje 1990– 1998 in 1999–2004, Dokumenti SJM 6 in 10, FDV, IDV, CJMMK, Ljubljana. Toš, N. (2007), Potek demokratične institucionalizacije, SJM 2006/2, Poročilo 2, E-dokumenti SJM, CJMMK, Ljubljana. Access: www.cjm.si/e-dokumenti. Toš, N. (2007), (Ne-)zaupanje v institucije, Potek demokratične institucionalizacije v Sloveniji (1991–2006), Teorija in praksa, 3–4, 367–395. Tourangeau, R., Rips, L.J., Rasinski, K. (2000), The Psychology of Survey Response. Cambridge: Cambridge University Press. Turner, S.B. (2000), Liberal Citizenship and Cosmopolitic Virtue, in: A. Vandenberg (ed.), Citizenship and Democracy in a Global Era. New York: St. Martins Press, 18–32. Underwood, G. (1996), Implicit Cognition. New York: Oxford University Press. UNDP, Human development report. http://hdr.undp.org/statistics/data Utterbeck, J.N. (1989), Dynamics of Industrial Innovation. Bal inger Publishers. Utterbeck, J.N. (1996), Mastering the Dynamics of Innovation: How Companies Can Seize Opportunities in the Race of Technological Change. Harvard: Harvard Business School Press. Václavíková, A. (2002), Cizinci na trhu práce v České republice na přelomu tisíciletí, in: O. Šrajerová, (ed.), Národnostní menšiny na přelomu tisíciletí. Sborník z mezinárodní vědecké konference. Opava: Slezský ústav Slezského zemského muzea, 271–275. 524 Bibliography Varela, F.J. (1976a), Not One, not Two: Position Paper for the Mind-Body Conference, in: Co-Evolution Quarterly 3, 62–67. Vecernik, J., Stepankova, P. (2002), The Czech Republic, in: C.Wal ace, Critical Review of Literature and Discourses about Flexibility, Vienna: Institute for Advanced Studies. Visser, P.S., Bizer, G.Y., & Krosnick, J.A. (2006), Exploring the latent structure of strengthrelated attitude attributes, in: M. Zanna (ed.), Advances in Experimental Social Psychology. New York: Academic Press. Waite, L.J., Gal agher, M. (2000), The Case for Marriage, New York: Broadway Books. Walker I., Pettigrew, T.F. (1984), Relative Deprivation Theory: An Overview and Conceptual Critique, in: British Journal of Social Psychology 23 , 301–310. Wal ace, C. (2002), Opening and Closing Borders: Migration and Mobility in East-Central Europe, Journal of Ethnic and Migration Studies 28 (4), 603–625. Wal ace, C. (2003) (ed.), Comparative Contextual Report. Vienna: Institute for Advanced Studies. Watzlawick, P. (1981) (ed.), Die erfundene Wirklichkeit. Wie wissen wir, was wir zu wissen glauben? Beiträge zum Konstruktivismus. Münchjen: Piper. Watzlawick, P., Krieg, P. (1991) (eds.), Das Auge des Betrachters. Beiträge zum Konstruktivismus. Festschrift für Heinz von Foerster. München: Piper. Weber, M. (1980), Wirtschaft und Gesellschaft. Grundriß der Verstehenden Soziologie, 5th edition. Tübingen: Mohr-Siebeck. Widrow, B. (1962), Generalization and Information Storage in Networks of Adaline ‘Neurons’, in: M.C. Yovits et al., op.cit. , 435–461. WIKIPEDIA: www.wikipedia.org Wilmoth, J., Koso, G. (2002), Does Marital History Matter? Marital Status and Wealth Outcomes Among Pre-retirement Adults, Journal of Marriage and the Family 64, 254–268. Wilson, T.D., Hodges, S.D. (1992), Attitudes as Temporary Constructions, in: L. Martin, A. Tesser (eds.), The Construction of Social Judgment. Hillsdale: Erlbaum, 37–65. Wojciszke B., (2004), Kobiety i mężczyźni: odmienne spojrzenie na różnice, Gdańskie Wydawnictwo Psychologiczne. Yovits, M.C., Jacobi, G.T., Goldstein, G.D. (1962) (eds.), Self-Organizing Systems. Washington: Spartan Books. Zaller, J.R. (1992), The Nature and Origins of Mass Opinion. Cambridge: Cambridge University Press. Zeilinger, A. (2005), Einsteins Schleier. Die neue Welt der Quantenphysik. Munich: Goldmann. Zentralarchiv für Empirische Sozialforscung (ZA) (2006), International Social Survey Programme (ISSP 2004): Citizenship (Codebook ZA Study No. 3950, 1st Complete ed. April 2006). Köln: Zentralarchiv für Empirische Sozialforschung an der Universität zu Köln. (Bachemer p. 40, D-50931 Köln Germany) (URL: http://www.gesis.org/ZA/). Bibliography 525 Zilian, H.