E - D O K U M E N T I S J M





Social Research: from Paradigmatic Divide to Pragmatic Eclecticism

Mitja Hafner Fink | Slavko Kurdija | Samo Uhan

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism

Mitja Hafner Fink | Slavko Kurdija | Samo Uhan E-DOKUMENTI SJM/Ljubljana, 2022

edition echoraum | Wien



Mitja Hafner Fink/Slavko Kurdija/Samo Uhan

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism 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 12

Editor: Slavko Kurdija

Peer-Review: Dr. Tina Kogovšek/Dr. Anton Kramberger/

Dr. Dinka Marinović Jerolimov

Text edition: Murray Bales, Mark Gough, Gertrud Hafner Layout: Werner Korn

Figures and Graphs: Armin Reautschnig

This work is licensed under a Creative Commons AttributionShareAlike 4.0 International License.

URL: https://knjigarna.fdv.si/ in www.cjm.si DOI: 10.51936/9789612950187

Kataložni zapis o publikaciji (CIP) pripravili v Narodni in univerzitetni knjižnici v Ljubljani COBISS.SI-ID 102545667

ISBN 978-961-295-018-7 (PDF)

Printed edition was published by Edition Echoraum, Wien, 2017

ISBN: 978-3-901941-51-1

Table of Content

Introdocution 7

Part I: Some Dilemmas of the Interpretative Approach 13

1 Grounded Theory and Inductive Ethnography: A Sensible Merging or a Failed Encounter?

17

1.1 The Theory and Practice of GT

21

1.2 Inductive Ethnography

26

1.3 Can the Position of Theory and Method in Ethnography be Strengthened by GT?

28

1.4 Can this Improve Ethnography?

30

1.5 Conclusion

31

2 Researching Communication: The Interpretive Approach Between Theory and Practice

33

2.1 Theory and Methodological Consequences

38

2.2 De(con)struction as the Methodological Means 40

2.3 Starting Points for a Different Methodology of Communication

44

2.4 How to Formulate Research Strategies?

49

Part II: Cross-National Comparative Research Between Theory and Application

53

3 Deductive and Inductive Logic in Developing Typologies in Cross-National Comparisons: The Case of Political Participation

57

3.1 The Problem

59

3.2 Political Participation Typologies in the Context of Cross-National Comparison

62

3.3 Summary: The Complexity of the Classification Process in Cross-National Comparative Research

75



Appendix 1

79

Appendix 2

81

6

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism 4 Selected Problems of International Comparative Social Survey Research: The Case of the ISSP

85

4.1 ISSP as a Global Social Infrastructural Programme 88

4.2 Some Aspects of Equivalence in International Social Science Surveys – the Case of ISSP

90

4.3 International Comparative Social Survey Research and the Problem of Combining Different Levels: From the Unit of Data to the Unit of Comparison

101

4.4 Some “Institutional” Aspects of International Social Survey Programmes, or why ISSP has Continued for over three Decades

105

Part III: The Quantitative Approach and some Problems of Social Surveys

107

5

Context Effects in Social Surveys: Between Instrument and Respondent

111

5.1 The Problem

114

5.2 The Research Model

117

5.3 Analysis and Results

118

5.4 Discussion

124

6 Development of a Telephone Survey within the Public Opinion and Mass Communication Research Centre – Comparative Aspects of the Method

127

6.1 Introduction

129

6.2 History

129

6.3 Technique

131

6.4 The Practice and Characteristics of Telephone Surveying within the Public Opinion and Mass Communication Research Centre (POMCRC) Research Programme

133

6.5 Quality

136

6.6 Comparison of the Approaches

141

6.7 Conclusion

143

References 149

Index

159



Introdocution

In the book, we link certain aspects of social research stemming from conflicting paradigmatic standpoints: on one hand, we consider issues of survey research as a typical representative of the quantitative approach based on positivism and, on the other, the problem of grounded theory as a paradigmatic reference of the qualitative approach. Today, when various approaches to the integration (combination) of methods from different paradigmatic positions are increasingly being applied in social research, this should not be seen as unusual. However, if we add that all of the book’s authors are primarily ‘involved’ in, at least at first glance, what is paradigmatically the most typical quantitative research, namely, in social surveys, this slightly changes the assessment. We proceed from our own experience in researching public opinion and from international comparative social surveys conducted by the Public Opinion and Mass Communication Research Centre (POMCRC) at the Faculty of Social Sciences of the University of Ljubljana.

We believe that for any successful social survey placed in the positivist tradition, certain principles of the interpretive approach and constructivism should also be taken considered. The survey situation is essentially a ‘social event’, mainly based on the communication between the researcher and the respondent. This ‘social event’ (the whole context of the survey interview) no doubt has an important role in how respondents form answers to survey questions. Therefore, the image of society we present with the data collected in the survey is partly also a result of this ‘social event’. The outcome of exploring people’s attitudes and opinions with a survey is therefore not merely a ‘neutral’ image of the actual situation in society, but we can say it is also a ‘construct’ of this situation ( cf. Krippendorff, 2005). If we consider the thesis on the social construction of reality (see Berger & Luckmann, 1988), then we can say that the phenomenon we reveal through opinion polls would not have existed without this survey. By doing so, we have certainly moved to an area which, at the time of the paradigmatic split, almost exclusively fits on the qualitative side and within the framework of constructivism.

If, accordingly, quantitative researchers pursue the quality of survey data as their key (methodological) goal and accept the thesis of constructing social reality, then, of course, they cannot avoid questions they are unable to answer simply by ‘juggling’ numbers or statistical calculations. They should also take the principles of the ‘opposite’ side into account, namely the interpretative paradigm and constructivism, and specifically a qualitative research strategy based on them. Thus, in developing and testing the survey questionnaire – the measurement instrument of the most typical quantitative social science research –

researchers often use cognitive interviews, which are a typical qualitative method ( cf. Willis, 2015). When accepting the thesis on the survey situation as a factor in the construction of social reality, it is also not surprising that researchers are dealing with the problem of the context of the survey situation, which includes

10

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism the questionnaire’s characteristics, the survey’s mode, the specific circumstances of the survey situation, along with the broader social and cultural circumstances of conducting the survey ( e.g. Schuman & Presser, 1981; Tourangeau & Rasinski, 1988; Schwartz & Sudman, 1992; Krosnick, 1992; Uhan, 1998, Johnson et al., 1999; de Leeuw, 2005).

As mentioned at the start, one trend in contemporary social science research is to combine qualitative and quantitative approaches (methods). This is also a reflection of the complexity of the research problems that social scientists are addressing. Let us consider just a few aspects of such complexity: a) multi-level research (micro

– macro); b) research problems often require an interdisciplinary approach; and c) applied research often involves other actors in addition to researchers. We can say that, even during a period of the greatest paradigmatic divisions, for purely pragmatic reasons researchers in practice often ‘forget’ about these divisions and use qualitative and quantitative methods within the same research. In this context, international comparative research is especially interesting, where the

‘qualitative–quantitative’ divide was particularly present, and at the same time, due to the complexity of comparative research problems the need to overcome this division has emerged ( cf. Ragin 1987). International comparative research is a typical example of combining different research levels – especially when considering the following: data are often collected at the micro level (for example, through a social survey of individuals), while research conclusions are presented at the macro level ( e.g. at the level of nations or countries). The complexity of cross-national comparisons (particularly when comparing complex macro-social units) is thus a key argument for combining methods or even synthesising qualitative and quantitative approaches in the comparative research context ( cf. Ragin 1987; Rihoux & Ragin, 2009).

Of course, it should be emphasised that the practice of combining approaches is much more demanding since researchers are expected to possess thorough knowledge about the principles (and specific methods in use) of both qualitative and quantitative key research approaches. Moreover, in the case of comparative research, understanding the importance of combining the two approaches is also expected. Therefore, this book’s structure is adapted to reflect this fact – the book chapters are grouped in three parts: the first part emphasises selected issues of the qualitative approach, the second part deals with questions of comparative research, while the third part examines certain aspects of quantitative research.

In the first part of the book, we deal with selected aspects of the qualitative approach. In Chapter 1, we examine the potential of the research complementarity of grounded theory (GT) and inductive ethnography in the context of merging and combining research approaches. In the second chapter, we discuss some aspects of poststructuralist theory relevant to methodological strategies in communication

Introdocution

11

studies. This discussion is based on Derrida’s concept of deconstruction and its possible methodological consequences. Here we emphasise its importance for research communication – a communication between researcher and respondent or research participant in general.

In the second part, we move on to problems of comparative approach, more precisely, problems of international social surveys. Moreover, by locating the part on the comparative approach in the middle between the qualitative and quantitative parts, we wish to stress how qualitative and quantitative approaches

‘meet’ in the comparative research framework. The discussion in Chapter 3

begins with this assumption, namely that comparative research is an approach that combines elements of quantitative and qualitative approaches. Using an empirical case of typology (classification) of political participation, we show how deductive and inductive reasoning are combined in comparative analysis based on data from international social surveys. In Chapter 4, we consider the question of whether and how the data collected in different socio-cultural contexts (usually countries) are comparable. We describe the ways in which researchers within the International Social Survey Programme (ISSP) are trying to improve the quality of data intended for cross-national comparisons.

In the third part of the book, we deal with specific methodological problems of social surveys as a paradigmatic quantitative research design. The book therefore concludes with research problems that constitute the focus of the activity of the research centre where the book’s authors work – that is, problems of the social survey. In the fifth chapter, we present an empirical test of context effects in social surveys and, as part of this framework, discuss the nature of the cognitive representations respondents create and utilise when answering survey questions.

In the final chapter, we address the problem of survey mode – more precisely, we deal with problems of telephone surveys. By describing the research practice of the POMCRC, we reflect on various aspects and dilemmas of telephone surveying.

The book’s ambition is not to present the entire range of the methodological and epistemological diversity on the spectrum from the positivist to the interpretative (constructivist) paradigm. Instead, it attempts to expose some of the characteristic points on this spectrum that have been significantly marked by the authors’ research institution – the POMCRC. Another central idea foundation on which the book is based is the belief that for any successful survey research placed in the positivist tradition some principles of the interpretative approach and constructivism must also be taken into account. In this sense, we believe the book can contribute to contemporary debates that are leading to the paradigmatic divisions in the social sciences being overcome.



Part I

Some Dilemmas of the Interpretative Approach

In the last decade, social research methodology has been experiencing rapid development. This is shown in both empirical research, with findings of the cognitive sciences gaining ground, as well as in the epistemology of the social sciences. In this process, an important place is increasingly being taken by qualitative research methods that were also foregrounded in fields which a couple of decades ago were exclusively reserved for quantitative research approaches. In recent years, attempts have been made to merge and combine research approaches as part of triangulation, along with more ambitious attempts to create so-called reflexive methodology.1 Researchers claim it is necessary to combine and complement research approaches because any research method based solely on measurable (quantified) data can be disputed by a more complex research design that includes teamwork, the participation of lay actors, and by practising the so-called double hermeneutics or the flow between theoretical and practical knowledge (Adam et al. 2012).

Based on this premise, this part will address two topics: grounded theory (hereinafter GT) together with inductive ethnography in the context of the merging these two research techniques (methods) and poststructuralist approaches, with each employing its own specific way of solving researchers’

quandaries when using qualitative methods in their work. This relates to the fact that the contemporary social sciences operate on one hand with open empirical material that lacks any clear structure and, on the other, with qualitative contents which are more suited to rigid categorisation. Accordingly, either uninterpreted facts in the form of “pure” data or the approach whereby even an everyday human experience is denoted as a “discursive construct” are offered as two extremes on a spectrum of answers to the question: What is the essence of research in the social sciences? Both approaches can be situated between the two methodological poles by taking account of the dangers all approaches face when leaving a conventional and safe environment.

1 The syntagm reflexive methodology is used by Alvesson and Sköldberg (2000) to denote the complex relationship between knowledge-development processes and variable contexts in which knowledge is developed, including all actors. Unlike “classical” methodology, reflexive methodology fully accepts the fact that language, culture, social structure, norms, ideology, discourses etc. form a constitutive part of the scientific process.



Grounded Theory and Inductive

Ethnography: A Sensible Merging or a

1 Failed Encounter?

The research focus of this chapter1 is to examine the potential of the research complementarity of GT and inductive ethnography in the context of the abovementioned merging and combining of research approaches and in light of the “rebirth of qualitative research” (see Adam et al. 2012).

Although both approaches are qualitative methods, at first sight the possibilities of their fruitful co-operation seem scarce. In particular, GT is considered an integrated research/scientific and cognitive approach rather than simply a type of analysis or a technique of data collection (Kavčič in Adam et.al. 2012).

As the title of their work ‘The Discovery of Grounded Theory: Strategies of Qualitative Research’ (1967) reveals, the basic intention of Barney G. Glaser and Anselm L. Strauss is to bridge this gap between big theories and empirical research. The authors claim that the principal aim in research is to discover theories rather than just verifying them as is the practice in quantitative research. Placing the generation of theories in the centre of research, they only see verification as part of the theory-generation process.

Glaser and Strauss proceed from their belief that theory has become a synonym for big theories, with the task of researchers being reduced to their more or less creative verification. Alternatively, the authors suggest a much less rigorous approach that enables virtually anyone to create their own theory, provided it is based on everyday experience of real life. While, of course, theories generated in this way also need to be tested, this mainly happens in terms of their modification rather than destruction or, in other words, a theory can only be replaced by another theory.

Glaser and Strauss insist on the inductive development of a theory against an empirical background. They claim this is the only way of preventing the separation of theory and reality which is a characteristic and also the biggest disadvantage of an approach based on logical deduction.

From its beginnings in the mid-1960s grounded theory has undergone deep changes. If initially it represented a positivist model within a qualitative approach related to symbolic interactionism or its sensibility towards the world, today it is primarily considered a useful and flexible analytical technique.2 Undoubtedly, 1 An earlier version of this chapter was published in journal Teorija in praksa, May-Aug. 2013, Vol. 50, No. 3/4, p. 642–657, 689.w

2 GT is rooted in the symbolic interactionism embodied in Strauss, as well as in statistical positivism which is part of Glaser’s “intellectual baggage”. Researchers who use GT draw their inspiration from the very concepts which undoubtedly belong to intellectual heritage of SI, with its foregrounded pragmaticism, idiographic research, qualitative methods, exploration, sensitising concepts, cognitive symbols, social action, empirical orientation and successive induction from empirical material. While GT is clearly not generated automatically as a sum of these concepts, it undoubtedly uses them in a way that provides the method with its

20

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism its beginnings are linked to ethnography much more than its modern use. Due to the complicated procedures in GT, numerous researchers more often tend to use in-depth interviews, focus groups and other ethnographic techniques, which of course raises questions about the justifiability of using GT techniques for ethnographic research. However, in light of the requirements to combine research approaches, the question arises of whether “self-limitation” in the choice of ethnographic research techniques makes any sense. In this context, we see an opportunity for the complementary use of both research approaches, with the assumption that the combining of approaches can help us reach beyond the contradictions and tensions which, according to Charmaz, emerge from the attempts to methodologically incorporate ethnography in GT.

In our view, any attempt at combining methods needs to answer the question about the “hermeneutic” justification of the attempt. Therefore, mainly the following questions seem unresolved in the relationship between GT and inductive ethnographic approaches:

1. From the viewpoint of GT, the question that remains unresolved is: what constitutes the basis of ethnography? While data obtained by researchers through in-depth interviews, focus groups or the analysis of archive material are considered to give at least ready-made answers to research questions, ethnographic field notes made on the basis of direct observation do not even meet this criterion.

2. The next question relates to the importance of theory in ethnographic research. Indeed, GT users have also never unanimously agreed on how it is that an inductively generated theory, which adapts to data, can concurrently contribute to the creation of the corpus of sociological theories. This applies to ethnographic research to the same or an even larger extent. In this view, realising the importance of theory is not enough.

3. The final simple question is: can GT improve ethnography? Here, Glaser and Strauss are entirely pragmatic – those methodological approaches that can

“serve” researchers are good quality. This paper also challenges the sufficiency of this pragmatic criterion.

Later in the chapter, the “tensions” identified by Charmaz in the relationship between GT and ethnography, and reformulated in the above research questions, will serve as the basis for formulating starting points for the complementary use of techniques from both qualitative approaches. First, we will look for answers by reflexively examining the properties of both approaches’ research techniques which in the concluding part of the paper will end by identifying empirical grounds (Alvesson and Sköldberg 2000:12).

Grounded Theory and Inductive Ethnography 21

the methodological context enabling the coherent use of the two research approaches.

1.1 The Theory and Practice of GT

Considering the emphasis it lays on the local (authentic), and its reservation towards big theories, in one part GT comes close to postmodernism while at the same time keeping a distance from postmodernism mainly due to GT’s pronounced rejection of postmodernism’s typical intertextuality. Regarding the role of empirical data, GT is in line with the principles developed by positivists.

Both approaches develop the ambition of separating theory from empirical data, along with a subsequent testing of their relatedness, even though for GT the role of verification only comes second. Apart from the status of verification, positivists match with GT in more ways than would be expected considering their formal conceptual starting points. GT and positivists virtually fully match in their requirements for generalisability, reproducibility and predictability (Strauss and Corbin 1990:29). Of course, these similarities do not eliminate the essential divergence of both approaches in their relation to theory. While symbolic meanings of interactions which generate theory are crucial for GT, positivists remain at the level of verification based on quantitative data (Alvesson and Sköldberg 2000).

1.1.1 Substantive and Formal Grounded Theory Glaser and Strauss define two forms of GT – substantive and formal – which differ in their respective fields for which theories are developed. Substantive theories explain particular aspects of social life such as questions related to the emergence of youth delinquency, teenage pregnancy and similar, while formal theories help researchers reach for higher levels that provide GT with theoretical abstraction (Marvasti 2004). A typical example of the use of formal GT is an analysis of social inequality requiring the simultaneous consideration of different factors.

Although recommending that formal theory be generated on the basis of substantive theory, the authors suggest that both types of theory can be developed inductively from data.

However, while crucial in the authors’ view (for example, they use this procedure to ground the syntagm of middle-range theories), this approach to generating theory has its downsides. Although Glaser and Strauss strive to clearly distinguish the two types of theories, they are only partly successful. It seems that in defining differences between both types, the authors presume that social entities are

22

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism empirical while interrelationships and properties are conceptual. They claim that in some cases the difference only exists at the “level of generality”. However, it is this claim that is most often the target of criticism given that differentiation between the levels of generality cannot equal the differentiation between the empirical and the conceptual. According to Alvesson (Alvesson in Sköldberg 2000), these two authors establish an absolute difference (although with fluid borders) between two phenomena which only differ in relative terms (in terms of a higher or lower level of generality). In their view, the choice of the level of generality only depends on the researchers’ aims.

As an example of such inconsistent use, different authors usually refer to Glaser-Strauss’ fundamental work Awareness of Dying (Glaser and Strauss 1965), an ethnographic study employed by the authors to develop a substantive theory of social loss (the case of providing care for a dying patient). In keeping with GT’s methodological starting points and taking researchers’ intentions into account the study of social loss could be expanded to the study of social values which is completely independent of the studied case and belongs to “formal” theory. It follows from this case that the substantial/formal dichotomy can be simplistically related to the two extreme positions on the scale of generality. In this case, a lower level of generality would mean focusing on the problem of a family or professional loss, while a higher level would require an expansion of study to the problem of general social values. This mainly suggests that Glaser and Strauss’

definition of the border between substantive and formal theories is completely arbitrary, and that they overlook the possibilities offered by the research process since the development of a theory, rather than necessarily ending at a certain level of abstraction or generality, has to continue to enable further research of properties rather than entities (Alvesson in Sköldberg 2000:32).

According to Alvesson and Sköldberg, the solution to the epistemological quandary leading GT to a dead-end can be efficiently solved by taking account of Bourdieu’s critique which is mainly targeted at positivists and Lazersfeld’s statistical school, although the basic thrust of the criticism can also apply to GT. On one side, Bourdieu (Bourdieu 2002) disputes the short-sightedness of an empirical approach which is unable to penetrate the deeper structures of social relations, and is thus limited to the “surface” and, on the other, the equally dangerous social sciences research reduced to common-sense concepts that is a common practice in qualitative research. That is why researchers in the social sciences should strive for a radical epistemological rupture centred around the demand to create social concepts (constructs of social objects) which are nether based on the referential framework of everyday worlds nor on superficial empiricism.

Grounded Theory and Inductive Ethnography 23

In terms of their interrelationship, grounded theory (GT) and Bourdieu’s constructionism represent extreme positions on a spectrum of possible methodological approaches, with their core resting on their attitude to the research object. In this view, the GT approach comes closer to the position of actors who are studied, while an “epistemological gap” exists between Bourdieu’s construction of a social object and its empirical background. Bourdieu’s construct is a representative of “big theories” (criticised by GT) with no empirical background which enables GT to “float on the surface, even though without revealing deeper structures”. While in this context Bourdieu’s constructionism and GT seem distant from each other, even incompatible, a distanced view of both approaches discloses their substantial research potentials which come to the fore when both methods are used parallel to each other and when (if ) they are related to concrete research aims. Alvesson proposes a somewhat heretic combination of both approaches, suggesting that the epistemological rupture would be achievable by considering Bourdieu’s “social object” to be GT’s “formal” theory which reaches beyond the mere empirical basis of an individual study. In this case, formal theory would represent the “deeper structure”, while substantive theory would represent the “superficial structure on which it leans” (Alvesson and Sköldberg 2000:34).

Therefore, introducing this rule would eliminate this basic deficiency of GT and at the same time reinforce its advantages – i.e. primarily researchers’ competencies to develop new theories with the emphasis on their creation rather than simply their verification. GT’s significance lies in its ambition to liberate methodology from rigid quantitative schemes.

1.1.2 Data

According to critics of the inconsistency of the methodological approach Glaser and Strauss developed within GT, what most conspicuously stands out is the question of what is considered data within GT. Glaser and Strauss use the term

“incident” (Alvesson and Sköldberg 2000) without, however, offering a precise definition of it. Later in his work Qualitative Analysis for Social Scientists (1987) Strauss uses the term “event” as a synonym for “incident”. Referring to the paradigm of symbolic interactionism, the term can be placed in the context of social interaction. Of course, such a placement does not provide all the solutions since some data do not correspond to the definition of event and, while some events are not incidents, not all incidents are social interactions.

24

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism 1.1.3 Categories

While in GT data represent the first research prerequisite, categories are the key link to data. Unfortunately, also when it comes to defining categories Glaser and Strauss are not very accurate and they do not give an unequivocal definition of category. The only uncontroversial fact is that the data-coding process includes a categorisation of data. Glaser and Strauss speak of two ways of developing categories: most commonly, categories are developed by conducting interviews with the actors who are being researched; categories can also be developed on the basis of an individual “incident” which in the research process is joined by other incidents that ultimately develop and create a category. In this case, researchers simply:



− read texts (field notes, interviews, documentary material);



− try to identify categories to which the data belong (especially in the case of common-sense concepts); and



− write memos on categorised data (Alvesson and Sköldberg 2000).

Strauss also proposes a special procedure which should consider the circumstances, the interaction between the actors, strategies and tactics, along with consequences.