-G., Flecker, J. (1998) (eds.), Flexibilisierung – Problem oder Lösung. Berlin: Sigma Verlag. Zimmermann, K.F. (1995), Tackling the European migration problem, Journal of Economic Perspective 9 (2), 45–692. Authors Christian Bischof (bischof@wisdom.at), social scientist, has graduated at the University of Vienna and works as data services manager and data analyst at the Vienna Institute for Social Science Documentation and Methodology (WISDOM). His work focuses on data archiving documentation methods and the research of the social stratification of cultural participation and cultural tastes. Jana Chaloupková (jana.chaloupkova@soc.cas.cz) is a member of the research team Value Orientations in Society at the Institute of Sociology of the Academy of Sciences in the Czech Republic. Currently, she is a Ph.D. candidate in sociology at the Faculty of Philosophy and Arts, Charles University in Prague. Her main research interests lie in family studies, particularly in work-family balance, as well as values and attitudes within society. Anna Domaradzka (anna.domaradzka@uw.edu.pl) is a Ph.D. student and full-time assistant at the Institute for Social Studies at Warsaw University, Poland. Her main areas of interest are social activity, non-governmental organizations and women’s movements, as well as gender issues in the private and public spheres. She is now finishing her Ph.D. dissertation which focuses on the leaders of women’s organizations in Poland. Zoltán Fábián (fabian@tarki.hu) is the director of databank and IT department at TÁRKI Social Research Center, Hungary since 1999. He was a fellow of the Hungarian Academy of Sciences, and became a candidate of sociology in 1999. His dissertation was published in a book titled Tekintélyelvűség és előítéletek (Prejudices and authoritarianism). He is the author and co-author of a number of articles and book chapters in the field of social psychology, social structure and information society. In 2003 he was the coordinator of a large-scale information society monitoring study in Hungary. He coordinated the social and market research on the introduction of digital terrestrial television platform (DVB-T) in Hungary in 2002. Fábián has been one of the coordinators of World Internet Project (WIP), Hungary since 2001. 528 Authors He has also been a consultant to the World Bank as a HHP database expert in 1998. Between 2000 and 2004 he was the Hungarian coordinator of the CHER (Consortium of Household Panels for European Socio-economic Research) project which was financially supported by the European Union in the Fifth Framework Program. Since 2004 he has been the president of the Political Behavior Section of the Hungarian Sociological Association and a member of the Sociology Subcommittee of the Hungarian Accreditation Committee. Mitja Hafner-Fink (mitja.hafner-fink@fdv.uni-lj.si) is Assistant Professor of Social Science Methodology at the Faculty of Social Sciences University of Ljubljana. He is a researcher at the Public Opinion and Mass Communication Research Centre; since 2004 he is Slovenian representative in the ISSP. He was president of Slovenian Sociological Association (1998-2001). His specialities include (political) participation, citizenship, (national) identity, and social science methodology. Dana Hamplová (dana.hamplova@soc.cas.cz) is a research fellow at the Institute of Sociology of the Academy of Sciences of the Czech Republic. In 2004, she graduated from the Faculty of Arts and Philosophy at the Charles University with a Ph.D. in Sociology. In 2006, she received a post-doctoral fellowship at McGill University. Her work focuses on comparative research, social demography, social stratification and religion. Jindřich Krejčí (jindrich.krejci@soc.cas.cz) is head and co-founder of the Sociological Data Archive (SDA) at the Institute of Sociology of the Academy of Sciences of the Czech Republic. He graduated in the field of sociology from the Faculty of Social Sciences of the Charles University in Prague. He is a member of the International Association for Social Science Information Systems and Technology (IASSIST) and the Steering Group for Social Sciences and Humanities at the European Strategy Forum on Research Infrastructures. He has been a principal investigator for three research and science infrastructure projects sponsored from Czech national grant agencies. In research he focuses on the survey methodology, research on data quality, and the study of political attitudes. He also participates in international comparative surveys of the International Social Survey Programme (ISSP) and the European Social Survey (ESS). He is the editor of book Kvalita vyzkumu volebnich preferenci (The Quality of Survey Research on Electoral Authors 529 Preferences) and author of number of articles in scientific journals and books. At the Faculty of Social Sciences of the Charles University he teaches courses in data management, survey methodology and methods of social data analyses. Slavko Kurdija (slavko.kurdija@guest.arnes.si) is assistant Professor of Methodology at the Department of Sociology on Faculty of Social Sciences in Ljubljana and researcher at Public Opinion and Mass Communications Research Center at the Faculty. His research and academic activities involve designing and fielding quantitative face to face and telephone opinion (national and cross-national) surveys. He works especially on data management and data analysis on a wide variety of research topics: social values; political attitudes and trust in institutions. Tomáš Lebeda (tomas.lebeda@soc.cas.cz) is a senior researcher for the team Value Orientations in Society at the Institute of Sociology of the Academy of Sciences of the Czech Republic. He teaches courses in electoral studies and comparative government at the Faculty of Social Science, Charles University Prague. He specializes in the study of elections, electoral systems, strategic voting, voting preferences and election forecasting. Vlado Miheljak (vlado.miheljak@fdv.uni-lj.si) studied psychology at the Faculty of Arts in Ljubljana, and at the Freie Universität in West Berlin and is actually Associate Professor of social and political psychology and a researcher at the Centre for Public Opinion Research and at the Centre for Social Psychology of the Faulty of Social Sciences, University of Ljubljana. His research is concentrated on the public opinion, voting behavior, youth population and values. Karl H. Müller (mueller@wisdom.at) is head of WISDOM, Austria’s infrastructural centre for the social sciences and humanities, president of the International Heinz von Foerster Society and scientific advisor of a research program RISC (Rare Incidents, Strong Consequences) at the University of Ljubljana, Slovenia. His main research interests range from issues in complex modelling within the social sciences and from interdisciplinary analyses of innovation processes in science, technology and economy to the history and the current potential 530 Authors of inter- and transdisciplinary research, to the frontiers