It follows from what has been said that techniques in GT’s research process can vary, with the coding process being its only constant. Coding is a constant comparative analysis of new data which are being categorised with previous data in the same categories so as to describe the properties of individual categories.

The coding continues to finally end in theoretical saturation which is reached when new data no longer bring anything new to a category. During this process researchers mainly face the problem of how “to represent reality in an unambiguous way”. The question here is of researchers interpreting what they are seeing in light of their own unreflected frames of reference (Alvesson and Sköldberg 2000:27). The problem lies in the pre-scientific categories grounded in common-sense thinking, which is what makes the category development process extremely important.

In practice, researchers tend to choose one of three ways of developing categories.

The first and most common way is taking field notes and writing memos on how categories’ properties are interrelated. As a rule, memos contain diagrams, matrices, tables and figures which help the discursive context to emerge for the researcher. The other way is identifying a core category or a central concept which permeates all others and represents the key to generating a theory. A core category is developed as follows (Strauss 1987):



− it has to be central and relate to other categories;



− it has to occur frequently;

Grounded Theory and Inductive Ethnography 25



− it has to be linked to other categories in a simple way;



− it has to have implications for a formal category; and



− it has to develop a new theory.

The third way is to make diagrams or models which illustrate how individual categories are interrelated on the basis of their properties. The aim of this process is to achieve so-called “conceptual density” or saturation. This way of coding is called focused coding by Charmaz who sees it as the possibility of expanding concepts’

level of abstraction and increasing their applicability, i.e. they become more theoretical and apply to a broader range of observations (Bryant and Charmaz 2002:686). By carrying out focused coding, researchers are able to reduce “the universe of meanings” to a manageable number of categories of meaning. In this sense, Charmaz is in favour of “action codes” which direct the researcher’s attention towards a continuing process of social interaction.

However, according to many authors the saturation and integration process comes at a “high price”. Closeness to or similarity with an actor’s views can cause a so-called over-formulation when at best we use different words to describe what is already (implicitly or explicitly) known ( e.g. a mastectomy as an invisible handicap).

1.1.4 Theoretical

Sampling

Considering the empirical nature of GT and its foregrounding of the importance of data in theory generation, the question of sampling proves to be relevant.

Here, it needs to be pointed out that sampling in GT is inadequately compared to quantitative probability sampling. In Strauss’ view, sampling as the central principle in the “data-category-theory” triad is “theoretical” (1987:38–39) and he describes it as: “a means whereby the analyst decides on analytic grounds what data to collect next and where to find them” (1987:38–39). Researchers deal with the question of which groups, sub-groups, events and activities come before others in the data collection process, “and for what theoretical purpose”

( ibid.). Therefore, the entire context of data collection relates to the context of the emerging theory.

GT is unique in that any groups at all can in principle be compared, while with classical comparative methods groups that are too different from one another are excluded from the comparison. This can be understood as an advantage if the comparison is sensible or grounded and not only depends on the principle of the distance between entities in some abstract conceptual space (Alvesson and Sköldberg 2000:27).

26

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism Theoretical sampling includes two steps. In the first step researchers minimise the differences between groups and, in the second, they maximise them. Sampling concurrently involves the process of a theory’s emergence. The purpose of the first step is to look for the basic categories and their properties. The procedure usually begins with an individual case, with a question which does not reach deep but is wide and non-specific. In the second step (maximising the differences between comparison groups) category properties are researched and linked together to make sensible contents (a theory) Alvesson and Sköldberg 2000:28).

In terms of technique, the first and second steps are about a constant comparison of data in order to generate and develop categories and their properties, which brings good results, while its disadvantages are revealed when events are arbitrarily ascribed to categories whereby organic relations between the incidents are broken ( ibid.).

1.2 Inductive Ethnography

Like GT, ethnography is also considered a “data-oriented” qualitative method which, however, is the point where its similarities with GT nearly end. Ethnographic research is concerned with the interpretation of society’s cosmogony in a way which takes account of actors’ participation along with their interpretation of the world they live in (Zoe Bray 2008:301). Through descriptive generalisation and development of explanatory interpretations of the “social world” researchers try to identify variability and common traits of societies in the studied period.

By allowing for the intrinsic interrelatedness of objective observation and actors’

subjective interpretation researchers explain (make sense of ) the process of the making of meaning that the actors confer on a social object and their own participation on the basis of their own beliefs and social conventions.

The basic idea of the ethnographic approach is contained in researchers’ tendency to understand social action in a specific environment from the perspective of another culture or the experience of the other. In these terms, Silverman (1993) reaches farthest by regarding as ethnographic every research technique which includes an observation of events, incidents, and participation in the natural context, based on the premise of the interdependence of theory and data. The key maxim of the ethnographic researcher is to “be here”, “merging” with the research object.

In contrast to GT, ethnography does not have a concrete and pre-established methodology. It develops and uses a relatively wide selection of research approaches commonly based on techniques such as observation in natural settings, case and artefact studies, interviews, projective tests, along with the

Grounded Theory and Inductive Ethnography 27

much asserted combination of ethnography and triangulation. Mapping is also typical of ethnographic studies (mapping of vast sections of local communities or groups with a focus on their symbolic thinking and behavioural patterns).

Ethnographic techniques vary in accordance with the aims of research. So-called inductive ethnography relies strongly on data, either quantitative or qualitative, while interpretative, critical and postmodern ethnography develops technical critical reflections, representations and narratives. Both approaches typically involve theoretical openness and strong self-reflection.

In technical terms, an ethnographic approach includes three basic steps:



− identification of the research object



− data collection



− analysis of empirical material

Here, the sequence of steps and a continuous reflection on the research work being carried out, including a reflection on one’s own culture and social position, are vital.

In the first step, researchers focus on so-called sensitising concepts (Ragin and Amoroso 2011; Bray in Della Porta et al. 2008) which help researchers indicate the direction of their research. In the data collection phase they focus on their research object in the widest sense, including their submission to the authentic context. So-called submissiveness to the object is established (Bray in Della Porta et al. 2008) along with the abandonment of any pre-established concepts.

At this point, researchers play the role of independent variables. The third step involves the analysis of empiric material in which researchers’ self-reflection comes to the fore, which according to Gadamer (Gadamer in Della Porta et al.

2008) always also includes a pre-understanding in tune with the researcher’s tradition, education and cultural environment.

In comparison with GT, ethnography foregrounds the researcher’s personality and allows for a more flexible attitude to data. A point shared by both approaches is their assumption that the data being studied are the key to the research result and that theory and interpretation are in second place relative to the data.

Apparently, however, due to the described properties of ethnography, this approach holds stronger implications for ethnographic research than for GT.

Fetterman therefore “recommends” (Fetterman in Alvesson and Sköldberg 2000:49) that researchers decide on a theory on the basis of its appropriateness, simplicity and explanatory power. In his opinion, theories’ ideological bases often “blind” researchers rather than guiding them to find ways to successfully process the complexity of data obtained during field work. In this respect, Fetterman’s instruction is simple: when the data do not support the theory, it is time to look for a different theory ( ibid.).

28

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism However, the standpoint that researchers can freely choose a theory, and that its usefulness is simply defined by data, seems somewhat overly naïve. In this respect, we agree with Alvesson that the choice of a theory cannot be simple in the way suggested by Fetterman due to the fact that a theory is always “paradigmatically determined”. Without concepts and theory, “nothing at all emerges as meaningful, as data” (Alvesson and Sköldberg 2000:47). Suitability, explanatory power, and the ability to guide the researcher do not “emerge” atheoretically and aparadigmatically by reference to data. In this respect, ethnographic researchers are faced with the fact that ethnographic data are often ambiguous and interpretatively open so several different theories may appear suitable as a response to a specific research question. However, according to Alvesson this fact is far from implying that a theory freely chosen according to the available data is the most suitable and will bring the optimum result.

1.3 Can the Position of Theory and Method in Ethnography be Strengthened by GT?

Essentially, grounded theory represents analytical choreography whose ultimate aim is to achieve a higher level of abstraction. This aim can be achieved through an in-depth examination of data carried out concurrently with an accurate, conceptual recording. The approach developed by grounded theory requires researchers to be focused on data, to continuously sharpen their sensibility and to gradually develop the final text. Perhaps the biggest advantage of GT is that it forces researchers to continuously reflect on their work and thereby carefully select the studied material.

If, typically, GT is methodologically convergent and linear, this is far from true for ethnographic approach with its typical sensory saturation, cacophony of information and the researcher’s personal involvement. According to Casper (Casper in Bryant and Charmaz 2007), an ethnographer does not create a report, but they instead live it, often with disastrous effects for the research results.

This automatically raises the question of the usefulness of ethnographic research or, in returning to the original question posed in the introductory part of this paper: what constitutes the basis of ethnography? The answer seems simple. The ethnographic approach is designed to understand interactions, and interaction is the “heart” of sociological research.

What is common to all variations of the ethnographic approach in sociology and anthropology is the researcher’s commitment to reconstructing the actors’

everyday world. Through observing the course of the actors’ day-to-day activities and their impact on changes, the researcher tries to explain the collective patterns

Grounded Theory and Inductive Ethnography 29

of social life, leaning on the actors’ “practical knowledge” of their functioning, and aiming to transfer the findings from “micro-situations” to the societal level.

Here, grounded theory can function methodologically and theoretically as a mediator between ethnography and its research subject, i.e. interaction. The methodological approach offered by Glaser and Strauss requires researchers to constantly verify “temporary” findings to finally create,discover a theory.

Researchers’ involvement in the reality of everyday life confers them with a mandate to form their final interpretation (theory).

Less experienced researchers often feel insecure when using the methods of grounded theory. Grounded theory offers a middle way between the “use of data to describe a popular theory” on one hand and field research using no theoretical starting point at all on the other. GT’s commitment to an inductive approach in research distances it from the classical research approach based on verification and the ability to scientifically confute (big theories), although Glaser and Strauss are far from denying the importance of the researcher’s “pre-existing theories” for the course of the research. On the contrary, researchers constantly move between the empirical world and the conceptual world full of abstractions and theories. Therefore, what is the difference between the usual ethnographic field work aiming to test theories and the approach advocated by GT? In response to this question, Bryant and Charmaz (2007) consider Popper’s wel -

known idea about science as falsification. According to Popper, (1998) a theory is scientific and sensible if it is refutable, i.e. if it assumes conceivable attempts which could prove its falseness. Popper thereby repudiates the inductive method which considers an assumption as proven simply through tests that corroborate it: Popper regards this kind of testing as always insecure, with a possibility always existing that the assumption is refuted by a future test. For Popper the scientific method is the hypothetic-deductive method: first you postulate a hypothesis, and then think of attempts to refute it (Popper 1998, Vattimo 2004). At first glance, it is seemingly impossible to bring the requirement of refutability close to GT’s inductive approach. However, Popper’s principle of refutability can also be understood in a way which does not exclude GT, if refutability works as a continual and internal method aiding discovery and theory building (Bryant and Charmaz 2007:11).

For researchers in practice this would mean that temporary micro-theories created within GT and based on empirical material constantly face instances (tests) which can prove them false. In this view, the aim of GT (and analytical induction) is to develop the theory of causality in a way that allows for continuous (internal) refutability (falsification).

For GT this process does not imply acceptance of the assumptions of the deductive approach. In its initial period, GT offered a new vision of how to generate theory

30

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism and new knowledge – through an innovative research practice (applied throughout from the research plan to the writing of a text for publication) – directly from data rather than through testing hypotheses arising from theories (Kavčič in Adam et al.

2012:165). Grounded theory’s abandoning of verification and the “improvement of big theories” enables it to make sense, make meaning and provide conceptual relevance to categories in the course of research, whereby it comes very close to ethnographic approaches.

1.4 Can this Improve Ethnography?

According to Bryant and Charmaz, (2007:15) the answer to this question is more complicated than it first appears. This is presumably due to GT being relatively rarely used in qualitative research, along with opinions about the usefulness of the GT method being strongly divided even among the researchers who use it. They agree more strongly that GT’s research purpose differs from original or classical ethnographic approaches, i.e. rather than mainly describing a phenomenon or process GT emphasises their study, thus resting on the attempt to conceptually interpret the functioning of the actor or the observed phenomenon. Compared to common ethnography, GT is much more analytical, and much less conventional in its use of sources, committed to developing middle range theories by laying emphasis on a processual rather than a structural approach (Bryant and Charmaz 2007). For GT science is not systematised “common sense”.

There are also problems regarding the provision of a description of the criteria defining what good ethnography is or how to understand its contribution (research scope). Researchers’ answers to this question depend on the research traditions or schools to which they belong. Within this spectrum there are researchers who see ethnography’s contribution in the context of the ethnographic method’s ability for political intervention or its potential for social change. In this sense, these researchers’ perspective of the role of ethnography is different from the perspective of researchers from the post-modern school who are mainly interested in ethnographic approaches for their ways of understanding and interpreting the everyday world, i.e. they are interested in what is called ethnographic reflection. Recently, so-called “realistic” ethnographers have been asserting themselves and they primarily point out ethnography’s methodological contribution to research or its insistence on the validity of research results.

Therefore, it can be agreed that having emerged as a reaction to “futile” research work limited to descriptions while contributing only a little to the development of theories, and by pointing out the significance of conceptual work and the generation of theories, GT does strengthen ethnography. However, since GT’s

Grounded Theory and Inductive Ethnography 31

approach has its own disadvantages it is important for ethnography in what way GT enters ethnography. According to Bryant and Charmaz (2007), GT should not be a short-cut in the process of collecting data or a handy excuse for the absence of a theory. Although time spent in the field is perhaps ethnography’s biggest advantage, it is at this very point that GT can be extremely helpful by being able to importantly contribute to structuring the time of the fieldwork phase of research by suggesting the research priorities (which area of fieldwork needs more attention), by cal ing attention to “grey areas” in the research process, directing towards the study of theoretical sources etc. By compelling researchers to organise their empirical material, grounded theory makes writing become a form of understanding and analysis rather than just an analytical prerequisite.

Moreover, the role of GT is no smaller in the education of ethnographers.

1.5 Conclusion

Silverman (1997) suggests the following two criteria for evaluating research: (a) the persuasiveness of the researchers in the substantiation and interpretation of their findings; and (b) the theoretical and practical relevance of the research question. We can agree with Silverman on the importance of these criteria, and add the requirement of methodological rigour which should also apply to approaches in qualitative research if the latter wants to avoid the “anything goes”

reproach. Good ethnography (and this is where we see the greatest contribution of GT considering data management) should thus include:



− credible empirical arguments;



− interpretative “openness” in explaining social phenomena;



− a critical reflection on the political and ideological contexts of the research;



− taking the indexicality of speech (observance of contexts) into account; and



− the generation of theory.

Perhaps the most important aspect of the ethnographic approach should be “richness in points” which, however, is rarely achieved through ethnographic methods due to their disadvantages described above. While drawing on empirical material, research “rich in points” concurrently reaches beyond it, through interpretation and conclusions. This is what Bourdieu calls the “epistemological rupture”, denoting the rupture with “every-day knowledge”. In this view, an interpretation that is “rich in points” refers to empirical material, but without necessarily taking it as “firm proof ”. Whilst the data support and inspire the interpretation and offer arguments, they do not require (or allow for) its unequivocalness. Empirical data can also prevent an interpretation or lead to its senselessness, all of which speaks

32

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism of the interpretation’s meaning. Typically, a research study rich in points involves a tension between the empirical material and the researcher’s imagination in creating an interpretative breadth and depth of the repertoire of data. In this view, GT can enable ethnography its interpretative creativity which the classical ethnographic descriptive cacophony does not. If limited only to what is common knowledge (and which is established after a lengthy observation) a research endeavour is senseless.

Avoiding making any definite statements about “how things are”, research rich in points emphasises the importance of “looking at things in some particular way”

(Alvesson in Sköldberg 2000:277) which allows for a new understanding.

According to the criteria of interpretative richness, a good research study enables a qualitatively new understanding of relevant fragments of social reality, and is capable of challenging the leading ideas and raising key questions about the way contemporary societies function, in turn changing the “taken-for-granted” model.

This is the only way to enable the creation of new research alternatives.



Researching Communication:

The Interpretive Approach Between

2 Theory and Practice

A proper understanding of communication research and the way it has been carried out cannot emerge without some consideration of the theoretical backgrounds of the different methodological approaches to communication analysis.1 Over the years, communication research has borrowed from and been dominated by social science disciplines such as political science, anthropology, psychology, linguistics, sociology and philosophy, which communication scholars have identified as interdisciplinary characteristics. In the past decade, the methodology of communication research has been experiencing rapid development directly related to the development of social science research.

The development of methods in communication studies has thus represented an upgrading of the development in wider social science research (Stempel and Westley 1989). This development is seen both in empirical research with the assertion of findings in cognitive sciences as well as in the epistemology of social sciences and through the development of so-called reflexive methodology.

According to Alvesson and Sköldberg, the syntagm reflexive methodology (2000) denotes complex relationships between the knowledge-development processes and variable contexts in which knowledge develops, including all actors. For research in communication studies and more broadly in social science, reflection on these relationships carries important methodological implications.

Typically, compared to “traditional” methodology, reflexive methodology fully accepts the fact that language, culture, social structure, norms, ideology, discourses etc. make up a constitutive part of the scientific process. These elements necessarily interfere with the relationship between the empirical reality and attempts to implement segments of this reality in research findings (text), which largely influences the research results or even creates conditions for their validity. While this realisation may be considered a justified criticism of unreflected empiricism,2 on the other hand a radical critique of empirical research triggers the “resignation” of research and questioning about whether empirical communication studies and also social science are at all theoretically justifiable.3 The larger part of the traditional methodological literature does not 1 The previous version of this text was published in journal Medijska istraživanja, 2013, Vol. 19, No. 1, p. 19–38.

2 In this context we link the term “empiricism” with those practices in research that use empirical research as the key criterion to reflect reality. In this view, research is primarily considered as the collection, processing and analysis of data, both qualitative and quantitative. Theory and data feature separately with the value of theory being tested against data (Alvesson and Sköldberg 2000:10).

3 In the background of this questioning lies the idea of research whose key and distinguishing characteristic is a continuous interpretation of and reflection on all research elements. A common mistake of researchers is that interpretation is understood as a “technical” element

36

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism offer satisfactory answers to these questions. Thus, either uninterpreted facts in the form of “pure” data or the approach whereby even an everyday human experience is denoted as a “discursive construct” are offered as two extremes of a range of answers to the question: what is the key element of social science research?

Therefore, researchers’ quandaries are related to the fact that contemporary communication studies operate both with open empirical material enabling its interpretive equivalence, and with qualitative contents which “succumb”

to rigid categorisation. As a result, researchers are growingly renouncing the precise separation of standardised and non-standardised research approaches, and thereby the key measure of differentiation between quantitative and qualitative methods. Alvesson and Sköldberg (2000:150) point this out when claiming that even with the highest level of methodological differentiation and rigour researchers cannot avoid what is believed to be the essence of research in communication studies and more broadly in social science; namely, that they form part of the society whose social relations they are studying, which in itself co-creates the research results.

Although in this context a justifiable and constant question is whether research results are not only a peculiar construct of researchers, this does not imply that we can legitimately distinguish methodological approaches whose research results are constructs, and those which do not “construct” research results.

Research practice shows that we can justifiably only speak of approaches that are not aware of the “construct”, and those which admit the “construct” and try to explain the nature of this fact and its impact on the research results. While being aware of the deficiency of this fairly rough simplification, later in the chapter we address only one of these approaches which is classified among the so-called epistemological fundaments of research in social sciences, namely the poststructuralist theory. However, in doing this, we have no intention to reduce the significance of other theories which constitute the epistemological arc of social sciences research; on the contrary. Contemporary reflexive methodology quite directly refers to hermeneutics, linguistic philosophy, critical theory, and to the recently particularly topical discourse analysis or contemporary discursive theory (comp. Vezovnik 2008). The emphasis laid in this context on post-structuralism is linked to the key element of reflexive methodology, i.e. the interpretation of research results.4 In the context of empirical research, reflexivity should be of research which follows the collection and processing of empirical material. In this case, the epistemological dimension of research is completely overlooked.

4 This chapter addresses “reflexivity” as a specificity (specific characteristic) of research which integrates the “reflexivity” of different research approaches, which are mainly asserted by

Researching Communication

37

understood as the interpretation of interpretation (Alvesson and Sköldberg 2000:6) or as the critical questioning of the researcher’s own interpretation of the empirical material. In this view, post-structuralism compared to other theories stands out with its requirement to separate the text from the “outside” reality.

For post-structuralists, a text has its own life without any outside reference, with only the influence of other texts being relevant to its understanding and interpretation.5 This assumption questions the authority of the authors (in our case the researchers) as well as the possibility of the “empirical reconstruction”

of reality. In other words, for contemporary methodology both the subject (researcher) and the object (empirical reality) of research are seen as problematic.

Consequently, the parsing (empirical) methodologies do not end with only the metric and interpretive. Just as the interpretive has entered the critical, the critical has likewise entered the empirical (Anderson 2012). During the past decade a corpus of critical- empirical methodologies has appeared. These are reflexive methodologies that apply cultural/critical interpretive methods to empirical texts. They are influenced by social communication studies now mostly known as the postmodern rejection of the universal narrative (Anderson 2012:18). They accept the requirement for some empirical grounding, but are less interested in telling the what or how of text than in understanding the cultural force of a class of texts or in promoting what ought to be the cultural consequences of their engagement ( ibid.). The reflexive methodologies go by a number of different names – discourse analysis, cultural studies, interpretative studies etc.

In this chapter we turn to the media research methods that can be considered under the category of interpretive analysis. Speaking in broad terms, methods of interpretive analysis seek to read media content of all kinds and draw from it particular socio-cultural meanings that the contents potentially create and circulate. Namely, we can think of media content as a text that offers lessons in how to understand the social, cultural and political worlds within which we live our daily lives.

ethnomethodological ethnography, critical phenomenology, post-modern sociology, and authors such as Gouldner and Giddens (double hermeneutics). The concepts of reflexivity and reflection mostly appear in relation to the processes of “developing” knowledge and different contexts, in which knowledge develops, including the actors (researchers). Here we lean on the concept created by Alvesson and Sköldberg who understand reflexivity as the intertwining of linguistic, social, political and theoretical elements in the process of

“developing” (constructing) and interpreting empirical material.

5 While a text is unquestionably an “echo” of another text, this does not mean that researchers can renounce its relationship with the “outside” reality. Thus, a text should more adequately be understood as a metaphor of a social phenomenon “containing” the tension between the research object and its representative (discourse, text), in which it is important for researchers to recognise this tension and “release” it at the level of interpretation.

2.1 Theory and Methodological Consequences The theory examined in the following paragraphs in terms of its impact on contemporary qualitative communication research dates back to the 1960s, to authors such as Levi-Strauss, Lacan, Barthes and others who are more or less justifiably classified first in the structuralist circle, and later with the arrival of Derrida (the concept of deconstruction), Lyotard (the concept of knowledge) and Foucault (the concept of subject and power) in the post-structuralist field.