of second order cybernetics and radical constructivism and, finally, to the newly emerging risk-potentials for contemporary societies in general. He has conducted and organized more than seventy national and international research projects and has published more than 150 scientific articles, books and reports in international as well as national journals or publishing houses. His recent publications include Advancing Socio-Economics (2002, together with J. Rogers Hollingsworth and Ellen Jane Hollingsworth (eds.)), An Unfinished Revolution? Heinz von Foerster and the Biological Computer Laboratory 1958–1976 (2007, together with Albert Müller (eds.)), Gordon Pask: Philosopher Mechanic (2007, together with Ranulph Glanville (eds.)), The New Science of Cybernetics. The Evolution of Living Research Designs, vol. I: Methodology. Wien:edition echoraum, “Methodologizing Radical Constructivism. Recipes for RC-Designs in the Social Sciences”, in: Constructivist Foundations 1(4), 50 – 61, Hollingsworth, J.R., K.H. Müller (2008), „Transforming Socio-Economics with a New Epistemology“, in: Socio-Economic Review 3 (6), 395 – 426, Hollingsworth, J.R., K.H. Müller, E.J. Hollingsworth (2008), “The End of the Science Superpowers”, in: Nature 454, 412 – 413. Filip Raciborski (raciborski@post.pl) is a Ph.D. student at the Institute for Social Studies, Warsaw University (ISS UW). His dissertation addresses the problem of gender in parliamentary elections. He also works for the Medical University of Warsaw (WUM) as a lecturer in the area of methodology of population studies. In addition to working with students, he manages the fieldwork for the largest and most important epidemiological research organization in Poland (ECAP), which focuses on allergies. It is a continuation of the European research, ECRHS II. He also collaborates with the Centers for Disease Control and Prevention (CDC) and the World Health Organization (WHO) on the Global Adult Tobacco Survey (GATS). Armin Reautschnig (armin@wisdom.at), information designer, has graduated at the Univeristy of Applied Sciences FH JOANNEUM in Graz at the department for Information Design and is currently a PH.D. student at Institute for Sociology at the University of Vienna. His work at the Wiener Institute for Social Science Documentation and Methodology (WISDOM) focuses on the development of new ways of data visualization such as the visualization program WISDOMIZE. Authors 531 Petra Šalamounová (petra.salamounova@soc.cas.cz) is a research associate for the team Value Orientations in Society at the Institute of Sociology of the Academy of Sciences of the Czech Republic. She graduated with a degree in demography from Charles University, Prague, where she is currently pursuing her Ph.D., also in demography. She studies demographic behavior, particularly non-marital fertility. Endre Sik (sik@tarki.hu), DSc, habil. Professor at ELTE University, Department of Minority Studies, project manager at TÁRKI, former director of the National Focal Point of the European Union Centre for Monitoring Race and Xenophobia. For ten years he was the chairman of Refuge – Association for Helping Migrants. He served as the president of the Hungarian Sociological Association. Bori Simonovits (simonovits@tarki.hu) (M.A. in Sociology, A.M.A. in Social Policy Analysis) has been a researcher at TÁRKI since 2001, and has research experience in the fields of discrimination, xenophobia, international migration and prejudice. Additionally, she has taught at the University of ELTE (Budapest) in the field of discrimination, migration and methodology. Currently, she is a Ph.D. candidate at the University of ELTE, working on her dissertation in the field of empirical research of discrimination. Niko Toš (niko.tos@fdv.uni.lj.si), is Professor of Sociology, University of Ljubljana, former Dean and Director of the Institute of Social Sciences in the Faculty of Social Sciences; he was the founder (1966) and continues to be the Head of the Public Opinion and Mass Communication Research Centre, and he acts as the Slovenian representative in cross-national, empirical projects (ISSP, WVS, EVS, CSES, ESS). He was President of the Slovenian (1965– 1973) and Yugoslav Sociological Association (1981-1985). His specializations include: the methodology of social sciences, social stratification, international migrations, democracy, voting behavior and electoral systems, and values. Samo Uhan (samo.uhan@guest.arnes.si) is an Assistant Professor of social science research methods at the Faculty of Social Sciences, University of Ljubljana. As a researcher is currently involved in a Slovenian general social survey project “Slovene Public Opinion”. He received his Ph.D. in sociology from the University of Ljubljana in 1993. His main research interests are cross national research, religious changes in transition countries, living and working conditions in candidate countries and EU integration process. He is engaged in ISSP and ESS projects. He has graduate and post graduate courses on methods of quantitative research in sociology at the Faculty of Social Sciences. He is also President of Slovenian Sociological Association. Klára Vlachová (klara.plecita@soc.cas.cz) is head of the research team Value Orientations in Society at the Institute of Sociology of the Academy of Sciences of the Czech Republic. Her research deals with the sociology of politics, particularly political values, ideas and attitudes, voting behavior, and national identity. From 1999 until 2003, she prepared the annual International Social Survey Programme (ISSP) surveys for the Czech Republic. Since 2001, she has been a European Social Survey National Coordinator. In 2002 and 2004 she prepared the ESS Round 1 and ESS Round 2 surveys. Claire Wallace (claire.wallace@abdn.ac.uk) is Professor of Sociology at the University of Aberdeen and President of the European Sociological Association. She was project co-ordinator for the Fifth Framework Programme project “Households, Work and Flexibility” from which this paper originated. She is currently co-ordinator of the Sixth Framework Programme project “Workcare. Social quality in work, care and welfare in Europe”. Marcin W. Zieliński (marcin.zielinski@uw.edu.pl), sociologist, works at the Center for Sociological Research at the Institute for Social Studies, an independent research unit of the University of Warsaw, which is also affiliated with the Institute for Social Research (ISR, University of Michigan, Ann Arbor). He is a member of a team that conducts the Polish General Social Survey (PGSS) and the Polish edition of the International Social Survey Programme (ISSP). He also acts as Deputy Director of the Polish Social Data Archive. His main research activity is concentrated on methodology, especially interviewer effect, item and unit non-responses. index 533 Index Abraham de Swaan 468 Czech Republic 13, 16, 38pp, 48pp, Allport, Gordon W. 463 62, 71, 80pp, 89pp, 105, 131pp, Almond, Gabriel A. 223, 262, 271 139pp, 148pp, 164pp, 175pp, atomic survey-dataset 425 186, 190, 206p, 228pp, 244pp, Atypical employment 37 259pp, 276pp, 283pp, 525pp Austria 16, 63, 67, 80pp, 131pp, 204, Dalton, Russell J. 272 248pp, 266, 276pp Denmark 61pp, 80p, 87, 90, 93, 96, 204, average happiness 135 208, 226, 248pp, 266, 275pp, Background knowledge 21, 422, 461, 281pp 466, 471pp Dennett, Daniel C. 501 Beck, Ulrich 