The “structuralist revolution” broke off with the philosophical tradition which had grounded “Western” thought from Descartes to Sartre by introducing structuralist analysis or concepts such as structure, rules, codes, system and others into the interpretation of social phenomena. The starting point for structuralists is a rejection of the concept of subject which had dominated the Western humanist tradition for decades. They see the subject as derived, secondary and marginal, the effect of language, the unconscious and culture, without any creative potential. By foregrounding the unconscious, the symbolic and communication-social relationships, structuralism positions the “methodological consequences”

accordingly: meanings and interpretations are not the result of transparent intentions of autonomous subjects; instead, the subject itself is the result of the linguistic relationship and subjectivity is a social and linguistic construct (Alvesson and Sköldberg 2000; Best and Kellner 1991; Vattimo 2004).

While post-structuralists adopted the key elements of the thesis of structural linguistics, they completely rejected the concept of the so-called dominant centre

“governing” the structure. Instead, centre stage is occupied by the text which post-structuralists regard as a mere play of signs without reference to an author (the subject) or the outside world. At the same time, this is post-structuralism’s most radical methodological innovation. For post-structuralists the text6 is separated from the outside world and thus unburdened of all references.7

6 “There’s nothing outside the text” states Derrida’s first commandment. On the other hand, the entire culture is a text in which everything we consider to be real has its own structure. As individuals we are permanently caught up in a network of political, linguistic, historical and other structures; which is why deconstruction aims at breaking through borders, opening up, expanding and increasing the complexity, creating space for what is only emerging (Campbel 2005).

7 Derrida’s key criticism refers to the so-called “metaphysics of presence”. The metaphysics of presence is a concept according to which there is always another, more important, more authentic, primal, either temporary or conceptual entity lying in the background of the observed phenomenon. Therefore, Derrida criticises the idea that there is always something behind what we observe that is hidden but crucial to the understanding of the phenomenon. For Derrida and post-structuralists in general the phenomenon does not have any “background” and so everything that is relevant for the researcher lies on the surface (Alvesson and Sköldberg 2000).

Researching Communication

39

This challenges both researchers’ authority in the research process and their possibility to reproduce outside reality in text.

Post-structuralists’ key reference becomes language which is ambivalent, ambiguous, metaphorical and constitutive rather than unambiguous, literal and descriptive. With such an understanding of language (speech) post-structuralists problematise the concept of objectivity, clarity and rationality of the research procedure. For the needs of research this problem can be reformulated into the question posed by Alvesson and Sköldberg (2000:152), i.e. can the researcher tell anything about “reality” which in terms of credibility and quality would reach beyond the opinions of others about the same reality?

Communication scholars and methodologists who lean on post-structuralist ideas are primarily interested in the concept of the “emergence” of theories, along with discursive strategies and the understanding of authority developed by post-structuralists. In this regard, post-structuralists do not create a theoretical frame of reference to guide researchers towards unambiguous logical results and interpretations. On the contrary, they strive for multiple and variable interpretations of results which should demonstrate the inconsistency and fragmentation of the (media) “text” (reality).

Indeed, the key post-structuralist authors do not even speak of methodology, which is why post-structuralism is believed to be anti-methodological. Yet it can be claimed that post-structuralism comes close to the method by asserting the concept of insightfulness, anti-objectivist interpretation and, of course, deconstruction – which is given more attention later in this chapter.

Naturally, the key to understanding the post-structuralist attitude to empirical research does not lie in the idea of “portraying reality” since for post-structuralists there is no such thing as neutral and simply interpretable reality. If in these terms we ask, what is the scholar’s key task then for post-structuralists the answer is clear: active work with the language and the text or their local, contextual and arbitrary nature.8 Obviously, for communication scholars this answer triggers more dilemmas than it solves. While it is impossible to address all of them here, we list three of them that can be considered inherent to contemporary methodological approaches with regard to the “specific” post-structuralist view, namely: 8 Scholars who rely on post-structuralist theory consider as the most important the metaphorical nature of language or the idea that in social sciences the text is the metaphor for a social phenomenon. The metaphor allows the scholar to gain an insight into the phenomenon by creating “tension” between the object of study (society, organisation, the subject) on one side and the “modifier” (text, discourse) of these objects on the other. Therefore, it is important for researchers to recognise this tension, yet without allowing the metaphor to define the meaning (Alvesson and Sköldberg 2000:179).

40

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism



− the attitude to empirical reality: post-structuralism consistently rejects the concept of a definite (irrefutable and absolute) truth, at both theoretical and interpretive levels;



− the attitude to the nature of social facts: post-structuralism primarily directs the attention of researchers to the disclosure of ambivalence, divergence and differentiation of communication and social facts which figure as self-understood constructs; and



− the attitude to the researcher–respondent relationship: post-structuralism problematizes the traditional relationship between researchers and respondents by foregrounding the unresolved question of “structural violence” or the fact that no matter how rigorous the methodology it does not eradicate the effects of social structure which are inscribed in the relationship between the researcher and the researched (Bourdieu et al. 2002).

In the rest of this chapter we will use these dilemmas as my point of departure for examining the frame of reference of contemporary (reflexive) methodology and the role of post-structuralism which we see as one of the key frame theories. We raise a theoretical research question about the influence of poststructuralist ideas on the formation of contemporary reflexive methodology (are they identifiable and how do they manifest themselves). Further on, we will be interested in whether the pinpointing of the influences can be used as the starting point for making recommendations for research strategies.

First, we look for answers to these questions by examining Derrida’s concept of deconstruction and the related two methodological dilemmas, namely the position of the humanistic and communication subject in research and the role of the researcher as an author. We link both dilemmas with those points in the post-structuralist theory which in the opinion of different authors hold direct implications for a contemporary reflexive, particularly qualitative methodology of communication.9

2.2 De(con)struction as the Methodological Means The starting point of Derrida’s deconstruction can be found in the metaphor of every, even the smallest and virtually invisible crack in the façade being a symptom of the flaw of the entire edifice (Alvesson and Sköldberg 2000:154).

9 In contemporary social science research a rigid distinction between qualitative and quantitative methods has become increasingly rare. Methodological discussions highlight the ontological and epistemological foundations of research, assuming that deciding on one method or another depends on the nature of the research question rather than belonging to a “school” or a theory.

Researching Communication

41

Hence the crack in the text, to use the same metaphor, lays bare the apparent unity of the text by unveiling the hidden, the repressed and the subjected in the text (relationship).

The first step in deconstruction requires that things be turned upside down, making the oppressed side the dominating one. Rather than implying the mere inversion of the hierarchical relationship between the two opposing sides, this procedure in the second step fully undermines the differences between these opposites. This is also how Derrida explains the word de-con-struction. The first step thus involves the destruction of the original dominating picture where what was hidden before now becomes dominating, and the next step implies the destruction of both opposing poles with their simultaneous “displacement” and the construction of something new and wider in which both sides constitute cases in their own right.

Through the assertion of the deconstruction concept, Derrida first criticised the apparent unity of the spoken word and reference. According to Derrida, in semiotics, linguistics and other theories of the meaning of communication the spoken word dominates over the written, which he terms “phonocentrism”

(meaning can be more immediately expressed through speech). Phonocentrism leads to the wider and better known notion of logocentrism which represents the tendency dominating the Western system of thought that everything has logical/rational grounds, and therefore such are (or have to be) also the goals and conclusions.

Derrida’s criticism of the privileged status of the spoken word leads to the demonstration of the importance of the written word or the inversion of the hierarchical relationship between the spoken and the written word, which some of his critics define as graphocentrism. Considering the requirement of deconstruction that the deconstruction process should not only imply the inversion of positions (in our case of the spoken and written word), according to Derrida a third phase of deconstruction is needed, namely a different way of creating a text with non-referencing signs, which should eradicate the hierarchical relationship. According to Derrida, texts should be read so as to dismantle the authoritarian game of which texts are the carriers each time they want to tell “the ultimate truth”, when trying to really speak about things as they are. This is about “laying bare” the intertextual play of presence and absence, such as the cause in the consequence, the intentions of an individual’s behaviour, the essence of power from its use etc., with which we dismantle a given and self-understood hierarchy. The analytical tool used in this context is the so-called conceptual triangle consisting of the word-meaning-object or the more well-known sign-signifier-signified. Therefore, conceptually, the word specifies a set of properties which together constitute its meaning and real phenomena

42

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism (referents) that possess these properties10 (Alvesson and Sköldberg 2000; Best and Kellner 1991; Derrida 1988, 1997; Kembel 2005).

Along with some other post-structuralist emphases (mainly Lyotard’s critique of

“the grand narratives”), Derrida’s concept of deconstruction generates two basic methodological dilemmas: (1) the question of the position of the humanistic subject in research; and (2) the question of the researcher as the author of research (Alvesson and Sköldberg 2000:164, 167).

2.2.1 Position of the Humanistic Subject in Research Post-modern authors and even more those who consider themselves poststructuralists consider the idea of the autonomous individual as an active individuum, the holder of sense and meaning, a missed invention of Western thought. Contrary to the prevailing idea of a coherent and fully integrated individual, post-structuralists wish to decentre the idea of individuum, and thus displace the emphasis from classical constructs, perceptions, emotions, and actions towards the so-called discursive context that constitutes the expression of subjectivity as limited in time and space. Subjectivity here is the conscious and the unconscious, emotions and perceptions, the individual’s self-insight and attitude to the surrounding world. In this sense, rather than language being the expression of subjectivity, on the contrary it constitutes subjectivity.

It is with language that we make the experience of the world (Vattimo 2004:91) and this is why subjectivity is unstable, contradictory and more of a process than a structure. It depends on the form of address which form of subjectivity will be constituted. An individual can be interpellated as a man, a journalist, a taxpayer, a drug addict, a Slovenian etc., in which the interpellation of the subject depends on the available discourses or discursive contexts that create different meanings of social roles, identities or relationships. Weedon (in Alvesson and Sköldberg 2000:165) notes that the individual’s experience has no innate essential meaning, but acquires its sense in language from “discursive systems of meaning” which often contain contradictory versions of how social reality should be described.

Accordingly, discourses do neither emerge freely nor randomly. Certain discourses are dominant, for example the discourses of gender roles, consumerism, normality or political participation, while others can be completely overlooked due to the nature of the social structure and the power relations.

10 Derrida believes words are the “doubling of absence” and, moreover, are built into a linguistic community that is historically and culturally determined, which causes an even further deviation from directness. That is why for Derrida the right model of language is writing (text) rather than the spoken word (Vattimo 2004:92).

Researching Communication

43

For post-structuralists the disclosure of subtle dimensions of the text is related to dismantling the real nature of authority and power. However, as opposed to critical theory post-structuralists are not interested in the emancipatory potential of knowledge. Margolis (1989) regards Derrida’s approach as conservative in his disinterest in reality or the insistence that nothing definite can be said about reality as such and it is even less possible to suggest a concrete social change. While the reproach of being apolitical applies to the entire opus of post-structuralism, this is, however, only partly true. Derrida’s approach can also be defined as subversive in it disclosing the prevailing forms of social relations or the violence of the social structure. However, the post-structuralist “obsession” with authority also has concrete methodological consequences. Post-structuralists’ fear of being authoritarian (with the scholar’s arbitrary intervention in the empirical material) is the same as their fear of the authoritarianism of the system. While for poststructuralists the Word is omnipotent and the key to understanding the world, at the same time they renounce contact with empirical material due to the fear of becoming “tainted” (Alvesson and Sköldberg 2000:240).

With the idea of the decentring of the subject, post-structuralism has in many ways “shaken” the foundations of social sciences. In this light, the poststructuralist requirement of rethinking the traditional sociological categories and primarily the concept of the autonomous rational subject has become legitimate.

2.2.2 The Question of the Researcher as the Author of Research With the idea that language or speech is the key element of research, the question of authorship also becomes crucial, especially in ethnographic communication studies. The fact that the abundance of empirical material prepared to be analysed requires a researcher’s radical intervention (reduction) has revealed a banal realisation that the final text (research result) is not determined by the empirical reality. In this light, post-structuralists see the research process as a

“totalising” description of a reality, where the researcher speaks on behalf of the “Other” so that, rather than being reinforced, the voice of others (ignored, repressed, subjected) is de facto made impossible (comp. Ragin and Amoroso 2011).

The key methodological question for post-structuralism therefore becomes the question of representation or the idea with which post-structuralism completely rejects traditional views according to which the recording of data is only one part of the research process which is determined by theory, data collection, analysis and interpretation. For post-structuralists there is no neutral social reality which can be objectively described and interpreted. Therefore, the researcher’s

44

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism task is anything but an objective description of reality: it is to make visible the text’s ambivalent and contextual nature by adopting an active attitude (values, interests) to it. Fictional elements or authors’ creative inventions have become an integral part of every interpretation that most post-structuralists see in two ways: on one side as the question of the reproduction/construction of a specific reality (re-presentation) in terms of whether it comprises the totality and how it explains it and, on the other side, as the question of what or whose interests predominate in the research. For researchers this chiefly means the recognition and taking into account of the fact that research is primarily an interpretive activity with the researcher as the key actor: interpretation cannot stand preconceived interpretive schemes, instead crucial elements are the researcher’s judgement, intuition and ability to see (more and beyond), in a more or less explicit dialogue with all of the actors in the research (comp. Uhan 1998).

2.3 Starting Points for a Different Methodology of Communication In this context, the examination of the above methodological dilemmas indicates clear starting points for the generation of a contemporary reflexive methodology. In the following paragraphs we first address them as the poststructuralists’ call for a different ethnography of communication studies, and then we present the mechanisms for the selection of research participants as the methodological means to prevent the reproduction of the asymmetry of social power of the actors in the research. Methodological implications also include the necessary understanding of the relationship between the researcher as the actor and the empirical material, while also highlighting the need to take account of alternative presentations and perspectives of all actors in research. We conclude the examination of methodological starting points with post-structuralist critical remark regarding interpretations of the contents of interviews as a technique of collecting empirical material.

2.3.1 The Post-Structuralist Call for a Different Ethnography One of the key post-structuralist emphases is their call for a different ethnography.

According to post-structuralists, the pronounced disadvantage of traditional ethnography lies in it being fixed to an empirical basis and not being sensitive enough to the ideological background of theories which researchers refer to in their interpretation of data. Post-structuralists believe the combination of both prevents the interpretation of a social cosmogony. An apparent opportunity is found in the ethnographic technique or so-called “giving word”. It is a technique

Researching Communication

45

which enables the disclosure of subtle aspects and properties of overlooked groups which should help researchers form better narratives of their experiences (Ragin and Amoroso 2011). The problem of traditional ethnography here is that, in the post-structuralist view, it totally overlooks the “intentional symbolic violence” or impacts of the social structure which become inscribed in the relationship between the researcher and the researched. The post-structuralist methodological innovation proposes researchers’ consistent self-limiting when attributing meanings to a researched phenomenon (transcripts are already interpretations!), while simultaneously abandoning the idea (the illusion!) of discourses speaking for themselves (Bourdieu 2002:620).11

In this sense, observation taking place in a natural context also does not bring a perfect solution due to it being under the strong influence of the indefinite nature of language – social interactions are constructed by speech and, in this context, actors’ involvement is only conferring meaning or sense to their own existence.

Compared to the traditional approach12 based on the coding and synthesis of findings (reduction), post-structuralists assert the norm of the “open text”, which in practice means that researchers look for variations in empirical material (observations, interviews) that are included in the interpretation in an equal way.

Only in this way is it possible to express the multiplicity of a subject’s identities that are not only dominant, but also particular and marginal. Therefore, the researcher must be aware that, by interfering with what has been said, they are creating a narrative and so they should think hard about how they do it. It is important that they recognise both the “unreflected sociological categories”

(such as a journalist, a woman, an audience member etc.) as well as discursive constructs which create effects. An example of poor research practice is the usual requirement of the researcher that the respondent describes, for example, their relationship with their superior; such a requirement creates a fixed identity which involves the relationship of dependence or superiority and subordination while at the same time excluding all other equally relevant individuals’ positions (Alvesson and Sköldberg 2000; Bray in Della Porta and Keating 2008).

11 Post-structuralists see the solution in taking into account (disclosing) the hidden dimensions of their own interpretation. This can be done by “confronting” different interpretations which

“disclose” the position of the dominant interpretation and the alternative one. In this way the text can remain “open”.

12 In this chapter, we use the concept “traditional” to denote those approaches based on making a consistent distinction between quantitative and qualitative methods or standardised and non-standardised techniques of data collection. In this sense, contemporary (non-traditional) approaches are those which “redirect” the researcher’s attention away from the “traditional”

examination of empirical data to taking the cognitive, theoretical, intertextual and linguistic contexts of the emergence (collecting) of data into account.

46

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism 2.3.2 The Selection of Research Participants as the Reproduction of Power Asymmetry

It is completely erroneous to assume that for the results in qualitative research to be valid it is important to consider all actors that can be relevantly categorised.

What is much more important for the validity of the findings is the process of so-called “reflected exclusion” or a well-grounded plurality of voices which can balance the perspectives of the actors included in empirical material to the greatest extent. Empirically, there are two dimensions of the exclusion problem: on one hand, it is about the (non)representation of different groups, categories and individuals in the research process or text and, on the other, there is the question of the (non)representation of these subjects within these groups or categories – thus, it is about whose voice is heard or disregarded.13 It is important to point out that “silencing” is an integral part of categorisation or “locking subjects into identities” (Alvesson and Sköldberg 2000:189).

For both post-structuralists and critical theorists, the question of exclusion is essentially a political question or a question of the nature of the social structure.

Post-structuralists believe the research instrument ( e.g. a questionnaire) reflects the relationship of social power among individual social groups. Authors within both approaches (critical social theory and post-structuralism) see the institute of the neutral research question as merely concealing the problem of a selective construction of the world. However, in their examination of political discourse post-structuralists move away from the engaged critical theory by mainly directing their interest to the problem of the researcher’s interpretation of reality, with the disclosure of asymmetries in power relations among social actors or the description of forms of ideological domination only coming in second place.

Thus, for post-structuralists the question of the exclusion and inclusion of participants in research is not primarily a question of which groups or categories are included in the research or how researchers influence the selection process.

For the research to have “weight” it is more important to find out the ways in which what actors say (what has been uttered) has been changed, disregarded or distorted with the interpretation. In describing their reality, subjects use different strategies and can also represent “voices” of others, which is why researchers have to encourage the ambivalence of empirical material rather than limiting it.

13 For Bourdieu, the individual’s history is never anything else than a “concrete speciality in the collective history of his stratum or class” (Bourdieu in Della Porta and Keating 2008:304).

Researching Communication

47

2.3.3 Complementarity of the Researcher and Empirical Material The selection of research participants and the way of the researcher’s interpretation of the material are the key elements of a plausible text. Poststructuralists see this fact as placing researchers side by side all other research actors. This is perhaps the most characteristic contribution of post-structuralist theory to understanding the processes in qualitative research. Post-structuralism strives for a well-thought-out process of exclusion or inclusion of subjects in research by representing and reading of empirical material. A frame of reference is needed for this process to enable the researcher to initially bring to the fore and strengthen, at the theoretical level, those relationships, roles or subject identities that are structurally discriminated; this is, for example, something traditional communication studies cannot do. For instance, if in an examination of media representation of sexual discrimination the discrimination occurs in a context which is not problematised by male or female journalists, this phenomenon cannot be characterised within traditional communication studies. This can only be done within a frame of reference in which theory can be used as the framework for the recognition and detection of unreflected structural violence expressed by the interviewees in the research.

In this sense, the researcher has a justified role to critically judge whose position is undervalued or whose voice in the research is not heard or is excluded as a result of intentional discursive violence.

However, it is necessary to point out that the critical reading which should disclose the ambivalence of the text and avoid attributing dominant meanings to individual phenomena is not the same as so-called emphasised reading. On the contrary, the reservation of researchers to form an unambiguous interpretation by pointing out one dimension (emphasised reading) of the text is not mutually exclusive of making an interpretation in which researchers intentionally strengthen the voice of an individual subject so as to compensate for the

“systemic loss” or the subordinated position of that subject in the social structure.

In practical terms, this dilemma can be solved if researchers expose their own interpretation to an alternative reading after the research or become informed about different views before conducting the research. In a post-structuralist light the researcher’s position is complementary to the empirical material, and not in any way an alternative to the empirical nature of the material.

2.3.4 Taking Alternative Representations and Actors’ Perspectives into Account Typically, the post-structuralist approach is sceptical of the theoretical frames of reference which should define the horizon of meaning. For post-structuralism

48

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism the relationship between the world and experience, text and reality, structure and action remains undefinable in terms of meaning, making the researcher’s reference to a theoretical frame of reference unproductive since a theoretical solution which would impose order onto the universe that is permeated with meanings is simply impossible (Marcous in Alvesson and Sköldberg 2000:191).

Although researchers are aware of the ambiguity, inconsistency and contradiction of empirical material, they most often miss with their interpretation because they try to harmonise the text by referring to a frame of reference. Post-structuralists see the solution in the confrontation of different interpretations, which initiates a play between a dominating and an alternative interpretation, leading to the opening up of the text. This is also what Derrida’s idea of conversion mentioned above refers to, which explains how a certain meaning depends on the repressed opposite or how the meaning of a specific sign is influenced by the absence of all other signs.14

Accordingly, the parallel interpretation and confrontation of perspectives are important for establishing the meaning as they enable the researcher to creatively combine all elements. With multiple readings a productive “tension” can be achieved, which can then only be released by thinking about which questions these perspectives require. The purpose of the whole procedure is to avoid making a synthesis of findings or form conclusions at too early a stage.

2.3.5 Taking Responsibility for the Text and its Interpretation One of the post-structuralist reprimands of empirical researchers is that they tend to hide behind an almost bureaucratic methodological procedure and the dominant conventions of writing scientific texts.

In this context, the post-structuralist theory of the subject as a discursive construction dominated by language and context holds major implications for the understanding of a research interview. For post-structuralists, how an interviewee represents reality in an interview has little to do with the reality itself. In the process of interviewing temporary subjectivities are formed which represent reality in relation to the local discursive context created by the interview (Alvesson and Sköldberg 2000:193).

The emphasis on the meaning of language primarily influences the understanding of interpretation. Post-structuralists substitute the modernist idea that language only reflects a complex reality with the idea of the constitutive (non-referential) nature of language. In this context, speech becomes the medium for 14 Similarly, Campbell says: “For Derrida every word is encircled by a fictitious “absence” of other words from which it differs. In a given system of signs, the meaning of an individual sign is what it is for the very absence of all other signs” (Kembel 2004:375).

Researching Communication

49

understanding what lies beyond empiricism and the theories that lean on it. In their analysis of speech, post-structuralists primarily focus on four dimensions which become the subject of interpretation, namely: (1) what we know and say (known uttered, said); (2) what we know but do not say (unknown uttered); (3) what we say but do not know (uttered unknown); and (4) what we do not say and do not know (unuttered, unsaid unknown) (Francois 2009).

What are crucial for interpretation are the researcher’s judgement, intuition and ability to recognise hidden dimensions of the problem which are reflected in what has been said. Therefore, interpretation is not based on self-evident, simple and unambiguous methodological procedures, but on explicit dialogue (Alvesson and Sköldberg 2000:248) between the subject, researcher and reader.

The main point of reflexive interpretation is that in this sense it “brings out”

the relationships among the participating actors and makes the research process and the power relations as transparent as possible. The challenge of interpretive analysis is to offer an insightful reading of the text, one that helps us better understand both the text itself and the issues of social, cultural and political significance of which the text speaks (Anderson 2012).