14, 35, 101, 125, 272 Descriptive Statistics 135 Belarus 228, 229 Domaradzka, Anna 14, 99, 525 Belgium 63, 80pp, 204, 226p, 248pp, East Germany 61, 68 276p, 285 Easton, David 223, 262 Bernard, Jessie 133 education 14p, 46pp, 71p, 187, 305, biomedical research 457 307, 312, 319pp, 326, 333pp, Bischof, Christian 11, 13, 55, 525 340, 347pp, 354, 361pp, 368, Bulgaria 13, 38, 41pp, 205, 228, 244pp 379pp, 386, 393pp, 400 Canada 78, 83, 89, 204, 208 Education System 237 categorization 432p EFTA 83, 89 Catholic countries 229, 274, 284 embedded cognition 57, 291, 466, 467, Catterberg, Gabriela 272 473, 477, 489, 490, 491 causal relation 481 empty nest syndrome 101pp, 123pp CEE see Central and Eastern Europe epistemological 462, 473pp, 487pp Central and Eastern Europe 101, Esping-Andersen 67p 105pp, 114pp, 306pp, 320pp, ESS see European Social Survey 334pp, 348pp, 362pp, 380pp, Estonia 62, 246, 252, 255p, 259p, 267 394pp, 412p EU 13, 16p, 40, 50, 82pp, 89, 173p, 201, Central Europe 11, 35p, 52p, 79, 131pp, 208, 230, 243pp, 252, 256, 259, 207, 276 264pp, 528 Chaloupková, Jana 14, 75, 525 Eurobarometer 425 citizenship 17p, 79p, 94p, 197pp, 525 European Social Survey 14p, 18p, 25, cognition 477p, 490p 57p, 63p, 73, 77, 83pp, 97, 129, cognitive neuro-sciences 452p, 457 134, 142, 198, 202, 243p, 253, cohabitation 16, 57, 60, 70, 134pp 260, 263, 266, 275, 286p 425, combined flexibility 40, 41, 42 434, 528 construct validity 433, 437, 439, 441 Eurosclerosis 35 contract flexibility 36, 40pp, 50, 52 evolutionary computation 493 Converse, Peter 464 Finland 61pp, 80pp, 87, 90, 93, 96, 105, corticosteroids 455 131, 204, 208, 226, 247pp, 257p, cumulative flexibility 40, 42 266, 275p, 285pp 534 index flexibility debate 35 Household Chores 153, 154, 156pp, flexibility policies 37 161, 167pp Foerster, Heinz von 452, 476, 478, 482, household labour 57, 58 488, 526 Households, Work and Flexibility 39 folk-psychology 478 Hungary 13pp, 37, 40pp, 62, 80p, 87pp, former USSR 83 105, 131, 134, 137p, 140pp, former Yugoslavia 83 155pp, 175pp, 186, 205, 207, France 16, 63, 67, 78, 80pp, 87, 90, 228, 248p, 252, 255pp, 266, 276p, 93pp, 105, 145, 148pp, 204, 226, 283p, 290 248pp, 266, 276p, 281, 283pp, Iceland 61pp, 226, 227 291 immigrant 14, 35, 78pp, 90pp, 174, Frey, Bruno S. 199 206pp fundamental incompleteness 484, 486 Income flexibility 40, 44 Fuwa, Makiko 57 independent orientation 177 Gallhofer, Irmtraud N. 462, 465, 467, information and communication tech- 470, 501 nologies 462, 468pp GA-system 493pp, 498, 500 Inglehart, Ronald 223, 226, 227, 272 GDP per capita 89 Ireland 16, 63, 67, 78, 80, 82p, 87, 90, GEM 60pp, 73 93, 96, 105, 131, 145p, 150pp, Gender Empowerment Measure 60, 162pp, 204, 226p, 248, 252, 66, 70 257pp, 266, 276p, 284p generative relation 481 ISSP 17p 104, 125, 146p, 153, 156, Germany 61pp, 67p, 73, 78, 80pp, 90, 158p, 163, 175, 187, 190, 198, 93pp, 175, 179, 181, 183, 186, 202pp, 214, 230, 525, 528p 204, 223, 226p, 248, 252, 257pp, Italy 68, 78, 80, 82, 87, 90, 93, 96, 208, 266, 275pp, 281, 283, 291 210, 226, 244, 246, 248p, 252, Germany (East) 62 257, 259, 267, 276, 283pp Germany (West) 63 Kaase, Max 272, 274 GINI 61pp, 70, 73 Korn, Werner 4, 11 GINI-Coefficient 61 Krosnick, Jon A. 466 Glasersfeld, Ernst von 452 Kuechler, Manfred 272 global governance 198, 203, 209 Kurdija, Slavko 18, 221, 266, 526 Great Britain 16, 67, 93, 105, 145, 148, Labour market 35pp, 42p, 45, 50pp, 58, 150pp, 162, 164pp, 204 61, 101, 110, 126, 145p, 166 Greece 14, 63pp, 80pp, 93, 96, 208, Labour Market 210, 244pp, 252, 259, 261, 267, Segmentation 35 276p, 280pp Stratification 45 Habermas, Jürgen 467 Theories 35 Hafner-Fink, Mitja 17, 195, 525 La Piere, Richard T. 463p Hamplová, Dana 