A frequent weakness of qualitative research practice is the fact that researchers consider or put too much weight on just one element in the research – either the empirical material and its interpretation or the sociolinguistic relations among the actors.

2.4 How to Formulate Research Strategies?

From a post-structuralist viewpoint the whole set of traditional methodological strategies is problematic. Reality cannot be grasped, described and explained by raising questions and translating answers into theory in the way that applies to empirical methods. While post-structuralist criticism has dissuaded numerous theoreticians from empirical research, many researchers suffer from “research resignation” which is mainly revealed as constant self-questioning about the sense of empirical research. Do post-structuralist “findings” and post-structuralism’s anti-methodological nature create a sufficient basis for abandoning the idea of traditional empirical sociological research? Not necessarily. The fact that “data handling” is not central to post-structuralist theoreticians can largely explain post-structuralist reservations regarding traditional methodological themes.

On the other hand, post-structuralist critiques of empirical research can strengthen researchers’ methodological ambitions by moving methodological attention away from looking for the “empirical truth” in data towards an interpretation of and reflection on data in the global context including the

50

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism ideological, metatheoretical, linguistic and political dimensions of research.

Alvesson and Sköldberg (2000:248) speak of reflexive interpretation which encourages researchers to look for answers to research questions beyond self-understood schemes at the interface of the positions of the research objects, researchers and “readers”. This means that in practice research occurs at three levels – empirical, interpretative and critical-interpretative. The latter means that reflexive interpretation does not simply examine the random structure of interaction, but interaction as the transaction of actors with the intention to realistically reconstruct the relationships among them. In this respect, reflexive interpretation rejects both pure empiricism or data reductionism as well as theoretical reductionism, which is shown as the domination of either gender, discursive, structural or cultural references of the chosen theory.

Based on this and in relation to the dilemmas and the research question we posed in the first part of this chapter, the following recommendations can be offered for a contemporary qualitative methodology in communication studies: 1. with regard to the dilemma or attitude of contemporary methodology regarding the empirical reality we suggest the employment of the practice of the consistent inclusion of alternative presentations of communication in the research procedure, including the critical and reflected use of different theoretical perspectives;

2. in relation to the requirement that contemporary methodology be perceptive to disclosing the ambivalence, divergence and differentiation of the studied social facts, we suggest the pluralism of perspectives of actors in research, which Alvesson and Sköldberg (2000:194) understand as the use of the potential of different identities (voices), associated with different groups and individuals and their positions or special interests which are the subject of research or a constitutive part of the researched (media) texts; and 3. with respect to understanding the relationship between the researcher and the respondent in contemporary methodology, we suggest the development of research approaches in communication studies that will increase receptiveness of variations in what the research subjects convey along with the possibility to accept the multiple representations an individual respondent can contribute to a topic. This suggestion sensibly relates to Bourdieu’s “recommendations”

for researchers regarding the choice of respondents and understanding the structure of relations between researchers and respondents. Bourdieu (Bourdieu et al. 2002:610) draws attention to the importance of social closeness (even affiliation) between researchers and respondents which he claims would enable researchers to conduct research in conditions of so-called non-violent communication. According to Bourdieu, non-violent communication solves

Researching Communication

51

the question of structural violence (the effects of the social structure which are inscribed in the relationship between the researcher and the researched) by enabling the exchangeability of social experience and thus reducing the danger of the researcher making subjective conclusions about “objective factors” (the functioning of the social structure). At the same time, social closeness enables an authentic interpretation by facilitating good knowledge of the communication context (slang, jargon etc.) (Bourdieu et al. 2002).

Nevertheless, the most important methodological commitment of the contemporary reflexive methodology of communication remains the creation of an open text which encourages active interpretation without a researcher’s constraining and methodologically rigid intervention. In communication studies the text is a metaphor for communication or a wider social phenomenon. The metaphorical nature of the text enables an insight by creating tension between the research object (society, media, and members of the audience) and the discourse which

“represents” the research object. In that context, new media forms are just beginning to have an impact on the old forms of research. Nowhere is this impact shown more dramatically than in ethnographic (interpretive) research.

The interpretive analyst is always, by definition, also a social critic – a media text provides us with critical windows revealing an endless range of social, cultural and political phenomena that are continuously in flux and often the sites of contestation. Stuart Hall (in Anderson 2012:326) suggested that media texts are often a location of struggle an argument in words, images, sound and story over meaning and value. If media texts are sites of struggle, then they equally are products of, or exercises in, power. To read a media text is also to confront the play of power in shaping the meanings and assumptions that delimit the boundaries of communities, cultures, and societies.

However, such an examination should be taken with a “grain of salt”. Researchers should not direct their attention to minor contradictions and incongruities in the text, but eradicate a strict line between two ontological positions – the extreme linguistic one which requires the deconstruction of text, and the objectivistic one which closes the interpretive space.



Part II

Cross-National Comparative Research Between

Theory and Application

Since the early 1980s, when the first round of the World Values Survey (WVS) was fielded, international social survey research has seen remarkable growth.

Many other similar international social surveys ( e.g. International Social Survey Programme, European Social Survey, and Asian Barometer) have emerged, all allowing open access to data. Thus, data from such survey programmes have practically become accessible to all researchers interested in comparative analysis of such data. However, in addition to the benefits, certain dilemmas and problems arise. In the next two chapters, we discuss some of these problems by considering specific cases from the International Social Survey Programme (ISSP) (see ISSP 1; Smith 2009).

One dilemma we deal with is the question of where comparative research, based on data from standardised social surveys, should be placed in the context of the relationship between qualitative and quantitative approaches. Our standpoint is that various paradigmatic elements of the two approaches often ‘meet’ or combine within comparative research. Let us look at an example: a) data collection is always carried out through a social survey as a typical quantitative method; while b) analyses and interpretations are usually conducted at the level of comparing particular countries, where often an in-depth qualitative insight into each specific case (country) is needed. We not only speak of qualitative and quantitative methods but can also describe the comparative approach as a synthesis of qualitative and quantitative approaches (see Ragin 1987).

In the third chapter, we describe a concrete example (based on ISSP data) of comparative analysis in which elements of both qualitative and quantitative approaches are present. We deal with the classification problem in the context of cross-national comparisons and show how the deductive (explanatory) and inductive (exploratory) logic interweaves within comparative analyses based on data from international surveys. In this way, we illustrate the thesis that (international) comparative research is an approach in which the paradigmatic gap between the qualitative and quantitative approaches is bridged.

Another problem facing international social research that we also consider in this part is the question of whether the data collected in different socio-cultural contexts (usually countries) are comparable and, as such, allow credible comparative analyses. In general, we speak of data quality but, more specifically, it is a problem of equivalence from the conceptual level to the empirical (measurement) level of collected data. Thus, in the fourth chapter, when showing how the ISSP

operates, we describe the procedures followed by researchers to ensure all aspects of equivalence within the framework of cross-national comparative analysis building on social survey data.



Deductive and Inductive Logic in Developing

Typologies in Cross-National Comparisons:

3 The Case of Political Participation

3.1 The Problem1

When social scientists discuss the epistemological differences between qualitative and quantitative approaches, they usually connect induction to the qualitative approach and deduction to the quantitative approach. However, just as the paradigmatic dispute between constructivism and positivism can be overcome and so called ‘pragmatism’ can emerge, it seems quite legitimate to combine induction and deduction within the same research process (Tashakkori and Teddlie 1999; Brewer and Hunter 2006, Teddlie and Tashakkori 2009; Ragin and Amoroso 2011; Cooper et al. 2012:1–11). The process of comparative research was convincingly demonstrated by Charles Ragin who regarded a comparative method as a convergence or even synthesis of qualitative and quantitative approaches (Ragin 1987). Ragin also understands social research to be a process of ‘retroduction’ rather than a process of pure induction or pure deduction (Ragin and Amoroso 2011).

As with the case of social research in general, we can speak of at least three levels of research and analysis in comparative research: (a) the descriptive level; (b) the classification (clustering) level, which can also be understood as a more sophisticated descriptive level; and (c) the level of explanations and/or understanding which primarily includes testing causal relations or discovering causal conditions (that lead to differences between comparison cases). When conducting comparative research, social scientists usually focus on discovering or testing the differences and/or similarities between the cases being compared.

This means that the classification or clustering of units of comparison occurs in almost all comparative research. However, comparative studies may vary considerably because the classification may be carried out in various ways.

Due to the wide range of social science literature – including both, empirical comparative studies, and methodological discussions on comparative research –

comparative research may include a continuum from…



− studies of only one case comparing it with an ‘ideal type’ or a theoretical construct, to



− studies with a large number of cases (all possible cases), which are relevant for the phenomenon studied ( cf. Ragin 1987; Pennings et al. 2006:20–23; Berg-Schlosser et al. 2009).

1 An initial version of the text for this chapter was prepared for the CINEFOGO WP11

workshop (part of the 6th FP) on ‘Methodological Challenges in Cross-National Participation Research’ The Hague, 16–17 January, 2009.

60

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism However, it is common for studies of a moderate number of cases to be ‘classified’

as comparative, while those with just few cases are identified as qualitative and those with a high number of cases as quantitative ( e.g. Ragin and Amoroso 2011).

Thus, the problem of classification is not always part of the problem of comparative research. If we compare two countries in their approaches to the welfare state, our result will be quite different to the result we would have obtained had we compared all EU members, or even a number of countries worldwide. In the first case, the concept of ‘classification’ is not used. The problem of classification is different when trying to discover typologies among the countries compared, as our comparative research demonstrates.

Before we proceed, let’s first consider some conceptual and terminological clarifications. The term classification is often closely linked to the term measurement process, within which measured cases are compared to each other and classified into categories of pre-prepared measurement instruments ( cf. Lazarsfeld 1958, Ben-Baruch 1980; Pawson 1989; Marradi 1990).2 However, here we are primarily interested in understanding the classification as an analytical process, within which cases are classified according to the differences and similarities between them in terms of more measured dimensions. In this respect, we use the term ‘classification’

to cover a general concept of classification which could be further broken-down into more specific ‘forms’ of classification. The literature employs several terms to identify the concept of classification: clustering, typology, taxonomy etc. (Marradi 1990; Eppler and Mengis 2011). Some authors use different terms in order to divide the general concept of classification into specific ‘sub-concepts’ to take into consideration the following two aspects of classification: a) the characteristics of the process (operation) of classification; and b) the result of this process (see Marradi 1990).

We will discuss both the process and the result. We will use the general term

‘classification’ for the process and the term ‘typology’ (or ‘taxonomy’) for the results of multidimensional classification.

As a process, c lassification could – at the most general level – be split into the following two processes:

a) Theoretical conceptualisation or the division of a complex theoretical concept into sub-concepts and categories. Within a quantitative approach, it is also the first step towards the measurement process. This understanding of classification is primarily based on deductive logic.

2 For example, Marradi argues that ‘the choice of a unit of measurement automatically breaks down a continuum into classes with the same characteristics as the classes of any classification scheme’ (Marradi 1990:150)

Deductive and Inductive Logic in Developing Typologies 61

b) Empirical classification (clustering) of observed cases (or the dimensions of these cases). This understanding of classification is usually closer to inductive logic. Cases are not classified into pre-prepared categories, but quite the inverse: the set of categories is the result of the ‘discovery’ (‘exploratory’) process.

Our focus is on the empirical classification of observed cases. We classify individuals according to their ‘style’ of political participation. Although our focus was on the classification of cases (individuals), in order to prepare the instruments to measure political participation, we first conducted a classification of the dimensions (the characteristics) of these cases. In general, we can say that we combined deduction and induction:



− When preparing dimensions and instruments to measure political participation, which are used to classify individuals, we mostly rely on deductive logic.

Applying the theory and findings of previous research, we divided political participation into certain key dimensions, or sub-concepts.



− After preparing measurements for these key dimensions – mostly following deductive logic – we ‘switch’ to inductive logic. In order to discover a typology of political participation, we classify individuals on the basis of these dimensions of political participation. To achieve this aim, we use an exploratory method for data analysis, namely hierarchical cluster analysis.

However, to make things even more complicated, we are interested in the development of such a typology in the context of cross-national comparisons.

Thus, one additional issue which comes into play is the problem of equivalence in cross-national comparative research (see Johnson 1998, van de Vijver 1998).

There are, in our case at least, two general questions that arise from the problem of equivalence:



− Does our conceptualisation of political participation fit the situations in all compared nations?



− Do the measurement instruments we use reliably measure across all compared nations?

Even if we can answer both questions positively, doubts remain about the equivalence of the discovered typologies. These typologies are the result of a classification process which includes a combination of various dimensions. Even if we can speak of equivalence at the level of the individual dimensions, it may not necessarily be the case when these dimensions are combined in typologies.

3.2 Political Participation Typologies in the Context of Cross-National Comparison

Social science research into political participation encounters the problem of the various types or forms of political participation. Until recently, political activities have usually been classified into two typical groups: conventional and unconventional types of political participation (Newton and Montero 2007, Linssen et al. 2015). Together with social changes, such as the development of late modernity, individualisation and the emergence of new communication technologies such as the internet, new forms of citizens’ political activities are emerging in the form of political consumerism and various on-line forms of participation. At the same time, the old forms of participation are being revised.

For example, many conventional forms of participation ( e.g. communication with politicians) are moving to the internet, while some unconventional forms of participation are in fact becoming conventional ( cf. Hafner Fink and Oblak Črnič 2014). We can also ask whether an interest in politics or, for instance, working in various voluntary associations, can be classified as forms of political participation (or at least as potential political participation). In short, the classification of different forms of political behaviour into the conceptual framework of political participation is clearly not a straightforward task (van Deth 2014). To answer these ‘empirical’ challenges, and in line with social theories on the process of individualisation ( e.g. Beck and Beck-Gernsheim 2002), there is evidently a need for new typologies which can transcend the conventional–

unconventional dichotomy. On this basis, there has been a shift in the field of political participation research toward typologies involving a differentiation between institutionalised and individualised forms of political participation (Dalton 1996, 2008; Fink-Hafner and Kropivnik 2006; Deželan et al. 2007).

There have also been attempts to reveal and provide theoretical justifications for new styles of citizenship that go beyond the dichotomy of institutionalised–

individualised participation. One example is the contribution of Henrik Bang, who speaks of ‘expert citizens’ and ‘everyday makers’ (Bang 2004; Li and Marsh 2011). In developed countries, a tendency towards the individualisation of politics is most visible and is revealed in the ‘shift away from electoral decision-making based on group and/or party cues towards a more individualised and inwardly-oriented style of political choice’ (Dalton 1996:11). This style is characterised by autonomous political decisions being made by individuals on specific political issues and their activities in various non-institutionalised forms of collective behaviour. Thus, in a cross-national comparative context, we can expect different results for the classification of political activities. For example: (a) a higher probability of clustering around the conventional-unconventional

Deductive and Inductive Logic in Developing Typologies 63

dichotomy in new democracies (countries with lagged modernization); or (b) a greater probability of clustering around the institutionalised–individualised dichotomy in countries experiencing late modernity or post-modernity.

Unlike the typology of forms of political participation, our contribution focuses on a typology of citizens who practice various forms of political participation.

Our primary goal was to develop types of individuals (citizens) with similar combinations of political participation and forms of membership. We tested the possibility of achieving this goal by employing hierarchical cluster analysis in an international comparative context using data from the International Social Survey Programme (ISSP) from 2004 (ISSP 2004 – Citizenship). Within this framework, we observed various forms of political participation. In addition to the indicators of political activity, we also included indicators of involvement in political parties and social organisations, specifically a feeling of belonging to and participation in the activities of an organisation. In so doing, we were limited by the data available from the ISSP research. For our analysis we prepared the following three dimensions of political participation: 1) Political activity was observed on the basis of former involvement in eight various kinds of activity (Appendix 1) which could be classified into one of the following three groups: (a) communication (contact with politicians, with the mass media, participation in Internet forums); (b) direct participation (boycotts, participation in demonstrations, participation in a political rally); and (c) ‘support’ for political ‘projects’ (signing a petition, monetary donations).

2) ‘Membership’ of organisations was observed according to membership of four types of organisations: (a) political parties; (b) trade unions and professional associations; (c) religious organisations; and (d) societies and other voluntary organisations (details can be found in Appendix 1). For each type, we prepared separate indicators of the level of involvement, whether active or passive on a scale from 0 to 2.

3) Interest in politics was measured using a single index (on a scale of 1 to 4) composed of three individual indicators: a clearly expressed interest in politics and two indicators measuring involvement in discussion of political issues (see Appendix 1).

Our approach involves a combination of deductive and inductive logic. We began with some theoretical ideas on the possible classification of political participation based on the results of various previous research. Within this framework, our theoretical starting point was the dichotomy of institutionalised and individualised participation. Apart from this, the only fixed starting point was the idea that citizens develop different styles of political participation combining different forms of political participation and involvement in

64

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism different associations. We didn’t use the hierarchical cluster analysis to test any particular hypothesis: neither the number of different ‘types’ (clusters of similar individuals) nor their characteristics. Rather, we set about exploring these ‘types’.

However, the first problem we encountered was to select the countries to be compared. We wanted to exclude possible sources of non-equivalence common to cross-national comparisons (Scheuch 1993 [1968], Scheuch 1989, Johnson 1998, van de Vijver 1998), while others we wanted to keep. We therefore limited our analysis to those countries with what we might call a ‘Western culture’. At the same time, we pursued diversity within this framework and included both new and old democracies, European and non-European countries, countries with different party systems etc. Thus, we included the following countries: Bulgaria, Poland, Slovenia, Portugal, France, Denmark, Sweden, Austria, the USA and Venezuela. In this way, we were able to identify whether our expectations of different typologies in differing social environments were justifiable.

3.2.1 Classification of Activities

We respected the principle of parsimony, which is especially important in cross-national comparative research where there is a vast amount of data. Our first step was therefore to reduce the number of empirical variables (indicators) for the abovementioned discovery process. This need for simplification was especially important for the list of different forms of political activities. The following eight forms of political activities are included from the ISSP 2004 data set:



− v17 – signed petition



− v18 – boycotted products etc.



− v19 – took part in a demonstration



− v20 – attended a political meeting



− v21 – contacted a politician



− v22 – donated money or raised funds



− v23 – contacted the media



− v24 – join an Internet forum.

Accordingly, we first investigated whether it was possible (in the cross-national comparative framework) to reduce these eight items down to two dimensions which might fit the idea of the ‘institutionalised-individualised’ typology.

However, in this kind of analytical problem, the problem of the functional (non-)equivalence of indicators of political participation is already evident.

For example, the meaning of ‘participation in demonstrations’ may differ considerably in consolidated democracies compared to in transitional societies in Eastern Europe or Latin America. The same could be said of contact with

Deductive and Inductive Logic in Developing Typologies 65

politicians or the mass media. Such differences (non-equivalence) when using exploratory classification procedures tend towards different typologies of political participation activities in different countries which are not comparable with standardised measures. In order to compare equivalent phenomena across countries with equivalent quantitative measurements, we need to exclude ‘non-equivalent’ forms of participation from further comparative analysis. This does not mean that these ‘non-equivalent’ forms are not relevant to the comparative analysis, but rather that they were not applicable to standardised quantitative comparisons, where cases are usually classified according to the quantity of a characteristic of interest. However, the same ‘non-equivalent’ forms may be important for those comparative approaches that try to reveal qualitative differences – or qualitative characteristics of typologies.

To test how the above forms of participation fit within a two-dimensional

‘institutionalised-individualised’ typology, we applied two types of exploratory analysis: a principal component analysis and a hierarchical cluster analysis (a classification of variables). Since we are interested in the possibilities of hierarchical cluster analysis, we will only discuss the results of the cluster analysis3. We prepared the variables in binary form (0 – if the participant had not done this in the past year; 1 – if they had done this in the past year). We then repeated both analyses, in the face place, for the pooled data of all ISSP countries involved, and then separately for each of the ten selected countries: Bulgaria, Poland, Slovenia, Portugal, France, Denmark, Sweden, Austria, the USA and Venezuela. Both the principal component analysis and the cluster analysis supported our expectations (see Figure 3.1) that the results for the pooled data would reveal two dimensions or groups of activities. The first group of activities includes all three forms of communication (v21, v23, v24) in connection with two forms of ‘mass’ or ‘collective’ political activities (demonstration, political meeting or rally). The second group includes the three forms of ‘individualised’

political actions (boycott, petition, and donation).

3 The results are presented in a dendrogram (clustering tree), which shows how objects (forms of political participations) join. The procedure starts with each individual object representing a ‘group’ and develops to a point where all objects are joined into a single group. The levels of joining represent the differences between groups that are joined in a particular step. These differences are relative (standardised on a scale of 0 to 25), which means that they are not directly comparable between different dendrograms.



66

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism FIGURE 3.1 The classification of forms of political participation



(ISSP 2004, pooled data – 39 countries)



After repeating the same procedure on datasets for individual countries (principal component analysis and whierarchical cluster analysis), we decided to exclude two items which were not clearly related to a single dimension or which most frequently fit into different groups of activities in different countries. These two items are demonstrations and internet forum. We can assume that these two forms are not functionally equivalent in a cross-national comparison of political participation, especially when they take place in different systems. From the six remaining items, we formed two indexes on a scale of 0 to 3: 1) ‘Individualised’ (political) activities, consisting of the following three items:

– signing a petition;

– boycotting products;

– donating money or raising funds;

2) ‘Party’ ‘institutionalised’) political activities, consisting of the following three items:

– attending a political meeting or rally;

– contacting a politician;

– contacting the media.

We can say that both indices provide comparability between countries at least on the level of structural equivalence: each individual index consists of the

Deductive and Inductive Logic in Developing Typologies 67

same individual forms of participation, although not necessarily in the same order in all the compared countries. This means that each of them measures the same phenomenon in each of the countries included in the analysis (see van de Vijver 1998:47–48). However, the measurement units are not necessarily comparable between indexes since they do not necessarily measure at the same level of accuracy. To this extent, the measured values are not equivalent: if any two countries return the same value for an activity of political participation, this does not necessarily mean the same level of intensity or level of participation.

The result of the classification also reveals an important difference between both groups of activities: institutionalised activities appear to be more closely related to each other, while ‘distances’ between individualised activities are substantially larger. This result is in line with our understanding of conventional and institutionalised political participation as a cumulative activity. In other words, more intense participation equates to the use of a higher number of different forms of (conventional or institutionalised) political participation ( cf.

Milbarth 1965). On the other hand, the results for individualised forms accord with the label ‘individualised’. While they are closer to each other than to forms of participation in the ‘institutionalised’ group, they still sufficiently differ from one another that they may be understood as being ‘representatives’ of different individual styles of political participation. This means that, in a more detailed analysis, these individual forms should not be classified in the same category (‘individualised forms’), but rather as separate categories of individualised participation.

3.2.2 Deductive Classification of Individuals (into a Pre-Prepared Typology) Based on the abovementioned dimensions and the types of political participation, we ‘constructed’ a typology of six categories of political participation patterns, which we were able to simplify into three categories. We started with the following dichotomies: institutionalised participation versus individualised participation; political organisations versus civil society organisations. In this way, the typology of six categories was formed on the following four binary variables: (i) practising individualised political activities; (ii) practising institutionalised (party) political activities; (iii) belonging to a political party; and (iv) belonging to other organisations or voluntary associations. The final result of the procedure is the pre-prepared typology shown in Table 3.1.