15, 129, 525 Latvia 105, 205 HDI 288, 289 Lebeda, Tomáš 19, 269, 526 health research 452, 454, 457 Life Chance 444 Holland, John 493 life-styles 448 Household Budget 159, 160 Lipset, Seymour Martin 223, 262 index 535 living systems 474pp, 488p Pattern loneliness 14p, 99, 101pp preservation 438 Luxembourg 63, 80pp, 86pp, 96, 244, preserving 438 246, 248, 250, 252, 258, 261, 267, Personal contacts 103 276p, 281, 283pp place flexibility 40, 42pp, 49 macro-level 57, 274 Poland 16, 62, 71, 80pp, 90pp, 105, Marsh, Alan 273p 145pp, 162pp, 175pp, 182pp, Maturana, Humberto R. 477 205pp, 228p, 244, 248p, 252, McAdam, Doug 272 255pp, 266, 276p, 281pp, 525p McCarthy, John D. 272 political participation 19, 57, 60, 199pp, McCulloch, Warren 480 213, 271pp, 280pp, 290p measurement process 464p, 470 Portugal 63pp, 80, 82, 86p, 90, 93, 96, micro-level 57, 274 205, 208, 210, 248p, 252, 255pp, Miheljak, Vlado 18, 221, 266, 526 264, 266, 276p, 280pp Milbrath, Lester 271, 273p principle of undifferentiated encodings Morocco 78 452 morphological approach 425pp, 450pp pro-children orientation 176p, 190 aggregation 425, 428, 430pp, 442, Putnam, Robert D. 272 445p, 449pp qualitative social research 463 elementary form 429, 435 Raciborski, Filip 16, 143, 527 formation 425, 428, 430p, 435, 439, Reautschnig, Armin 4, 11, 301, 407, 442, 444, 449pp 423, 375 morphological construct 425, 430, 433, recursive designs 483 443, 454 regimes of regulation 37, 53 Müller, Johannes P. 452 regulated flexibility 13, 37, 38 Müller, Karl H. 1, 3, 12, 20p, 25, 29, Regulation of Flexibility 51 295, 421, 423, 459, 526 reliability 432, 437, 439pp Netherlands 13, 36p, 41, 45, 47pp, 62, Remote contacts 103 78, 80, 82, 87, 93pp, 204, 226p, Research Design 58 248p, 257p, 266, 276p, 284p retention rate 437pp Net Reproduction Rates 173 Risk 444 new immigrant countries 78 Romania 13, 38, 40p, 44p, 48pp, 228p Newton, Kenneth 272 Russia 105, 205, 228p, 243pp, 255p, non-trivial 473, 487p, 499 259, 263p, 267 Norris, Pippa 272, 274 Šalamounová, Petra 14, 75, 527 Norway 61pp, 80p, 87, 90, 93, 96, 105, Saris, Willem E. 462, 464p, 467, 470, 175, 177, 179, 182, 183, 186, 190, 484, 501 204, 208, 226, 248p, 257p, 266, Satisfaction with Democracy 254 275p, 281, 284p Satisfaction with Life 255 OECD 77, 79p, 84, 89, 92 scaling 444, 449 old immigrant countries 78 Scandinavian countries 17pp, 57, 61, OLS Regressions 137p, 140p 67, 80, 95, 97, 179, 182p, 208p, Orthodox countries 229 210, 248p, 255p, 258, 262p, 275p, over-learned 21, 490pp, 496, 498p, 503p 281, 291 536 index Scheler, Max 104, 126 Trust in Institutions 18, 221, 228, 231, Schneider, William 223, 262 237, 239p, 243, 245, 248, 253, Science I 21, 461, 470pp, 480pp, 503p 257p, 260, 266, 347pp, 379pp, Science II 21, 461, 470pp, 480pp, 503p 411pp secularization process 229 Turkey 63pp, 78 self-inclusive 480 Turner, Bryan S. 198p self-referential 480, 482 Uhan, Samo 17, 195, 528 Shore, Cris 197, 201 UK see United Kingdom Sik, Endre 13, 33, 527 Ukraine 62, 228p, 243pp, 250pp, 263pp Simonovits, Bori 13, 33, 527 under-learned 21, 490pp, 500pp situated cognition 477 UNDP 60, 63, 73 Slovakia 16, 62, 145, 150pp, 160pp, 205, United Kingdom 13, 37, 40pp, 50pp, 228, 230, 244pp, 252pp, 261, 267 62, 78, 80p, 94pp, 155, 226, 248, Slovenia 4, 13, 18, 38, 41pp, 62, 77, 105, 252pp, 261, 266, 280pp 131pp, 175pp, 204pp, 221pp, United Nations Development Program 230pp, 240pp, 252pp, 260pp, 60, 288 276p, 284, 290 United States of America 35, 67 78, 83, Smith, Tom W. 36, 198, 272 89, 134, 204, 223, 291 Social capital 434 USA see United States of America Social Networks 15, 101, 105, 109, 125 validity 432p, 437, 439, 441 Spain 16, 63pp, 78pp, 90pp, 105, 145pp, Verba, Sidney 223, 262, 271 151pp, 204pp, 226p, 248, 252, Vlachová, Klára 19, 269, 528 257p, 261, 266, 276p, 280, 283, Wallace, Claire 13, 528p 284pp Warren McCulloch 480 stressors 454pp WE see Western Europe Swaan, Abraham de 468 Weber, Max 467 Sweden 13p, 36pp, 40pp, 61pp, 78pp, Western Europe 13, 33pp, 52, 53, 78, 90pp, 175, 179, 182p, 186, 204, 80pp, 97, 101pp, 112pp, 120pp, 208, 226, 249pp, 257p, 266, 275p, 197, 227pp, 251, 276, 291 280pp Western Germany 85 Switzerland 14, 63, 80pp, 86p, 90pp, workforce 13, 35p, 42, 51, 80, 145 105, 204, 210, 248p, 255pp, World Value Survey 425 261pp, 276p, 281pp Zaid, Mayer N. 272 test-design 430, 433, 442, 444 Zaller, John R. 497, 500 Time Budget 158 Zieliński, Marcin W. 17, 171, 529 time flexibility 36, 43pp Time flexibility 40, 44 Topf, Richard 272 Toš, Niko 1, 3, 12, 18, 25, 29, 230, 266, 295, 527p Trade Unions 235 trans-national democracy 198 triangulation 463 trivial research designs 487p WISDOM-Schriftenreihe in der edition echoraum Programm 2007 Band 1 Karl H. Müller : The New Science of Cybernetics. The Evolution of Living Research Designs. Vol. I Methodology Release : Mai 2008 ISBN 978-3-901941-10-8 Band 2 Niko Toš | Karl H. Müller [eds.] : Political Faces of Slovenia Political Orientations and Values at the End of the Century WISDOM | edition echoraum, Vienna, 2005 [ISBN-10 3-901941-11-8] ISBN-13 978-3-901941-11-5 Band 3 Albert Müller | Karl H. Müller [eds.] : An Unfinished Revolution? Heinz von Foerster and the Biological Computer Laboratory (BCL), 1958–76 WISDOM | edition echoraum, Vienna, 2007 ISBN 978-3-901941-12-2 Band 4 Richard Jung : Experience and Action. Selected Items in Systems Theory. WISDOM | edition echoraum, Vienna, 2007 ISBN 978-3-901941-13-9 Band 5 Ranulph Glanville | Alexander Riegler [eds.] : The Importance of Being Ernst. Festschrift for Ernst von Glasersfeld. WISDOM | edition echoraum, Vienna, 2007 ISBN 978-3-901941-14-6 Band 6 Ranulph Glanville | Karl H. Müller [eds.] : Gordon Pask, Philosopher Mechanic An Introduction to the Cybernetician's Cybernetician WISDOM | edition echoraum, Vienna, 2007 ISBN 978-3-901941-15-3 www.wisdom.at Zu beziehen bei / www.echoraum.at To be ordered from : edition echoraum Gerhard Grössing Gerhard Grössing Ernst von Glasersfeld et al. [Hg.] et al. [Hg.] Radikaler Heinz von Foerster 90 Heinz von Foerster Konstruktuvismus Aus Anlaß des The Stanford Lectures Herausgegeben von 90. Geburtstags von According to a dictum Albert Müller und Heinz von Foerster am of Heinz von Foerster Karl H. Müller 13. November 01 ist in telling us that his life Ein Vortrag von einer Gemeinschafts-and work were Ernst von Glasersfeld, produktion von Heinz inseparable, this gehalten am 10. Mai von Foerster-Gesell- booklet combines 2005 an der schaft und echoraum Heinz von Foerster’s Universität Wien ein Buch mit CD-ROM ‘Stanford lectures als Text und erschienen, das in on constructivist Tondokument, ergänzt einer Serie von bislang epistemology’ from durch ein Gespräch unveröffentlichten 1983 and their original von Albert Müller Materialien und manuscripts an und Karl H. Müller mit Fotografien Heinz von CD-ROM with a series Ernst von Glasersfeld Foersters eindrucks- of photographs über biografische volle Synthese von documenting his life. und werkspezifische »Leben und Werk« Aspeke. dokumentiert. © 2005, 96 p, 65 Abb. teilw. 4c. incl. Buch mit Audio-CD. © 2001, 96 Seiten, CD-ROM (MAC/PC/MP3) © 2005, 96 Seiten, 65 Abb., teilw. 4c. incl. Euro 15,– 5 s/w-Photos. Euro 15,– CD-ROM (MAC/PC/MP3) ISBN 3-901941-28-2 ISBN 3-901941-29-0 Euro 15,– ISBN 3-901941-08-8 [Engl. Ausgabe von »Heinz von Foerster 90«] Zu beziehen bei / To be ordered from : www.echoraum.at Document Outline Blank Page