68

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism TABLE 3.1

A pre-prepared typology of political participation ‘styles’

Party / Institutionalised

1.

Party belonging and other institutionalised activities Activities

2.

Party belonging combined with social (and/or

individualised) activities

Combined Activities

3.

Combined activities (individualised and institutionalised) (no Party)

– not belonging to a party

4.

Individualised and/or civil society activities Excluded / Observers

5.

No activities, only belonging (or membership)

6.

Alienated (excluded) – no activities and no membership (belonging)

When applying this pre-constructed typology to all the countries investigated, we can identify a general cross-national pattern: (a) the category of excluded is the greatest in the new democracies (Bulgaria, Poland, Slovenia and Portugal)

– above 50%; (b) the category of ‘non-party’ active citizens is the greatest in old European democracies (France, Denmark, Sweden and Austria) – above 50%; and (c) the category of party/institutionalised activities is the greatest in the USA and Venezuela (above 20%) (see Table 3.2).

A deductive approach such as this, in which the observed cases are classified into pre-prepared categories, would seem to be the simplest and clearest way of comparing countries and testing the hypotheses deducted from theory. Since the same (‘standardised’) typology was used for all countries, we compared simple frequency distributions of individuals within each individual country. In fact, we can say that this classification process represents a process of measuring individual styles of political participation. However, the assumption embedded in this procedure is that the indicators used in the preparation of the typology need to be absolutely equivalent. This is not only the case for measurement equivalence (as the technical characteristic of the measurement instrument), but is even more important for functional equivalence. For example, when we ask about activities (or work) in a political party, we have to be sure that it means the same thing – conceptually and functionally – in all national contexts. If this assumption is false, then such an approach becomes questionable. This concern may be particularly relevant when comparing the US and European countries due to the different party systems. This problem could, at least to a certain extent, be resolved by taking an inductive approach which would usually proceed from the assumption that each country compared is idiosyncratic and each country needs to be classified individually. This also means that classifications may be conducted on the basis of common dimensions for the countries compared. The

Deductive and Inductive Logic in Developing Typologies 69

final step would be an examination of the differences and similarities between the typologies identified in each country.

TABLE 3.2

Types of citizens’ political participation – 10 countries



compared (in %; ISSP 2004) Results of the classification on



the basis of the pre-constructed typology of 6 (3) categories BLG

POL

SLO

POR

FRA

DEN

SWE

AUT

USA

VEN

Type of

participation

Party /institutionalised activities

Party belonging

activities and

institutionalised

4.0

0.9

1.5

0.9

1.6

0.7

1.2

2.5

2.2

6.7

activities

Party belonging

combined with

social (and/or

1.7

0.5

4.3

3.0

3.7

6.7

8.2 16.7 37.4 15.2

individualised)

activities

Combined activities (no party)

Combined

activities (indi-

vidual and insti-

tutional) – not

1.2

2.8

5.1

2.4

12.4

9.8

9.6 12.4

9.4

10.0

belonging to a

party

Individualised

and/or civil

9.7

40.0

37.5

35.5

51.9

63.1 55.7 40.2 29.5 29.8

society activities

Excluded/observers

Only belonging

(or membership) 10.5 28.7

20.9

14.8

7.6

17.4 19.4 23.5

9.9

20.5

Alienated

(excluded) – no

activities or

72.8 27.1

30.7

43.3

22.8

2.4

5.9

4.8 11.6 17.8

belonging

3.2.3 The Empirical Clustering of Individuals – Inductive Classification In the next step, we proposed an inductive approach, but one which is combined with elements of deductive reasoning. We based our analysis of political participation patterns of theoretical assumptions about the differentiation between institutionalised and individualised forms of political participation.

70

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism However, no typology of political participation patterns (types) was prepared in advance; all were ‘developed’ or ‘discovered’ empirically, on the basis of data analysis for each country separately. We applied the procedure of hierarchical clustering (using the SPSS) within the ten abovementioned countries. We clustered individuals according to values of the following three groups of variables (7 variables): (a) two indexes of political actions (‘individualised’ political activities,

‘party’ or ‘institutionalised’ political activities) as presented in the previous section; (b) involvement in four types of organisations (political party, trade union or professional association, religious or church organisation, other voluntary organisations or associations); and (c) a composite index of interest in politics. We used Ward’s method4 and we measured the differences and similarities between the units using the squared Euclidian distance for the standardised data. Our criterion for the selection of a reasonable and meaningful number of groups was a dendrogram (clustering tree), which was cut for each country on the relatively similar level of joining of clusters (a value between 10 and 15 on a scale from 0 to 25) (for an illustration see Appendix 2). Using this criterion, we obtained various typologies with a different number of clusters in various countries. With a few exceptions, there are a smaller number of clusters in old and developed democracies: (a) there are 4 clusters in France, Denmark, Sweden, the USA, Portugal and Venezuela; (b) we obtained 5 clusters in Austria and Slovenia; and (c) there are 6 clusters in Bulgaria and 7 clusters in Poland (see Table 3.3).

It is interesting that, when distinguishing between the clusters, neither the type of activities nor any interest in politics appears to be an important determiner.

What matters, first of al , is the general level of participation (whether citizens are involved in greater or smaller numbers in both types of political participation) and the differences in their involvement in various types of organisation –

especially political parties. This is how the two following clusters characterise all the countries investigated:

1) The group of respondents who make up the largest group in most of the investigated countries is the least politically active. We conditionally named this group ‘the excluded’ or ‘observers’ (see Table 3.3). This group was characterised by a lower level of interest in politics coupled with the lowest amount of involvement in political and social organisations. This group represented the lowest proportion of the surveyed who had been involved in any kind of political activity in the past (as a rule, a score of below 2 is a low average number of various kinds of political activity).

4 Ward’s method relies on an analysis of variance approach to asses the distances between clusters. At each step of joining, it seeks to join clusters (objects that are clustered) so as to minimise the increase of the Error Sum of Squares of group means (see Ward 1963).

Deductive and Inductive Logic in Developing Typologies 71

2) The group of respondents who tended to link their political activity with party membership (belonging) we conditionally called the ‘party activists’

(or ‘party soldiers’, ‘politically active elite’) (see Table 3.3). The key difference in relation to the other groups was that this group easily had the most party members. In the majority of the investigated countries this was also the only group which included party members (significant exceptions are the USA, Venezuela and Austria). As a rule, all indicators of political participation in this group had high values (a wide range of activity, a large number of members in organisations and a deep interest in politics).

TABLE 3.3

Types of political participation – 10 countries compared



relative size of each cluster in %; ISSP 2004) Type of participation

(cluster):**

BLG

POL

SLO

POR

FRA

DEN

SWE

AUT

USA

VEN

Party (polit.) activists

4.1

7.4

9.2

4.5

8.0

6.5

Party ‘soldiers’ (members, voters) 4.8

0.7

4.5

Politically active (not only party) elite 1.3

2.0

4.1

9.0

Politically active elite (no or low

party)

3.7

8.9

Active citizens (diff. org., also party)

15.5

23.2

30.6 39.6

Civil society (org.) activists (‘mixed’)

44.2 26.0 14.9

Active citizens (trade union/mixed)

13.4

Individually active (low in org.)

12.1

Trade Union (members, some act.)

7.4

21.4 13.4

8.5

Soc. assoc. (members, some act.) 3.9

8.1

28.8

Religious org. (members, some act) 3.6

28.7 14.7

Observers (org., some act.)

58.0 39.5 55.8

39.1

Observers (relig. org., some act.)

27.0

Observers (org.not relig., some act)

23.1

Excluded – observers (traces of

activity)

56.3

Excluded – only voters

37.0

Excluded – only relig. org.

42.5

Excluded – alienated

49.4 25.6 56.7 53.8

No. of clusters

6

7

5

4

4

4

4

5

4

4

Association with ‘pre-constructed’

typology of 6 categories

0.47* 0.76* 0.59* 0.71 0.69 0.80 0.69 0.43 0.39 0.42

(Cramer’s V)

** Types are labelled empirically according to their structure with working names.

* More than 25% of the cells have an expected frequency of less than 5

Results of the hierarchical cluster analysis – the number of clusters for each country was individually established by cutting the dendrogram on the same (similar) level of the joining of clusters.

72

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism The groups described represent the two extremes from the countries investigated.

Between these two extreme groups several other groups were formed in the various countries. As a rule, only two additional groups were formed in developed countries with a longer democratic tradition. Both were characterised by relatively high values for both types of activities. Particularly high (the highest compared with the other groups) was the amount of membership of societies and other similar social organisations. At the same time, these groups rarely included party members (the USA and Austria being the most significant exceptions). In less developed countries with less experience of democracy, more groups were formed between the two extreme groups. They differed primarily according in their levels of activity and in particular in their involvement in various kinds of social organisations. As a rule, there was no involvement in political parties in these groups.

The different results in the number of clusters and composition of individual clusters raise a number of questions of comparability. Here we will discuss two questions: what are we comparing? and how do we decide the appropriate number of clusters (participation types)?

We should perhaps first ask What is being compared? Should we compare individual clusters from different countries (contexts), compare countries regarding the number of clusters obtained etc., or should such results be understood in a holistic way – the typology as a whole representing a type of political culture (or perhaps a type of political system) in a particular country? If we adopt this second option, we would need to compare combinations of all aspects of the result for each country (number of clusters, composition of each cluster individually, size of clusters etc.) in connection with the social and political context ‘presented’ by a particular country. Thus, the result can only achieve its full meaning in relation to the context. Let us briefly consider an example. Generally, we obtained similar results in all developed old democracies (Austria, Denmark, France, Sweden, and the USA). We identified the following: a small but active group of political party

‘activists’; different groups of active citizens (less connected to party politics); and groups of observers (the results were similar in new democracies).

Yet there was one important difference which distinguishes European democracies and the USA. Belonging to a political party proved to be a delineating point in European countries: usually all (or most) respondents who belonged to a political party were classified in one cluster (‘party activists’), while in the USA we can find party adherents in all clusters. The difference becomes meaningful when we take into account another difference: political party adherence in European democracies is mostly defined through membership and a membership fee, while in the USA we can speak of party adherence as a kind of social (political) identity (‘I’m a Democrat’ or ‘I’m a Republican’).

Deductive and Inductive Logic in Developing Typologies 73

An explorative approach is typical of how the hierarchical clustering method is used: the number of clusters is not fixed in advance but is the result of a thorough inspection and analysis of the results of clustering. Therefore, the second question is about the criteria for selecting the appropriate number of clusters. In general, separate clusters (groups of individuals) are obtained by cutting the clustering tree (dendrogram) at a selected level. When the classification is conducted in one country only, this is a less demanding task than in the cross-national comparative context. In the comparative context we face the following dilemma: should we cut the dendrogram at the same level in all countries? – which would mean that number of clusters can differ from country to country – or should we follow the criteria of the same number of clusters in all countries? In our case, we decided to check both. In addition to the existing criterion for fixing the number of clusters (the same or similar relative level of the joining of clusters in all countries), we additionally applied the criterion of the same number of clusters in all countries. The results presented above suggest that the most parsimonious solution might be the solution with three clusters. We thus cut the clustering tree in all investigated countries at such a level as to obtain three clusters (for an illustration see Appendix 2), presenting the following three general types of citizens’ political participation: (i) party activists; (ii) active citizens; and (iii) observers or excluded (see Table 3.4).

With this simplified solution of using three clusters in each country, we also encounter the problem of cross-national comparability (or equivalence) in terms of the structure of the whole cluster setting on the one hand, and the structure of apparently matching clusters in various countries on the other. Thus, whether we chose the solution of a different number of clusters or the solution of the same number of clusters in each country, while drawing cross-national comparisons we in fact obtained a much wider variety of participation ‘patterns’

than could be expected from the number of clusters within countries. Where we cut the clustering tree at a similar level in all countries we obtained at least 18 different types (or ‘patterns’) of participation (see Table 3.3). And where we cut the clustering tree at the level to obtain three clusters per country, we can recognise at least 13 different ‘patterns’ (types) (see Table 3.4). Despite this kind of diversification, we can still observe some regularity in the differences and similarities between the countries observed. We examined some of those which support our argument that the differences (or similarities) discovered are not just a methodological construct but probably indicate a meaningful general pattern of differences between countries. It seems there is a wider gap between the ‘party’

group and the group of ‘excluded / observers’ in new democracies. As a rule, the group of ‘excluded / observers’ is significantly larger in new democracies (see Table 3.3). The group of ‘active citizens’ is clearly recognisable and less dispersed

74

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism in developed democracies: a smaller number of clusters were formed between

‘party activists’ on the top and ‘excluded / observers’ at the bottom (see Tables 3.3 and 3.4). This may also be an indicator of stronger civil society in countries with a longer democratic tradition.

TABLE 3.4

Types of citizens’ political participation – 10 countries compared

(the relative presence of each cluster in %; ISSP

2004)

Type of participation

(cluster):*

BLG

POL

SLO

POR

FRA

DEN

SWE

AUT

USA

VEN

Party activists

6.5

4.1

7.4

9.2

4.5

6.5

Activists (party, trade union)

21.3

Politically active citizens

(also party)

4.1

39.6

Some activity – party

9.9

Some activity – different

org. (also party)

37.3

Trade union (some party

belonging)

8.1

Politically active (no party

belonging)

3.7

Active citizens (civil. soc.

activists)

37.9

53.1

35.0 45.5

Some activity – different

org. (no party)

36.8

Some activity – trade union

13.4

Some activity – relig. org.

3.6

Observers (some activity)

39.5

55.8 50.1 39.1 56.3

Excluded

86.5

88.2

56.7

82.5 58.0

Association with

‘constructed’ typology of 3 0.39 0.32

0.62

0.37 0.60

0.74

0.66 0.40 0.38 0.41

categories (Cramer’s V)

* Types are labelled empirically (according to their structure) with working names Results of cluster analysis –dendrograms were cut at the level to obtain 3 clusters in all countries.

There are also indications of a stronger association between the pre-constructed and discovered typologies in developed European democracies – namely France, Denmark and Sweden (see the values of Cramer’s V in Tables 3 and 4). We can say that the empirical situation in these countries more than others closely corresponds to the idea of the ‘institutionalised-individualised’ political participation typology which was built into our conceptualisation and operationalisation of the variables included in the analyses.

3.3 Summary: The Complexity of the Classification Process in Cross-National Comparative Research

The use of classification (from a general research approach to a specific method of data analysis) within cross-national research design entails various general methodological and epistemological problems. Based on the case presented, we can identify the following problems:



− mixing qualitative and quantitative approaches;



− combining deductive and inductive reasoning;



− problems of equivalence (comparability);



− the benefits of using a particular classification method (hierarchical clustering);



− the placement of classification methods within the framework of the induction-deduction dichotomy.

How is a combination of the quantitative and qualitative approach (including a combination or synthesis of deductive and inductive reasoning) present in the comparative approach? What does our example of political participation typologies tell us about this?

The whole research problem is based on the ISSP data collected by the typical quantitative method of highly standardised survey interviews. We used deduction to develop the measurement instrument (survey questionnaire) for this survey, which was a result of the operationalisation of theoretical concepts.

Our starting idea of political participation typologies was also developed on the basis of theoretical discussions and based on the results of previous research (again, we mostly relied on deduction). We selected countries for comparative analysis in the same way. However, our main goal was not to test theoretical assumptions but rather to discover country specific typologies of political participation patterns. Our method of data analysis (hierarchical cluster analysis) is typically quantitative, but also explorative: typologies are discovered and described ‘post festum’ on the basis of the results of the data analysis and not on the basis of theoretical conceptualisation. This means that this method also contains elements of inductive reasoning.

The traditional understanding of what it means to mix qualitative and quantitative approaches – whether in general or specifically in comparative research designs

– entails a reversed sequence of approach: the qualitative approach is useful at the beginning, in the explorative stage of the research process, while the quantitative approach is used to answer the main research questions or to confirm the main hypothesis (see Al ardt 1990:189). However, given the proliferation of

76

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism international social surveys,5 which creates a great quantity of quantitative data, the opposite approach seems to be more fruitful for international comparative studies. Quantitative methods are used to explore the big data, while qualitative studies, which researchers can use to ‘dig deeply and thoroughly into the subject matter’ ( ibid.:189), bring additional relevant knowledge and clarifications.

In our comparative research, we have shown how deductive and inductive reasoning are applied while analysing the same quantitative data through different quantitative methods. When classifying individuals into pre-prepared categories of political participation types we started with a typically quantitative approach

– the quantitative method and deductive reasoning. Then we proceeded with the qualitative approach: we continued to use the same quantitative data, but by applying hierarchical cluster analysis, which is still a quantitative method, our approach became rather explorative and inductive. The most meaningful next step in this research process, in accordance with Al ard’s (1989) recommendations, would be a more thorough qualitative study of specific cases (nations) to deepen our understanding of the typologies discovered and the differences between nations.

We will not take this step here, but merely suggest this as a possible future direction.

Both, the results presented and the experience of employing hierarchical cluster analysis also reveal certain strengths and problems when applied the cluster analysis to a cross-national comparative research of political participation patterns.

Talking of strengths, we can identify the following:



− When we use cluster analysis in the context of cross-national research, we can identify idiosyncratic national patterns (combinations, distributions, relationship) in the phenomenon investigated. In our case we observed that the pre-constructed classification scheme only allowed us to compare the relative size of each pre-constructed category in the various countries, while the use of a hierarchical cluster analysis revealed country-specific political participation typologies with specific political participation types (clusters, or categories).



− This insight offers an opportunity to link the discovered patterns to country-specific socio-political and cultural contexts.



− Hierarchical cluster analysis is a good explorative tool where there is a large number of classification dimensions ( e.g. forms of political participation), as this offers an optimal way of searching for general patterns and, in the cross-national context, also of identifying national idiosyncrasies.

5 For example, the ISSP is conducted annually and it started with the first round in 1985 in six countries and now involves approximately 50 countries. Apart from the ISSP, certain other large international surveys also provide data for various (quantitative) comparative analyses.

Three which are conducted regularly are: the European Social Survey (ESS), the World Values Survey (WVS), and the European Values Study (EVS).

Deductive and Inductive Logic in Developing Typologies 77



− It is also more ‘robust’ than some other procedures such as factor analysis or discriminant analysis and it is less sensitive to the measurement quality of the variables used.

Among the open problems we encountered, we should mention the following:



− When using hierarchical cluster analysis in the context of cross-national research, fixing the appropriate number of clusters for each country involved becomes difficult. There is no strict general rule as to how we may determine this number, but it depends heavily on the researcher’s theoretical background, his/her knowledge of the phenomena investigated, and his/her familiarity with the comparative cases. Of course, it is important that the criteria are transparent and common to all cases in the study.



− Due to the explorative (inductive) nature of hierarchical cluster analysis to identify country-specific patterns, the task of defining ‘units’ of comparison is much more complex. We can compare individual clusters, country-specific typologies (in our case, patterns of participation), and relevant national contexts. Moreover, it should be noted that all three possible ‘units’ of comparison are interlinked.



− There are certain other problems of a general nature with regard to cross-national comparative analysis. For instance: (a) the adequacy of the operationalisation of theoretical concepts (political participation and involvement) in cross-national surveys, which are used within the selected classification procedure; and (b) the national samples are usually too smal , which means that certain specific and relevant empirical manifestations of the phenomenon investigated might not be properly covered by the sample due to a low frequency within the population.

Specifically, we should mention the problem of equivalence, which is one of the key issues of cross-national comparative research, and especially within international cross-national social surveys. When testing theoretical assumptions about the phenomena investigated, the problem of equivalence becomes even more important. In our case, we briefly discussed some of the common aspects of equivalence: conceptual and functional equivalence ( e.g. the meaning of participation in demonstrations, and of party membership) and measurement equivalence – the reliability of measurements in a cross-national context –

specifically, whether indexes of participation are equally accurate in all countries.

In our political participation typology, we demonstrated that in the context of cross-national comparative research these problems are not trivial. To illustrate the point, if we are not sure that all mentioned types of equivalence are provided, then we also cannot be sure that the concepts in our research question or in our hypothesis really refer to the same phenomena in all investigated countries.

78

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism In our case, where we used hierarchical cluster analysis to discover types of citizens with the same ‘style’ of political participation, the question of how to compare the results of clustering in different countries can be also understood as a question of equivalence. We can ask whether the discovered typologies are equivalent across nations to the extent that they can be compared with standardised quantitative methods ( e.g. measurement equivalence), or whether the equivalence is ‘only’ structural, in other words, whether we should apply a more interpretative – qualitative – approach to the comparison.

To conclude, let us summarise the epistemological role of classification in the scientific process of building knowledge. We have shown both that classification may be carried out as a process of deductive reasoning – a process of conceptualisation – and also as an inductive process – an explorative ‘revealing’

of typologies. Although we can apply classification as a deductive process, we can reasonably question whether it is also possible to use a classification method, such as hierarchical cluster analysis in the classical process of scientific explanation, in order to verify (support or falsify) a scientific claim or a theoretical hypothesis.

Our case offers some arguments against this option. Namely, the approach and especially the method per se strongly rely on an exploratory logic. Although the selection of criteria for clustering (dimensions of cases to be classified) is theoretically based, the clustering procedure is typically explorative, which is further accentuated in the framework of cross-national comparative studies. In theoretical discussions, we can also reject the possibility that classification may be used as an analytical tool to verify scientific claims (see Marradi 1990).

Does this mean that the clustering method should be understood only as an inductive method, and may not be used as a tool for testing theoretically based hypotheses? If we are pragmatic, we can say that certain hypotheses could be tested by some kind of clustering method. These are hypotheses concerning…



− the number of dimensions of a phenomenon,



− the number of types (clusters) of similar individuals,



− the structure of these clusters,



− the differences and/or similarities of typologies between countries.

We can expect there to be a conceptual (theoretical) framework for these kinds of hypotheses. However, it is more common that such hypotheses are not directly deducted from general theoretical claims but are rather ‘moderated’ by previous data analysis or exploration.

Deductive and Inductive Logic in Developing Typologies 79

Appendix 1

Indicators for Political Participation (ISSP 2004 – Citizenship): Action, Belonging, and Interest

1. Political (and Social) Action

Here are some different forms of political and social action that people can take.

For each one, please indicate:



→ 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 you have not done it and would never, under any circumstances, do it.

Have done Have done it Have not

Have not done Can’t

it in the

in the more

done it but it and would

choose

past year

distant past

might do it

never do it

a) communication

V21 Contacted, or

attempted to

contact, a politician

1

2

3

4

8

or a civil servant to

express your views

V23 Contacted or

appeared in the

media to express

1

2

3

4

8

your views

V24 Joined an internet

political forum or a

1

2

3

4

8

discussion group

b) participation in ‘actions’

V18 Boycotted, or

deliberately bought

certain products

for political, ethical

1

2

3

4

8

or environmental

reasons

V19 Took part in a

demonstration

1

2

3

4

8

V20 Attended a political

meeting or rally

1

2

3

4

8

c) ‘Support’

V17 Signed a petition

1

2

3

4

8

V22 Donated money or

raised funds for a

social or political

1

2

3

4

8

activity

80

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism 2. Belonging to Organisations (‘Membership’ and Activities) People sometimes belong to different kinds of groups or associations. For each type of group, please indicate whether you:

→ belong and actively participate,

→ belong but don’t actively participate,

→ used to belong but do not any more,

→ or have never belonged to it.

Belong and Belong

Used to Never

Can’t

actively

but don’t

belong

belonged

choose

participate

participate

V25

a political party

1

2

3

4

8

V26

trade union, business, or

professional association

1

2

3

4

8

V27

a church or other

religious organisation

1

2

3

4

8

V28

a sports, leisure or

cultural group

1

2

3

4

8

V29

another voluntary

association

1

2

3

4

8

3. Interest in Politics (One Composed Index)

V42

How interested would you say you personally are in politics?

Very interested

1

Fairly interested

2

Not very interested

3

Not at all interested

4

Undecided

8

V47

When you get together with your friends, relatives or fellow workers, how often do you discuss politics?

Often

1

Sometimes

2

Rarely

3

Never

4

Undecided

8

Deductive and Inductive Logic in Developing Typologies 81

V48

When you hold a strong opinion about politics, how often do you try to persuade your friends, relatives or fellow workers to share your views?

Often

1

Sometimes

2

Rarely

3

Never

4

Undecided

8

Appendix 2

The results of hierarchical cluster analysis in two countries (showing how classification trees were cut) – clustering individuals on the basis of two types of political participation: involvement in four organisations and an interest in politics (ISSP 2004 – Citizenship)



82

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism France



Deductive and Inductive Logic in Developing Typologies 83

Poland



Selected Problems of International

Comparative Social Survey Research:

4 The Case of the ISSP

In today’s globalised world the social sciences have become supra-national in nature. However, the desire to discover universal social laws that would enable the prediction of (global) social processes was already present in the minds of the founders of modern sociology ( e.g. Comte, Durkheim). It is this desire that serves to foster international connections among social scientists when conducting international (cross-cultural, cross-national) social surveys. As such, the social sciences seem to have become closer in their research practices and epistemological principles to the natural sciences. However, the problems emerging with regard to cross-national or cross-cultural research support the thesis about essential methodological and general epistemological differences between the research of nature and the research of society, as also seen in some of the paradigmatic conflicts that marked the development of social-science research in the 20th century. The complexity and fluidity of different cultural contexts (social, economic, political, etc.) that importantly define (seemingly homologous) social phenomena and processes can cause problems when seeking to carry out international social surveys. Wide diversities still exist among countries, despite the disappearance of borders, cross-cultural intertwinement and the constant global exchange of goods and ideas. On one hand diversities are a “blessing” for social scientists, who believe that international comparative research can reveal and explain these and so in turn help discover certain general laws of social processes. On the other hand, these diversities seem like a “curse”

when researchers try to ensure adequate quality and comparability (equivalence) of data within international comparative surveys ( e.g. Jowel et al. 2007).

When speaking of comparative research, social scientists can today refer to a very diverse set of research practices, such as an “in-depth” (qualitative) comparison of two specific cases of litigation in the European Court of Human Rights; a comprehensive comparison of the election campaigns of two candidates running for the president of the United States; the classification of European countries according to the characteristics of their welfare systems; or an international opinion survey. Indeed, such a wide range of research practices also points to the difficulty of delimiting a comparative study of society from a more general one. Therefore, debates about the problems of comparative sociological (and general social scientific) research tend to open up the question of whether the notion of “comparative” social research may itself be superfluous, with sociology as such being considered a comparative science (see for example Nowak 1989; Øyen 1990; Arts & Hallman 1999). We would claim that such views are a mere continuation of Durkheim’s, when he pointed out that “the comparative method is the sole one suitable for sociology”) (Durkheim 1982:147), and noted that

“comparative sociology is not a special branch of sociology; it is sociology itself ”

( ibid. 157).

88

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism In this chapter the discussion will be limited to the above mentioned practices of comparative research, i.e. to international (cross-cultural) social survey research.

The relevance of this kind of research in the study of social laws and testing of (universal) social theories will be shown in the case of the International Social Survey Programme (ISSP),1 and at the same time we will point to some of the key aspects of cross-cultural (cross-national) comparative research that are important in this context. We will consider the problem of equivalence as a particularly important aspect of quality in the context of international (cross-cultural) comparative surveys. Moreover, this article will touch upon the problem of the “identification” of the unit of analysis or unit of comparison that is especially relevant in international social survey research, involving first, the collection of data about individuals (through interviewing), followed by a comparison of the countries in which these data were collected.

4.1 ISSP as a Global Social Infrastructural Programme From the very beginning of its operation the ISSP2 (the first survey of which was conducted in 1985) was developed with the goal of creating a widely accessible social survey database that could be used for secondary analysis by researchers without the resources to conduct their own surveys, and enabling both international comparisons and monitoring of changes over time. Over the years, the ISSP’s development and working principles have been grounded on the following four basic (founding) principles: 1. to develop jointly topical modules dealing with important areas of social science;

2. to field the modules as a fifteen-minute supplement to the regular national surveys (or a special survey if necessary);

3. to include an extensive common core of background variables; 4. to make the data available to the entire social-science community as soon as possible. (Smith 2009:3)

In 1985 the first ISSP survey was carried out in six countries (Australia, Austria, Italy, Great Britain, Western Germany and the USA), and perhaps its founders did not imagine that in thirty years a real global “movement” would develop 1 Besides the ISSP there are several other similar (academic) international social survey programmes, such as the European Social Survey (ESS), European Values Study (EVS) and World Values Survey (WVS).

2 For more see the ISSP website: http://www.issp.org/.

Selected Problems of International Comparative Social Survey Research 89

from such modest beginnings, one that now connects research institutions from almost 50 countries on all the continents.3 By insisting on the continued annual conduct of the survey, and maintaining a keen eye on the methodological adequacy of its practices and the quality of data collected, the ISSP programme has also provided “inspiration” for the foundation of the elite European Social Survey – ESS. The ISSP has undoubtedly played a central role in the development of modern international comparative social research, as shown in both the existence of a rich, publicly accessible database collected over the 30

years of the programme’s operation, as well as in the wealth of methodological and organisational experiences thus obtained, which provide an example for other international social surveys.

The ISSP’s measurement instrument (the module of survey questions) basically consists of two parts. The substantial part that is prepared annually, usually with 60 indicators. The instrument also consists of a block of over twenty demographic and “contextual” questions (“background variables”) that is replicated annually (gender, age and marital status, household structure, education, professional status, income, and political orientation). By 2016, eleven different topic modules had been developed within ISSP that are relevant from both aspects, that of social science disciplines (mainly sociology and political science) as well as social practices or current social processes. Most modules have been replicated at least once. Data for all modules are (freely) accessible for academic purposes through a database (ZACAT) held by the German institute Gesis – Leibniz Institute for the Social Sciences (at the website http://zacat.gesis.org/webview).

Researchers thus have the opportunity to run international comparisons in individual time points (cross-sectional study) or cross-time comparisons (for individual countries or a group of countries) within individual topical modules.

So far, the following modules have been developed (with the years in the brackets being those in which the modules were carried out or for which they are planned):



− Role of Government (1985, 1990, 1996, 2006, 2016),



− Social Networks (1986, 2001, 2017),

3 A total of 53 countries have been included in ISSP over the 30 years of its operation. By 2016 the ISSP had come to consist of research organisations coming from 46 member countries, including: Australia, Austria, Belgium, Bolivia, Bulgaria, Chile, China, Croatia, Czech Republic, Denmark, Estonia, Finland, France, Georgia, Germany, Great Britain, Hungary, Iceland, India, Ireland, Israel, Japan, Latvia, Lithuania, Mexico, the Netherlands, New Zealand, Norway, Philippines, Poland, Portugal, Romania, Russia, Slovakia, Slovenia, South Africa, South Korea, Spain, Suriname, Sweden, Switzerland, Taiwan, Thailand, Turkey, the USA, and Venezuela (ISSP 1).

90

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism



− Social Inequalities (1987, 1992, 1999, 2009, 2019),



− Family and Changing Gender Roles (1988, 1994, 2002, 2012),



− Work Orientations (1989, 1997, 2005, 2015),



− Religion (1991, 1998, 2008, 2018),



− Environment (1993, 2000, 2010),



− National Identity (1995, 2003, 2013),



− Citizenship (2004, 2014),



− Leisure Time and Sports (2007),



− Health and Health Care (2011).

Besides the replication of individual modules that enable cross-time comparisons, particularly important for ISSP is the enlargement of its membership and in turn the ISSP database, which has led to the thematic enrichment of the empirical material and increased the possibilities for comparative analyses, which now reach beyond the developed world and so allow more global views of important social (and social scientific) topics to emerge. While during the first period of ISSP’s operation the validity of some of the social science laws that were observed could only be verified through comparing developed countries with a long democratic tradition, from the very beginning of the 1990s, with the accession of a large number of countries of the former Eastern Bloc, this possibility extended to these once socialist countries that were on the path towards democracy and faster social development. Moreover, this enlargement process did not stop there, as the project was then joined by various Latin American countries (Mexico, Venezuela, Chile, Argentina and Uruguay), and more recently China and India have also become members of the ISSP.

4.2 Some Aspects of Equivalence in International Social Science Surveys – the Case of ISSP

Of course, the social and cultural diversity of the countries participating in international social surveys, like ISSP, also raises certain issues. Problems of equivalence in the operationalisation of key concepts, which are faced by researchers in preparing the survey questionnaire (as a research instrument) and running their national surveys, always present the question of whether the final results, stored at the end in the shared database, are really comparable ( e.g. Teune 1990). More specifically: Are we really comparing the same type of phenomena?

Are the repeated measurements actually equivalent? Are the measurements equally and sufficiently precise in all social contexts, thus enabling them to serve as a basis for adequate classification of the comparative units (countries)?

This problem is particularly acute in international research that surveys social

Selected Problems of International Comparative Social Survey Research 91

attitudes, such as the ISSP. The fact is that differences between countries regarding an attitude (for example, differences in the degree of perceived corruption in public services, as can be seen in Figure 4.1), as identified on the basis of the survey data, do not express neither absolute values of the observed phenomenon nor the relative relations between the units of comparison (countries), but at best the ranges (sequences) of the units of comparison regarding the analysed phenomenon.

FIGURE 4.1 Perception of corruption in public services*) 10

0,

10

20

30

40

50

60

70

80

90

0,

0

,0

,0

,0

,0

,0

,0

,0

,0

,0

0

*) (the sum total of three answers to the question about how many public servants are involved in such practices: a moderate number is involved, a lot of people are involved, almost everyone is involved; in %) (ISSP 2014 – Citizenship II) Source of data: ISSP 2014 – Citizenship II (ISSP research group, 2016)

92

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism Of course, even this can be true only in cases when phenomena of the same kind were actually measured in all the units of comparison (countries), and when approximately the same level of precision ( conceptual and measurement equivalence) was reached in all countries. In the context of international comparative surveys, a measurement is equivalent when validity is ensured within each country (system), while reliability is provided at the cross-national (cross-system) level ( e.g. Przeworski & Teune 1970 [1985]:107–108).

In the framework of international comparative survey research, the problem of equivalence is related with practically all the dimensions of the research process.

More generally, it involves the degree of control over all potential factors that could influence the quality of the collected data. Problems with the quality and equivalence of the data obtained from international surveys, such as ISSP, occur (and are solved) at least at three levels:

1. The preparation of the survey at the level of international co-ordination, that is within the ISSP working bodies such as: the secretariat, Methodology Committee with sub-committees, drafting groups for each topic module, and annual General Meeting;

2. research practice and work at the level of national research groups, which involve the preparation and conduct of the national survey, and include procedures such as translation of the original questionnaire, sampling, and fielding;

3. in both cases, within the preparation of the survey at the ISSP co-ordination level, as well as the level of national research groups, there is a third “level”

involving the problem of values and culture at the level of the population of the participating countries, which impacts the conduct of the survey through the dominant norms (culture) of interpersonal communication (that also form part of the survey situation) in the society (country) which is being surveyed. On one hand, this is implied in the “protocol” of the entire survey and the proceedings of each interview, and on the other in the (expected) reactions and participation of the respondents.

The key topical and methodological (technical) problems (factors), that influence data comparability, and which are confronted by researchers when working to achieve equivalence within international surveys, such as ISSP, include the following:



− selection of the research problem and conceptualisation;



− operationalisation of the theoretical concepts in the form of a common survey questionnaire;



− translation into the languages of the participating countries;

Selected Problems of International Comparative Social Survey Research 93



− harmonisation of the sample design and the methods of fielding the survey;



− the problem of archiving the acquired data.

The provision of comparability (equivalence) starts with the selection of the research problem and the conceptual (theoretical) design of the research (definition of the key concepts) – that is, in the conceptualisation phase. When a large number of countries that are diverse in almost every respect (culturally, economically, politically, and so on) col aborates in this phase as is the case with the ISSP, there are two questions to be raised at the very selection of the main research topic (module) and in further selection of concrete themes in this framework: 1) whether the themes selected within the module are relevant for all the participating countries – whether the researched phenomenon actually exists (and is measurable) in all the countries; and 2) whether enough common points exist in terms of the general and concrete research topics in the selected module across all the participating countries, to enable research based on standardised data collection within the social survey. In short, so-called conceptual equivalence should be provided, which allows us to say that the same kinds of phenomena, or phenomena that allow meaningful discussion in all the participating countries (cultures), are researched in the survey ( e.g. Johnson 1998:6–7). If we go back to the example of the survey question about the range of corruption in public services (Figure 4.1), then at least the following issues arise with regard to conceptual equivalence: a) Did the respondents in all the countries understand corruption in the same way – did they understand the same phenomena as corrupt? b) Can the question asked really be understood as a completely unambiguous measurement instrument which assesses the degree of corruption in a country, or does it perhaps also measure the population’s

“sensibility” to corruption? c) Did the respondents in all countries understand the answers offered in the form of an ordinal scale in the same way, and did they also react equally to them? Moreover, the second and third questions also relate to the problem of operationalisation and equivalence of the indicators.

When conducting an international comparative survey, the definition of the research problem is followed by the operationalisation of theoretical concepts – the design of the measuring instruments in the form of questions in the survey questionnaire.

In the context of international surveys, besides the usual requirements to ensure validity and reliability of the designed measuring instruments, the international comparability (equivalence) of the data acquired through these instruments should also be provided. Therefore, the survey questions in different societal (cultural, economic, political) circumstances of all the participating countries need to guarantee that the same (expected) phenomena or theoretical variables are really measured at a comparable level of precision. Therefore, measurement

94

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism equivalence or the equivalence of indicators ( e.g. Scheuch 185–188, van de Vijver 1998:1968:47–48) should be ensured beside the aforementioned conceptual equivalence. Each individual question (indicator) should have the same meaning for all the respondents, and at the same time should enable measurement at the same level of precision in all the participating countries (societies, cultures, etc.).

Although ISSP’s operation does not have any guaranteed central funding, its working principles and practice allow it to deal relatively efficiently with problems regarding various aspects of equivalence in the phase of substantive conceptualisation and operationalisation of each module. This is due to its democratic principle of operation, which allows each participating country (its researchers) equal weight (one voice) in the selection of the module and concrete sets of topics within the selected module. The rules (and practice) of ISSP’s operation guarantees that the design of individual modules is not simply left to an individual country or a research group from a country. The selection and preparation of the thematic module take place at several levels and in several phases. First, a narrow group of researchers (from one or several countries) prepares a theoretically grounded proposal, followed by a “public” debate of all the ISSP members that adopt the module at the annual general meeting. A special working group (drafting group) is then formed consisting of researchers from no fewer than three (and no more than six) ISSP member countries (usually five or six, in practice) (ISSP 2). According to the topic principle being considered, the drafting group normally consists of researchers from countries with different cultural or socio-historical contexts. The group then makes the first draft of the module – preparing a kind of analytical model that connects different sets of topics. Priorities for concrete topical sets of the selected module are defined at the following annual meeting of the ISSP members (based on debates, remarks, proposals, and so on). On this basis the drafting group prepares its proposal of the final version of the draft questionnaire (in British English), drawing on the results of the pilot tests of the questionnaire in different countries. In this phase the practice of instrument design also follows the working principles according to which all ISSP members have to be informed about the working process: for example, no later than two months before the final adoption of the last version at the ISSP annual general meeting, all member countries need to receive the draft questionnaire to formulate their remarks and be able to participate in the discussion at the meeting (ISSP 2; Smith 2009:22–23). As such, in practice researchers from all the ISSP member countries participate in instrument development. This method of work can be said to guarantee a high degree of conceptual equivalence of the selected indicators. Although a standardised measuring instrument for culturally and developmentally very diverse environments is designed in this process, the appropriate level of

Selected Problems of International Comparative Social Survey Research 95

measurement equivalence (comparability) can be achieved due to the way in which the instrument is designed.

International social survey research has two general strategies to prepare a measuring instrument (the questionnaire): a) one is foregrounding the comparability of the concepts (concept equivalence), that thus do not need to be measured with identical questions; b) or alternatively, providing international comparability through the use of the same questions that are adequately translated (see Harkness 2007:80–81). Most major international surveys (including ESS, WVS and EVS, besides ISSP) use the second strategy, that is, they use the same questions that have been adequately translated. They generally use the “sequential” approach, where the “original” (primary) version of the survey questionnaire is first prepared, which researchers from the participating countries use as the source for translation into the language(s) used for fielding the instrument in their respective countries ( ibid.).

As with many other international survey studies, with ISSP the original version of the questionnaire is prepared in British English. This is why when carrying out the translation of the (usually English) original version of the questionnaire into the languages of the participating countries it is important to provide comparability (equivalence) of the measurements or data. However, even more important than linguistic adequacy of translations is their so called functional equivalence. This means that rigid insistence on the precise or literal linguistic translation (“lexical equivalence”) can do more harm than good. The key point here is that what is not to be lost in translation is the meaning of each question, as was defined within the conceptualisation and operationalisation. This is why in the ISSP’s established practice the original version of the questionnaire includes “translator’s notes” at all the points at which potential misunderstandings in translation could occur, and which call attention to the meanings that might be “hidden” behind certain expressions.

Of course, in the end the translation and in turn the quality of the measuring instrument depend on the translation practice within each country’s research group, as mistakes in translation can completely change the meaning of a question.

It is not unusual that a translation causes problems with regard to measurement reliability, and thus measurement equivalence. For example, problems often occur in translating the categories of the ordinal measurement scale expressing different degrees of intensity of the measured phenomenon. Let us look at an example of a seemingly very simple question used in all the replications of the ISSP module the Role of Government (1985, 1990, 1996, 2006 and 2016).

The respondents were asked about the degree of their interest in politics. They were offered five possible answers, classified on the ordinal scale from the highest degree of interest to no interest at all (Table 4.1).

96

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism TABLE 4.1

The translation of individual categories on the ordinal scale



of answers to the question about interest in politics (an example from the ISSP research – Role of Government) English original

Direction of

Slovenian translation

translation

How interested would you

Kaj bi vi rekli, v kolikšni meri vas

say you personally are in

osebno zanima politika?

politics?

1 – Very interested

→

1 – zelo me zanima

2 – Fairly interested

→

2 – kar precej me zanima

(2 – Quite interested)

←

3 – Somewhat interested

→

3 – nekoliko me zanima

4 – Not very interested

→

4 – le malo me zanima

(4 – Hardly interested)

←

5 – Not at all interested

→

5 – čisto nič me zanima

The question that emerges in the English original itself is whether different degrees of intensity of the measured phenomenon are discernible enough. For example, do categories 3 and 4 (“somewhat interested” and “not very interested”) really differ enough from one another to be understood as such (in the same order) by all the respondents of the same country (Table 4.1)? In the next step this raises the question of whether the translation can preserve the sequence, the distinctive abilities of the entire scale and the “intensity” of the phenomenon defined by individual categories, thereby ensuring comparability of data between the countries. Therefore, if in the English original, category 3 (“somewhat interested”) means a higher degree of interest in politics than category 4 (“not very interested”), then in the Slovenian translation, category 3 (“nekoliko me zanima”) should also unambiguously signify a higher degree of interest than category 4 (“le malo me zanima”) (Table 4.1). These problems are also mirrored when reverse translation is attempted and the results are not necessary identical to the original; for example, the Slovenian category 2 – kar precej me zanima, could also be translated as 2 – quite interested; or 4 – le malo me zanima could be translated as 4 – hardly interested (see Table 4.1).

With the researchers from all ISSP member states participating in the design of the measuring instrument (the questionnaire) for each ISSP module, numerous potential problems in translation can be prevented with the design of the original (English) version of the questionnaire. Even before the final “confirmation” of the formulation of individual questions, the researchers from all the participating countries meet at a joint meeting of researchers, where they can call attention to any translation problems, which in turn can lead to the improvement of the original (English) version.

Selected Problems of International Comparative Social Survey Research 97

The problem of equivalence in international survey research also relates to the problem of sampling. The question is: are the sample designs used in all the participating countries such that they provide both “representativeness” in each country, and international comparability of data. An adequate (high enough) and comparable response rate (the planned sample realisation) should be provided in addition to the adequacy and comparability of the sample design. Apart from the usual problems that can arise in survey research, today, social scientists aiming to carry out opinion survey research also face the growing problem of persuading people to take part in such projects in the first place. The response rate in public opinion telephone surveys is usually under 20 per cent. The top quality international social surveys, such as the European Social Survey (ESS), which is grounded in the ISSP tradition, invest enormous efforts in achieving the minimum target response rate of 70 per cent in the field survey through direct interviewing (ESS 2; Koch et al. 2014). It is in international surveys that the problems related to refusals to participate can have even more serious consequences, as comparability of data between the participating countries is not only expected, but required. We are thus justified in asking whether data comparability and equivalence can be provided at the international level, if the response rates differ significantly from country to country. Also the ISSP

researchers face substantial difficulties related to the provision of adequate response rate: for example, in the data for the ISSP 2007 survey – Leisure Time and Sports, the response rates among the countries range between 20 and 90

per cent (Scholz & Heller 2009:11).4 High refusal rates may lead to biases in results that can have different “orientations” from country to country due to different sampling and fielding practices. This makes data comparability between the countries become even more questionable. While ISSP has always aimed to standardise these procedures, its efforts are limited by the fact that ISSP, unlike the ESS, has no financial means for its co-ordination and the completion of so-called common tasks. However, despite these problems ISSP

has always managed to assert some common rules regarding sampling: only probability (random) sampling is acceptable (after several years of debate, the practice of quota sampling in individual countries was prohibited), the use of supplementary respondents has been severely limited (in an attempt to eliminate it), and sample realisation data are collected and published, in compliance with international standards.

4 Difficulties in achieving the target response rate (70%) occur even in the better funded ESS

survey, where, for example, the response rates for the fourth round of the survey (for 2008) range between 31.4 % and 74.1 %, with only one country (of 21) managing to achieve the target of 70% (source: ESS 1).

98

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism Sample realisation also depends on the mode (form) of data collection. (see for example Billiet et al. 2007:121). Direct “face-to-face” interviewing is understood as the mode giving the best results, and is therefore included also in the ESS

standards ( ibid. 115). However, recently ESS has also been looking for some compromise solutions in the form of combined research methods, or web panels (ESS 3). Apart from face-to-face interviewing, the participating ISSP countries also use the so called self-completion format, where the respondents fill-in a questionnaire left with them by the interviewer or that they receive through the regular mail service. Some countries also use a combination of face-to-face interviewing and self-completion, especially in cases when the ISSP survey is carried out with another national survey. While telephone surveying is not seen as acceptable, it is sometimes used if there is no other way to achieve an adequate response rate.5 The existence of varied practices regarding the modes of interviewing used in different countries can of course cause problems regarding comparability, which is why within the ISSP’s Methodology Committee a special sub-group was created to deal with the mode-effects which has recently been dealing with the problems caused by combining different modes of data collection. As such, the ISSP Citizenship module for 2014 was carried out by some countries using the combined survey method, namely, with some of the respondents completing the questionnaire (as was usual before) sent to them through the mail, while others took part via the on-line questionnaire (Joy and Sapin 2016). These efforts are not only related to verifying the equivalence of the data acquired through different modes of interviewing, but also aim to improve the response rate.

The social circumstances in which the surveys in individual countries were carried out represent an important factor that also influences the comparability (equivalence) of the data in international social-science surveys. However, we must distinguish the general social (cultural, political, historical, and so on) context from concrete events that happened just before the beginning of the survey and during it. Indeed, researchers expect different social contexts to exist within international comparative (survey) studies, or often even consciously choose the kind of comparative units (countries, societies, systems etc.) that are as diverse as possible. This is especially important when their research aim is to test the universal character of a scientific claim in different social circumstances. Of course, these differences need to be considered when 5 For example in the ISSP 2007 – Leisure and Sports survey in the USA, 19 % of people were interviewed through telephone surveying. This information is accessible in the form of a special variable containing information about the mode of data collection, and this is a standard part of the ISSP data base (see Scholz & Heller 2009:10).

Selected Problems of International Comparative Social Survey Research 99

designing the measuring instrument. More serious problems can be created by the circumstances that occur due to concrete (unexpected) events just before and during interviewing. Of course, these events cannot be considered in the design of the survey questionnaire, but we do need to be aware of them and consider them as a possibility in such a way that the maximum quantity of data is collected about the conduct of the survey itself. It is important for the researchers who will later use data from international social surveys for secondary analysis to have adequate information about when the data was collected in the individual participating countries. This enables them to examine for each country that they included in their analysis, whether during this time there were any important events that could have influenced the respondents’ answers.6 Namely, rather than measuring current public opinion as a response to concrete current events in individual countries, the aim of international social surveys, such as ISSP (or ESS, EVS, and so on), and the main interest of the researchers, are the differences (and similarities) between countries regarding more general (and stable) value orientations, as well as their citizens’ attitudes or actions, that is, regarding those characteristics that can be placed in the framework of different social theories.

With offering the data collected to the whole community of social researchers for secondary analysis being ISSP’s (and similar projects’) basic aim, it is particularly important that these data are adequately archived. This means that apart from the survey data researchers also have access to metadata, or “data about data”. Within ISSP, the procedures for archiving the data collected take place according to a set of pre-determined rules. ISSP’s “official” archive is the German Data Archive from Cologne7, whose representative regularly participates in ISSP activities.

The ISSP working principles also define data archiving procedures. As the last research step, these enable an adequate level of quality and comparability (equivalence) of data collected in the varied conditions of the participating countries, before the data are given to researchers for secondary analysis. It is especially important that data from national surveys that were collected in different ways and in different languages are finally stored in a standardised form that enables the formation of a common database. Moreover, information about the specificities of the individual participating countries still has to be available.

Therefore, archiving has to accomplish two seemingly incompatible tasks: on one hand it has to ensure a high level of data standardisation, and on the other it has to prevent specific information regarding individual participating countries 6 Within the ESS survey, the practice was developed of monitoring current events in each participating country based on mass media reports. Data about such events are then accessible to researchers, along with the data from the survey itself (see Stoop 2007).

7 For more see ISSP and GESIS web page (http://www.issp.org/; http://zacat.gesis.org/).

100

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism from getting lost. This is possible in at least two ways: a) some “nationally specific” information (some demographic variables, such as the degree of education in the form of completed school) is available in the form of specific

“national” variables in a common database; b) and the metadata that include important (methodological) information about the survey conduct in each country are available separately, such as data about sampling, survey methods, survey time, concrete problems in the translation of the original version of the questionnaire, and so on.

Therefore, equivalence problems are solved by the ISSP researchers’ constant care to maintain high methodological standards (from the preparation of the instrument, through translations and sampling, to fielding and archiving), certain working principles and the continuous monitoring and evaluation of the conduct of each survey in each country. Finally, the responsibility for quality rests with each individual researcher who uses the data in their research, as international comparative research equivalence (comparability) problems should also be minimised during the analysis (or use) of the acquired data.

First, through researchers’ adequate (methodological) use of all data that is available to them in the archive, including the metadata. Second, in terms of contents, potential problems of (non)equivalence can be avoided by comparing the relations between the phenomena (variables) in different countries, by comparing changes over time or, rather than engaging in observation (analysis) of individual questions (variables), by trying to achieve conceptual (theoretical) equivalence, through a complex comparison at the level of theoretically defined groups of variables (Teune 1990:54–55). The research design made within the ISSP offers all the above-mentioned possibilities:



− replications of individual modules enable comparisons of changes over time;



− individual modules are always prepared on the basis of (theoretical) models of the association between different variables at the individual level (such as the relation between professional position and attitude to social inequalities) that can later be verified in different contexts (countries);



− most theoretical variables (concepts) are measured with several indicators (survey questions);



− the aforementioned method for the preparation of individual modules and the inclusion of a large number of diverse social contexts (countries) in ISSP

enable complex and theoretically grounded comparisons.



4.3 International Comparative Social Survey Research and the Problem of Combining Different Levels: From the Unit of Data to the Unit of Comparison

After the technical and methodological problems of the comparability and equivalence of data in international social surveys are overcome, new problems start to arise. One of them stems from the fact that countries (systems, societies, cultures) are compared on the basis of data collected at the level of individuals (through interviewing). This is a general characteristic of international comparative surveys within which the analysis takes place by combining different levels: from the data on the individuals, through the (complex) phenomena within individual systems or countries ( e.g. inter-generational mobility, elections, social capital, and so on) to cross-system comparisons. Particular caution is needed in cases when survey data about individuals are merged (aggregated) and observed in the form of percentages (averages or similar statistics) as data about countries. On one hand, the problem lies in the fact that while collecting (and analysing) the survey data about individuals, the heterogeneity of societies is presumed ( i.e., individuals are expected to have different opinions with regard to the observed phenomena), and on the other hand, when societies are compared on the basis of the aggregated individual data, the homogeneity of these societies is often presumed. One of the solutions to this problem is to produce aggregated data from the collected individual data that would point to this heterogeneity.

Even when general methodological (and technical) and certain specific methodological problems in comparative survey research (such as different aspects of providing equivalence) have been successfully resolved, caution is still needed in the analysis and interpretation of the differences identified between the units of comparison. Research programmes, such as ISSP, that enable both dimensions of comparison (“cross-time” and “cross-system”) and interpretation of differences (and similarities), put in researchers’ way at least two temptations that mainly have theoretical or conceptual backgrounds rather than purely methodological and technical ones:



− having identified changes in time, researchers can be tempted to explain these differences with structural changes at the level of society. However, the time-range that is usually covered by the replicated measurements is (still) not wide enough to allow them to also make, on the basis of any changes identified, justifiable conclusions about the essential structural changes at the level of the units of comparison (countries or societies) – such as about any fundamental shifts in the prevalent values in a society. It is, of course, important that the replicated measurements are really equivalent to allow us to speak about the stability (reliability) of the measuring instruments.

102

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism



− Perhaps even more dangerous is the second temptation, which is trying to make conclusions about certain general social (cross-time) trends on the basis of some regularities in the identified differences between the units of comparison (countries or societies). An example of this is given in the analysis based on the research data of the ISSP 2004 – Citizenship. In classifying countries on the basis of similarities and differences regarding the citizenship values of their populations, and the practicing of citizenship in the form of citizens’ political participation, three relatively homogeneous groups of countries were formed that seem to correspond fairly well with their potential classification on the basis of their economic placement, “democratic experience”, and geographic position: a) “the first world” (highly developed old democracies, based on European and North American cultural traditions; b) “the second world” (mainly young European democracies, including Slovenia); c) “the third world” (mainly Latin American countries that are still dealing with the initial problems of democratisation) (see Deželan et al. 2007:109–114). This result and the combination of a Eurocentric bias and forgetting about the necessary distinction between cross-time and cross-system comparisons can quickly lead us to jump to conclusions about the phases of citizenship values and practices through which the democratisation process supposedly proceeds. Failing to consider the “individual” historical (cultural, political) experiences and specificities of current circumstances in the respective units of comparison, means that this kind of conclusion would undoubtedly be a gross simplification, to say the least.

Such problems can be avoided by having an in-depth knowledge about the units of comparison, that is, the countries themselves. However, in research programmes, such as ISSP, that include a relatively large number of countries, researchers cannot always be expected to possess an in-depth knowledge of each individual nation. The problem of the (lack of ) knowledge or understanding of the units of comparison is mainly solved by establishing international research groups, along with providing adequate theoretical grounding for the selection of countries participating in the analysis. We can thus agree with the thesis that the problem of international comparative research is not so much technical (methodological) as it is theoretical in nature ( e.g. Scheuch 1990). It is a theoretical question of what, within the international comparative survey research, a) is considered the system or context of the researched phenomenon, or what b) is the unit of comparison.

The above example calls attention to problems regarding the understanding of the meaning of the (national) state within international comparative research.

Within ISSP, this is further reflected in certain “inconsistencies” regarding respect for the “integrity” of a (national) state as the context for data collection

Selected Problems of International Comparative Social Survey Research 103

or the of comparison: a) while the Slovenian research group was at first rejected by the ISSP (during the late 1980s) for not conducting the survey in the entire territory of the then recognised (federal) state of Yugoslavia, in the late 1990s, another research institution was accepted by the ISSP, although it only carried out surveys in the Flemish part of Belgium (with the entire territory of Belgium being covered in the more recent period); b) after the unification of Germany, the eastern part of Germany was also included in ISSP (with data collection in both parts of the country carried out by the same institution), but analytically, data still exist (in the data base) that enables a separate consideration of eastern and western parts of Germany; c) the ISSP enabled a separate consideration of the Jewish and Arabic parts of Israel, and, of course, all three parts of Belgium (Wallonia, Flanders and Brussels).

All this points to the fact that the unit of comparison or the interpretative context of the studied phenomenon cannot be sufficiently defined merely by the geographical framework of a country. For reasons of survey conduct – such as its sampling framework – the notion of a country with its geographic and administrative borders is usually considered a starting point for international social survey research ( e.g. Scheuch 1989:152). However, different meanings can be ascribed to this “geographic framework”: nations (countries), cultures or societies ( ibid.) can be compared. A specific issue of international social survey research in today’s globalised world is the so-called Galton problem, which appears as a general characteristic of any international social comparative analysis. This is when the factors of the studied phenomenon cannot be exclusively limited to the phenomena within the “geographical framework” of the unit of comparison ( e.g. a country), because cross-border or even cross-culture impacts are also involved (see for example Scheuch 1989:153–154). This means that the unit within which the data are collected ( i.e. a country participating in international comparative research) is not always the unit of analysis and interpretation – that is, the unit to which the theoretical conclusions of the study refer.

Essentially, this is a problem with regard to understanding the context (geographical space of data collection) in which the studied phenomenon is observed – whether it is understood as a political-administrative unit, or as a society or culture. However, none of the ways of understanding the context necessarily correspond to the geographical and political-administrative borders of the (national) state. It is thus also possible to say that Slovenia, as a former federal republic, was a political-administrative unit within the former Yugoslavia (much the same as Yugoslavia in its entirety), although its territory did not correspond to the geographical borders of the then internationally recognised state. Moreover, some studies of the social structure of the former Yugoslavia showed that in this framework it was possible to speak about several societies

104

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism that were (at least partly) delimited by the very borders of its federal units (republics) (see for example Hafner Fink 1994). In sociological international comparative analyses countries can naturally appear in different “roles”.

Melvin Kohn, for example, speaks about four possible types of international comparative research with regard to how countries are considered: a) a country is considered a research object, when we are interested in the selected countries themselves (and the analysed phenomena in them) that are compared; b) countries can be understood as a (social, cultural, historical) context, when they are included in comparisons due to a specific configuration of conditions that they represent ( e.g. consolidated democracies and the countries in the process of democratisation); c) when an international comparative research only focuses on the analysis of connections between variables ( i.e., some characteristics of the countries), similar to classical quantitative research, the country is considered a unit of analysis; d) a country can be understood as a constituent of a wider international (supra-national, global) system that we are essentially interested in (Kohn 1989:20–24).

If we assume that one of social research’s key epistemological aims is developing (and improving) theories, and that this goal is best achieved through comparative research (see Ragin & Amoroso 2011:138–143), then the most sensible approach for international survey research, such as the ISSP, seems one that understands the researched country (or geographical unit) as a context in which individuals experiencing the intertwined work of various factors form (and express) their own viewpoints and act as social actors. As such, what is important in international comparative research, which should be theoretically very well grounded, are no longer the number and size of participating countries alone, but primarily also the fact that these countries represent different social (cultural, political, historical) contexts or configurations of relevant factors of the studied phenomenon. The principles of ISSP’s operation can also be understood in this sense, based on which a) data for some countries are also shown “regionally” (such as Western and Eastern Germany) or b) the research programme also includes institutions from countries (such as Taiwan), whose sovereignty is disputed in international politics (community).

4.4 Some “Institutional” Aspects of International Social Survey Programmes, or why ISSP has Continued for over three Decades Which characteristics have enabled ISSP’s continued operation for over 30

years, and the uninterrupted growth in the number of participating research institutions? In addition to mentioning its working principles, it should be noted that it is an organisation established primarily (if not exclusively) for academic motives. Its key goal is thus the preparation of a data infrastructure that is widely accessible to all interested researchers. Moreover, inclusion in the ISSP programme enables researchers and research institutions to form academic connections and carry out col aborations with other projects. While other international social survey programmes, such as WVS, EVS and ESS, operate according to similar general principles and with similar goals, compared to these ISSP has some additional features regarding its organisation and working principles.

Rather than working as a “classical” (international) research project, ISSP is actually an international social-science research “movement” consisting of the network of social science research institutions and their researchers. Its foundation was not encouraged through a tender issued by an international foundation, but instead it “spontaneously” developed from the “grassroots”

based on the ideas of researchers from research institutions that were already doing similar work at the national level. Therefore, the selection and contents of ISSP annual modules are not the result of a “political” or “bureaucratic” project, but of academic debates among social scientists from different political, social and cultural environments across the world. It is thus not surprising that ISSP

data have been very intensely used throughout the project’s existence as the basis of various social scientific projects, and that this has led to numerous papers for scientific conferences, academic articles, doctoral dissertations, conference proceedings and monographs, as well as other achievements.8 Of course, one could argue that these are also the characteristics of other international social science programmes, yet within ISSP these are further “consolidated” because the activities are carried out annually, while, for example, WVS and EVS are repeated after much longer intervals.

Moreover, the operation of the programme does not depend on any common financial resource that would guarantee at least the carrying out of common tasks and co-ordination, as applies to the ESS programme. Instead, all activities within ISSP are financed from the resources acquired by the ISSP members 8 According to the ISSP Bibliography prepared by Tom Smith, 5225 scientific publications were recorded until 2012 that were based on the analysis of the ISSP data (ISSP 3).

106

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism (research institutions) in the framework of research programmes carried out in the countries they come from. This is why the relations within the ISSP are never threatened by conflicts that would be based on the allocation of funds. All the costs of ISSP’s operations (secretariat, working bodies, the organisation of annual general meetings, and so on) are always covered by the participants from the resources they have available within their own research institutions. However, there are also downsides to this, which show in the stronger role played by larger research institutions coming from richer, often “Western” countries that have more stable funding. Therefore, biases, which here can be defined as “Western-centrism”,9 can occur in the selection and design of the topics of ISSP modules.

To avoid this, the working principles are designed to guarantee the equal participation of all ISSP members in conceptual and methodological discussions.

This is shown both in the heterogeneous composition of commissions and working groups, as well as in the democratic principles of making thematic and organisational decisions.

One aspect that is undoubtedly important for overcoming “Western-centrism”

is the informal or social nature of ISSP’s operations. Having been formed as a network of research institutions, ISSP is not only based on the connections among research institutions, but also those among researchers as individuals. It is the ISSP’s very working principles that enable a basically informal form of col aboration to occur among institutions and researchers from various cultural environments, which is not grounded in “bureaucratic” logics, but instead in academic principles as well as the ethics implied in its founding principles. It can thus be concluded that to a considerable extent ISSP works based on the

“ethical drive” of its members, with important reasons for col aboration in the

“movement” including academic motives and academic “altruism”, as ISSP’s central goal is to create social scientific databases that would be accessible to all interested researchers. However, there is also a “practical” side to ISSP’s ethics: for example, its working principles can be said to facilitate an adequate level of data quality in terms of equivalence, as discussed above. Finally, the ISSP’s working principles reflect an important belief and general goal to be followed by social scientists in their research work that the overall aim of such efforts is working for the benefit of humanity in general.

9 The concept “Western-centrism” here defines biases that come from the domination of theoretical concepts grounded in Western-European and Northern-American cultural spaces.



Part III

The Quantitative Approach and some Problems

of Social Surveys

The content in this part ranges from testing specific methodological concepts through to experiences with practical research methods within social surveys.

In the fifth chapter, we address the nature of the cognitive representations respondents create and utilise when making evaluative judgements in surveys.

More precisely, the chapter presents an empirical test of the existence of context effects in the framework of actual survey data from the Slovenian Public Opinion Survey’s research programme. The starting points of our analysis are two hypothetical statements: (1) the order of items affects both the cognitive representation of the underlying dimension and the factor structure of the answers; and (2) the different levels of respondents’ cognitive sophistication affect their recognition of the concepts measured and the nature of their answers. A multidimensional scale of the concept of ‘negative nationalism’ is analysed. The research results partly confirm these hypothetical statements: the level of a respondent’s education (as an indicator of the level of their cognitive sophistication) influences their recognition of the concept measured, while other classical context characteristics (such as item order) did not. We can only confirm the possible effect of item order on the nature of their response, which results in the two-dimensionality or unidimensionality of the concept being measured.

The last chapter brings a shift from specific methodological issues to the wider field of research approaches or research designs. Regarding the era gone by that was significantly more favourable for conducting telephone surveys, in the last chapter of the present edition we follow the genesis of telephone surveying as developed within the research programme at the Public Opinion and Mass Communication Research Centre (POMCRC). It presents the circumstances in which work on improving this particular data acquisition method was undertaken, illustrates the related needs and expectations existing at the time of its development, and touches on both the history of telephone surveying and this research method’s possible prospects in contemporary empirical social sciences. Beyond what telephone research can still offer us today (a dilemma worthy of further reflection), we are interested in what this approach permitted during the heyday of this research practice. Highlighting the connections between the progress of modern communication and computer technology and the possibilities they have created for this kind of research, the overview presents the characteristics of telephone surveying, and its comparative advantages and disadvantages over other methods. It goes on to describe the research practice at the POMCRC, the research centre that, by entering the sphere of telephone surveying relatively early, was able to establish this approach as one of its own fundamental methodological and research tool during the Slovenian transition.



Context Effects in Social Surveys:

Between Instrument and Respondent

5

Social survey researchers tend to reduce or control as far as possible (systematic) context effects on the replies of their respondents.1 To this end, researchers must determine whether respondents – as a result of the particular context effect-are able to identify a measured concept ‘hidden’ within the questions asked. In this chapter, we discuss two aspects of context effects: a) so-called ‘local context effects’ (the effects of the measurement instrument); and b) so-called ‘global context effects’, which include the motivational and cognitive basis of attitudes (see Uhan 1998). We observe the local context as a question-order effect and the global context as the effect of respondents’ cognitive sophistication (level of education) (Krosnick 1992; Johnson et al. 1999).

When measuring a concept (such as nationalism, in/tolerance, social distance, prejudice) within a social survey, researchers often use statements with which respondents express their agreement or disagreement. These usually take one of two forms: a) a balanced list of positive and negative items that cover the two poles of the same dimension ( e.g. tolerant – intolerant) (bipolarity); and b) a list of items covering two or more (sub)dimensions of the same concept, which are not necessarily antagonistic (multidimensionality) ( cf. Hafner-Fink and Uhan 2013). In this chapter, we focus on the second form.

Context effects are not usually identified directly in respondents’ replies, but are rather based on hypothesis. Context effects can be defined as those changes in responses to survey questions that are a consequence of the characteristics of the questionnaire or of the circumstances in which the survey takes place. If these effects were not present, the replies would be different, namely unaffected by context. The key to understanding context effects is the mental representation, or the model of information processing that the respondent references when forming his/her response. The replies of respondents can thus be influenced by the formal characteristics of the instrument used; for instance, the question order, the type of scale, etc. (see Schuman & Presser 1981; Tourangeau & Rasinski 1988; Smith 1988; Sudman, Bradburn & Schwarz 1996). On the other hand, some authors have considered other possible effects based on the specific circumstances in which the survey interview takes place – such as the personal, cultural and social context of those involved – which have an explicit or implicit influence (see Zaller & Feldman 1992; Turangeau, Rips & Rasinski 2000; Hair 2005). In this respect, the most frequent question researchers ask is whether the context effects should be treated as a ‘temporary disturbance’, or whether they represent a serious, systemic fault that may diminish the significance of the survey results or findings.

1 The chapter was previously published in journal Teorija in praksa, Jan.-Feb. 2013, Vol. 50, No.

1, p. 233–248.

114

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism Although the methodological literature cites many instances of research into context effects, most are experimental studies constructed in order to demonstrate context effects. This does not diminish their validity, but it does raise the question, how often and in what way do context effects appear in non-experimental circumstances?

Researchers often point to the well-known study by Schuman and Presser (1981), who describe the influence of context in researching standpoints in ‘normal’

non-experimental circumstances. On the basis of their analysis of results in the DAS (Detroit Area Study), they establish that the likelihood of the occurrence of context effects is scarcely any greater than coincidence. Smith (1988) reaches a similar conclusion following his analysis of the replies in the GSS (General Social Survey). Smith establishes that a random rotation of questions leads to context effects in only four percent of cases. However, these findings which indicate a low or coincidental effect of context on responses can be misleading.

Tourangeau, Rips and Rasinski (2000) demonstrate that context effects are more frequent than one might expect. They contend that researchers often overlook the fact that the appearance of context effects depends on the conceptual links among questions. This means that heterogeneous research often conceals context effects.

In this paper we will verify the appearance of context effects in the ‘natural’

environment of a public opinion survey. In our research we will check for the traditional effect of survey context on survey responses, i.e. the influence of the order in which questions are asked. At the same time, we will also test the influence of context in the case of related questions or statements, based on the hypothesis that a respondent’s cognitive sophistication influences their ability to identify what is being measured (socio-cultural context). And in this situation we expect socio-cultural context (the ability of respondents to identify the concept measured) to override the expected effects of the survey context.

5.1 The Problem

In order to operationalise concepts, the standard design of social surveys includes two basic types of questions that can be referred to as objective and subjective.

As a rule, respondents have fewer problems framing responses to objective questions than to subjective questions, since rather than the recognition and linking of facts, the latter demands the creative processing of information. This awareness has led to a debate as to whether it is theoretically appropriate to view the underlying variable as a unidimensional continuum.

Context Effects in Social Surveys: Between Instrument and Respondent 115

Bipolar survey items have long been held to be appropriate instruments for the operationalisation of theoretical models. This is because it is relatively simple for them to prompt the respondent to a cognitive representation of the object of research that is close to the relevant concept expected by the researchers.

The application of a bipolar scale, however, raises a number of methodological questions. In the face of ‘faulty’ results, cognitive psychologists have drawn attention to the fact that the actual instrument, namely the bipolar scale of statements, can itself be the object of the cognitive representations that the respondents form of the research and the context in which it is carried out. If cognitive representations play a crucial role in models of information processing, this would also imply the hypothesis that these representations have a direct influence on the formation of responses to survey questions and, consequently, that they reduce the validity of the results. The experimental findings relating to this methodological problem (presented below) emphasise the importance of taking account of both the socio-cultural context (cultural and social) and the context in which the survey is conducted (the characteristics of the instrument).

Traditionally, social surveys address respondents’ attitudes as a latent variable that can be adequately measured on a bipolar response scale. These models assume that the cognitive structures underlying the responses to a bipolar survey scale are both unidimensional and continuous. The term ‘unidimensional’

refers to the notion that there is a single dimension of variability for the class of stimuli being judged. The term ‘continuous’ refers to the fact that there are no breaks in the dimension: that is, that there is a seamless gradient from one end of the latent variable to the other. A cognitive structure that has the properties of unidimensionality and continuity should produce judgments that are reciprocally antagonistic: that is, as one moves away from one pole of the response continuum, one will inevitably move towards the opposite pole (Ostrom et al. 1992:298). If, for example, one were measuring the ‘subjective’

dimension of ‘(in)tolerance’ then one end of the bipolar scale would consist of a statement that expressed extreme intolerance, while at the other end would be an highly tolerant statement.

Early on, researchers noted that the concept of a unidimensional latent variable was unsuitable for evaluating complex social phenomena. For instance, researchers encountered difficulties when attempting to assess subjective standpoints as to whether the death penalty is morally justifiable. The measuring of such a standpoint proved possible only when the respondents combined evaluations for different aspects of the death penalty (ethical, legal, cultural, historical, etc.). According to Ostrom (Ostrom et al. 1992), respondents face a similar task every time they evaluate the properties of objects that differ from each other but are not mutually exclusive (when measuring a complex,

116

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism subjective standpoint). As an example of an ineffective scale, Ostrom cites the traditional American presidential question that assesses the conservatism or liberalism of the candidate. The scale includes two poles on what appears to be the same continuum – liberal and conservative – which the survey participants are expected to perceive as mutually exclusive, although much research shows the opposite.

A newer approach to social psychology (above all cognitive theory) reconceptualises the traditional psycho-physiological models for presenting opinion, or, as Ostrom (Ostrom et al. 1992:298) notes, ‘The attitude construct, traditionally viewed as a bipolar continuum, can be reconceptualised in terms of two discrete categories that are separately stored in a semantic network.’

The essence of this approach is the desire to identify the discrete cognitive representations triggered by the survey questions. The respondent has two kinds of cognitive representations available: those that relate to the object of evaluation and those that relate to the instrument of measurement (for instance, the abovementioned scale of conservatism or liberalism of the presidential candidate).

Prior to giving a response, the survey participant’s task is to find the highest level of correspondence between the two cognitive representations, the object and the instrument of measurement. Continuing with our the example of evaluating the presidential candidate on the conservative-liberal scale, it is reasonable to assume that the extreme poles of this scale trigger separate cognitive structures or representations (a prototypical conservative or liberal president). The task of the respondents is to decide which of the prototypical representations (or both together) best describe the candidate.

In spite of the apparently unidimensional bipolar questioning, the likelihood that, in giving their answers, the respondents will make use of separate cognitive categories that can ‘support’ both poles of the scale (simultaneously conservative and liberal) has prompted theorists to surmise that the latent cognitive structures are dualistic and discrete (Ostrom et al. 1992:298). In discrete categories, the two poles are linked by content but are nevertheless independent of each other; each pole is not necessarily a negation or inverse property of the category at the other pole. Dualistic categories are linked but not necessarily antagonistic.

In the conservative-liberal example, the respondent can express his or her attitude as non-conservative, but not at the same time liberal. In other words, we are no longer talking about one dimension on a bipolar conservative-liberal continuum, but about two separate dimensions – liberal and conservative. Based on the assumption of duality, Ostrom designed an experiment to reject the hypothesis that the latent variables were merely unidimensional and continuous.

The results of the experiment confirm duality only in both block sequences, while bipolarity is present in the random ordering of statements (Ostrom et al.

Context Effects in Social Surveys: Between Instrument and Respondent 117

1992).The starting point for Ostrom’s experiment was the ‘Donald case’, which is known in socio-psychological research and which describes an average day in the life of a fictitious character known as Donald. Donald has both the positive and negative features of an average person. In the original experiment, the researchers tested a number of hypotheses with a pre-prepared evaluation test on the unidimensional continuum of ‘friendly-unfriendly (malicious)’. To this end, Ostrom made use of a special procedure that involved a short presentation of Donald, which incorporated a ‘disturbing story’. The purpose of this story was to negate the effect of recentness.

In carrying out his experiment, Ostrom used three different sequences of questions: each sequence incorporated the same twelve statements which served to prompt the respondents to express their impressions of the target person. Six statements fitted the latent dimension ‘friendly-unfriendly (malicious)’, while the other six were construct-irrelevant. Ostrom used an eleven-step scale of agreement presented on the two pages of the questionnaire, which each carried six statements. The sequences in Ostrom’s experiment differed in terms of the location of the relevant and irrelevant statements. In the random ordering sequence, the twelve statements were ordered at random with three relevant and three irrelevant statements on each page. In the 3-6-3 sequence, the first three statements on the first page were relevant to one of the poles of the scale, while the last three statements were relevant to the opposite pole; the middle six statements in this sequence were irrelevant. The 6-6 sequence placed all the relevant statements on one page. According to Ostrom’s findings, duality (two-dimensionality) can be expected when the measurement instrument enables the direct identification of the concept due to the ordering of the questions in both block sequences. In the case of the random sequence of items, bi-polarity (unidimensionality) would be expected. The findings of other researchers suggest that when the instrument enables the direct identification of the concept, the opposite effect (a one-dimensional structure) can be expected (Hafner-Fink & Uhan 2013). It seems that the latter indicates the ‘global context’ effect (the content of the concept and the cognitive sophistication of respondents), which may override the question order effect ( ibid. 850–851).

5.2 The Research Model

These findings formed the starting point of our research. Apart from question order effect (‘local context’), we were interested in the ability of respondents to identify what was being measured (‘global context’). The test was performed in the autumn of 2003, within the framework of the Slovenian Public Opinion

118

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism Survey (SJM 2003/3 and SJM 2003/4; see Toš et al 2004a; 2004b). We obtained 1777 questionnaires which answered all the relevant questions (2002 adult residents of Slovenia were surveyed).

The respondents were divided into three sub-groups, each of which responded to a different version of the questionnaire.2 With the help of a five-step (Likert-like) scale, the participants expressed agreement or disagreement with a group of twelve statements, six of which were designed to express two dimensions of negative nationalism (protectionism and xenophobia), while the remaining six were construct-irrelevant. In each version of the questionnaire, the position of the construct-relevant items were changed in line with Ostrom’s sequence, so that, in the first version, the first three statements expressed xenophobia and the last three protectionism, while the six intermediate statements were construct-irrelevant ( i.e. using a 3-6-3 sequence). In the second version, all six construct-relevant statements appeared in one block that comprised the concept ‘negative nationalism’, while the six irrelevant statements appeared separately on the other page of the questionnaire ( i.e. a 6-6 sequence). In the third version, the twelve statements appeared in random order, with three relevant and three irrelevant statements on each side of the questionnaire, irrespective of the concept.

5.3 Analysis and Results

We supplemented Ostrom’s thesis on the influence of the measuring instrument (bipolarity and item order) with the thesis on the influence of the ‘content’ of the concept measured. We thus carried out a test in which all the statements were unidirectional (which means that we excluded the possible influence of bipolarity) and we researched only the influence of the item order on the ‘formation’ of the two-dimensionality (or unidimensionality) of the concept of ‘negative nationalism’.

A Likert scale was used to measure the level of negative nationalism, which was hypothetically divided into two dimensions: xenophobia and protectionism. Here, each dimension was represented by three statements that were combined in the questionnaire with six irrelevant statements.3 The scale was constructed as follows: 2 The respondents completed a survey sheet which was attached to the basic SJM questionnaire.

3 The six irrelevant statements covered concepts (dimensions) of authoritarianism and traditionalism that are theoretically (conceptually) related to the relevant dimension (concept) of nationalism. Because of this affinity of the relevant and irrelevant concepts, the ‘task’ of

‘identification’ of the relevant dimension was not trivial for respondents. The following six irrelevant items were included in the questionnaire:

- Homosexuals should not be allowed to express their sexual orientation in public. (x7)

- Old customs are being destroyed by modern times. (x8)

Context Effects in Social Surveys: Between Instrument and Respondent 119

a) the dimension of ‘protectionism’ was represented by the following three statements:



– Slovenia should limit the import of foreign products to protect its economy. (x1)



– Foreigners should not be allowed to buy land in Slovenia. (x2)



– Slovenian television stations should give precedence to Slovenian films and programmes. (x3)

b) the dimension of ‘xenophobia’ was represented by the following three statements:



– The crime rate is increasing because of the number of immigrants. (x4)



– Non-Slovenes should not be allowed to hold public posts. (x5)



– Mixing people of different ethnic and cultural backgrounds brings only problems. (x6)

To test both models (unidimensionality and two-dimensionality) for each of the three methods of ordering (3-6-3; 6-6; and random), we used confirmatory factor analysis (CFA) on the basis of a correlation matrix of the six relevant statements shown. We first carried out an exploratory factor analysis (the principal axis method) which in this case yielded similar results to those arrived at by Ostrom and his colleagues: in both block sequences there is a clear two-dimensional solution (protectionism, xenophobia); while in the case of the random ordering of statements, a unidimensional solution makes more sense.4

Using confirmatory factor analysis for all three methods of ordering the relevant statements, we tested two measurement model variants: a) a two-factor model (a thesis of the two-dimensionality of the concept of negative nationalism) –

protectionism (ξ ) and xenophobia (ξ ); and b) a single-factor model, which 1

2

presupposes the unidimensionality of the concept of negative nationalism (ξ).

Here too, we tested using both models, and for each model we tested the three conditions of ordering the relevant statements (Figure 5.1). We compared the results of the testing of all six models to the same degrees of freedom (df ).

- I am always prepared to support new things. (x9)

- Human life is determined by destiny. (x10)

- In general, a child will benefit if he/she accepts his/her parents’ way of thinking. (x11)

- A community that tolerates large differences in beliefs cannot survive in the long run. (x12) 4 In the instances when irrelevant items were also included in the model, the results were structurally the same: a) in both block sequences, the two expected factors of nationalism (xenophobia and protectionism) and the two factors of irrelevant items were formed; b) in the random sequence, two factors were formed – one factor of relevant items (nationalism) and one factor of irrelevant items. There are only a few deviations: a) one irrelevant item (about homosexuals) ‘joined’ the factor of xenophobia in both block sequences; and b) one irrelevant item (about homosexuals) ‘exchanged’ positions with one relevant item.



120

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism To evaluate the models’ fit we applied two statistics: the ratio χ2/df and the root mean square error of approximation (RMSEA).5

FIGURE 5.1 General hypothetical models for the scale of ‘negative nationalism’



5 There are different views in the literature as to the threshold value for each statistic. For the χ 2/df ratio, ‘different researchers have recommended using ratios as low as 2 or as high as 5 to indicate a reasonable fit’ (Marsh and Hocevar 1985:567). There is also disagreement regarding the RMSEA value: some authors are more conservative and put the value of 0.05 as the upper boundary, while others are more liberal and put the value of 0.08 as the upper boundary for a good model fit ( e.g. MacIntosh 1998:87; Li and Wehr 2007:376).

Context Effects in Social Surveys: Between Instrument and Respondent 121

5.3.1 The Question-Order Effect

Only when allowing for correlations of measurement errors in all three methods of ordering were we able to obtain a good fit between the model and the data in the case of both the single-factor and the two-factor models. However, when comparing the results of testing for all six models to the same degrees of freedom, differences in the model’s fit are present ( Tables 5.1a, 5. 1b, and 5. 1c).

These differences and the specific results for each model do indicate item-order effects. In the case of the two block sequences, the fit between the single-factor model and the data is significantly worse than the fit between the two-factor model and the data. Thus, the single-factor model’s fit with the data in the case of the random ordering of statements is stronger than in both block sequences, while the opposite applies for the two-factor model; random ordering means a worse fit between the model and the data. So, for the block sequences, a two-factor solution makes most sense and for the random distribution, a single-factor solution. Similarly, given the correlations among the factors (ϕ ) in the case of 21

the two-factor model, we can conclude that a single-factor solution makes more sense for the random ordering, as the correlation is considerably higher (ϕ =

21

0.866) than in either of the block sequences (0.624 and 0.573). Moreover, the correlations among the measurement errors (δx ) shows the logic of a two-factor ij

solution in the two block sequences: to obtain a good fit for the single-factor model, we were obliged, in both block sequences, to allow for correlations among the measurement errors of the statements within the same measured dimension (specifically, for protectionism); however, there was no correlation among the measurement errors for the statements that we included in the different dimensions (Tables 5.1a and 5.1b). In the case of random ordering, to obtain a good fit between the model and the data we had to allow for correlations among the measurement errors; although, in this instance, we were not dealing with systematic correlations within particular dimensions, but rather correlations of errors between ‘protectionist’ and ‘xenophobic’ statements.

The results thus show possible fits between the data and both the single-factor and two-factor models, but only when allowing for correlations between measurement errors for particular statements. On this basis, we can conclude that, behind the measured concepts relating to negative nationalism (common to both), there is another, more general dimension for which we had to allow certain correlations of errors which we initially assumed to be illogical. Namely, we must consider the possibility that irrelevant statements could be a by-product of associated attitudes, such as authoritarianism and traditionalism. For instance, a nationalistic attitude in the form of protectionism or xenophobia can also be imbued with authoritarianism.

122

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism TABLE 5.1A The results of confirmatory factor analysis (CFA) – test of the



scale of ‘nationalism’, sequence 3-6-3 (SJM 2003/3-4)*

one-factor model

two-factor model

Statements:

lambda - x

R2

lambda - x

R2

neg. nationalism

protect.

xenoph.

(ξ)

(ξ )

(ξ )

1

2

x1 – no foreign products

0.417

0.177

0.661

- -

0.337

x2 – no land for foreigners 0.490

0.240

0.653

- -

0.426

x3 – precedence to

0.401

0.161

0.578

- -

0.334

Slovenian films...

x4 – more crime because 0.679

0.461

- -

0.755

0.570

of immigrants

x5 – no public posts for

0.701

0.492

- -

0.655

0.429

Non-Slovenes

x6 – mixing of people –

0.799

0.639

- -

0.881

0.776

problems

φ = 0.624

21

δx = 0.190; δx = 0.205

δx = -0.135

21

31

64

df = 7; χ2 = 35.660 (P = 0.000)

df = 7; χ2 = 9.302 (P = 0.232)

RMSEA = 0.083; χ2/df = 5.094

RMSEA = 0.023; χ2/df = 1.329

* Results are shown in a standardised version

TABLE 5.1B The results of confirmatory factor analysis (CFA) – test of the



scale of ‘nationalism’, sequence 6-6 (SJM 2003/3–4)*

one-factor model

two-factor model

Statements:

lambda - x

R2

lambda - x

R2

neg. nationalism

protect.

xenoph.

(ξ)

(ξ )

(ξ )

1

2

x1 – no foreign

0.362

0.134

0.603

- -

0.363

products

x2 – no land for

0.465

0.217

0.754

- -

0.568

foreigners

x3 – precedence to

0.403

0.162

0.669

- -

0.448

Slovenian films...

x4 – more crime

0.684

0.467

- -

0.685

0.469

because of immigrants

x5 – no public posts for 0.710

0.503

- -

0.717

0.514

Non-Slovenes

x6 – mixing of people – 0.762

0.581

- -

0.767

0.588

problems

φ = 0.573

21

δx = 0.241; δx = 0.218

δx = -0.166

21

31

32

df = 7; χ2 = 32.254 (P = 0.000)

df = 7; χ2 = 9.875 (P = 0.196)

RMSEA = 0.079; χ2/df = 4.608

RMSEA = 0.027; χ2/df = 1.411

* Results are shown in a standardised version

Context Effects in Social Surveys: Between Instrument and Respondent 123

TABLE 5.1C The results of confirmatory factor analysis (CFA) – test of the



scale of ‘nationalism’,random sequence (SJM 2003/3-4) *

one-factor model

two-factor model

Statements:

lambda - x

R2

lambda - x

R2

neg. nationalism

protect.

xenoph.

(ξ)

(ξ )

(ξ )

1

2

x1 – no foreign products

0.594

0.352

0.554

- -

0.307

x2 – no land for foreigners 0.623

0.388

0.684

- -

0.468

x3 – precedence to

0.491

0.241

0.528

- -

0.278

Slovenian films...

x4 – more crime because 0.621

0.385

- -

0.658

0.433

of immigrants

x5 – no public posts for

0.691

0.477

- -

0.702

0.493

Non-Slovenes

x6 – mixing of people –

0.762

0.581

- -

0.745

0.555

problems



φ = 0.866

21

δx = -0.221; δx = -0.132

δx = -0.137

51

61

51

df = 7; χ2 = 21.090 (P = 0.004)

df = 7; χ2 = 19.950 (P = 0.006)

RMSEA = 0.058; χ2/df = 3.013

RMSEA = 0.056; χ2/df = 2.850

* Results are shown in a standardised version

5.3.2 Empirical Grounds for the Socio-Cultural Context Effect Hypothesis In general, the results support the hypothesis of item order-effect (survey context), but some findings also give rise to certain doubts: a) to obtain a good fit for a two-factor model in both block sequences, a correlation between the factors was included in the model; b) to obtain good fits for all models, some correlations between measurement errors were necessary. Owing to these facts, we argue that the item-order effect (survey context) was corrupted with certain other effects. We have already mentioned the possible effect of irrelevant statements.

But we are more interested in the possible effect of the varying cognitive sophistication of the respondents (socio-cultural context), which could result in varying abilities to identify the measured concept and (al ) relevant items for the concept. This would mean that not only do the technical characteristics of the instrument (item-order) affect respondents’ answers, but also the content of relevant (and irrelevant) items. Our experiment questionnaire also included questions which would allow us to test the abilities of respondents to identify relevant items from among all 12 relevant and irrelevant items. Respondents were asked to select 6 items which were relevant for the measured concept. The mean value of the number of correctly identified items is 4.20 and there are only minor (statistically non-significant) differences between the three methods of

124

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism ordering: a) 4.13 for the block sequence 6-6; b) 4.20 for the block sequence 3-6-3, and; c) 4.24 for the random sequence.

We then tested the correlation between the respondents’ level of education and their ability to identify the relevant items. For the purposes of our investigation, we take the level of education to be an indicator of general competence or cognitive sophistication as one of the dimensions of the socio-cultural context.

The variations in the different education levels are statistically significant (p < 0.01) – a higher level of education indicates a higher level of identification: the mean value of correctly identified items within the group of respondents with elementary school (or less) is 4.03 (SEM = 0.079). This rises to 4.50 (SEM =

0.089) in the group of respondents with college or university education. To compare both effects (respondents’ education and item-order), we carried out a linear regression analysis with the number of correctly identified relevant items as a dependent variable and a) education level (4-point scale) and b) item-order binary variable (non/random) as predictors. Only the effect of respondents’

education has proved to be statistically significant, albeit weakly so6 (Beta = 0.10; p < 0.01) (see Table 5.2).

TABLE 5.2: The effects of education and question order on item



recognition – results of multiple linear regression analysis



(SJM 2003/3-4)

B

Beta

P

intercept

3.808

0.000

– education

0.150

0.100

0.000

– question order (non/random)

0.078

0.024

0.310

N

1730

Adjusted R2

0.010

F

9.354

0.000

5.4 Discussion

The general premise of our research was to test the presence of the context effect of survey questions on the responses obtained, as a consequence of the manipulation of the order in which they occur. The results we have obtained allow us to draw the following conclusions about the influence of context. We 6 When the level of education was increased by one degree, the level of correctly identified items increased by 0.15 (at the intercept value of 3.80).

Context Effects in Social Surveys: Between Instrument and Respondent 125

can confirm the work of authors who emphasise the importance of taking context into account when interpreting research findings ( cf. Tourangeau et al. 2003), as well as those who assert that these effects are not as common as might be concluded on the basis of experimental findings ( cf. Smith 1988). Our findings suggest that, in specific situations, the question-order (survey context) effect can be overridden by the socio-cultural context. The survey participants responded to the content of the questionnaire more than to its technical specificities and in doing so made use of their representations of the concept being measured. This can happen only when respondents identify the concept.

The results confirm both the hypotheses that: a) the technical properties of the instrument (local context) are likely to exert an influence on the participants’

responses; and b) that cognitive sophistication (socio-cultural or global context) may influence the results. Regarding the local context: the two-dimensionality of the concept of ‘negative nationalism’ was formed when the structure of the questionnaire clearly showed it in the two block sequences. When the statements appeared in random order, the two-dimensionality was not ‘identified’. We can confirm the general findings regarding the influence of the local context (already mentioned in our introduction), especially the following: 1. Respondents form representations of the measuring instrument, which has an influence on their response process or choice of modalities; 2. Above al , the findings confirm the influence of survey context or item order on the formation of dimensionality of the concept being measured.

Nevertheless, there are important additional effects which could not be simply ascribed to the context as it is usually understood in survey research. The understanding of the survey context should be broadened by applying the new dimension which we referred to as the socio-cultural or global context. In our research, this dimension is firstly present as an effect of the content of research (or the concept measured) which stimulates some responses according to the level of cognitive sophistication of respondents. This is also related to our key finding, that the level of education of the respondents influences their recognition of the (content of the) concept being measured, while other classical context characteristics (such as item order) does not influence the recognition of the concept. Based on the results of the analysis, we should stress that the effect of education is also rather weak.

This result also has direct research implications. We can see that respondents with a higher level of education are more likely to recognise the concept being measured, which means that they are also more capable of identifying the semantic variability of the different wording of the question. The opposite is true for less educated respondents, who express significantly lower semantic sensibility ( cf.

Uhan 1998). The problem is that cognitive sophistication (in our case indicated

126

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism by the level of education) affects the identification and interpretation of the researcher’s intention ( cf. Hippler 1989). Semantic variations in conceptually and logically equivalent questions trigger different perceptions and interpretations of these questions by different categories of respondents (see Uhan 1998:99).

This consequently creates difficulties in drawing comparisons and equivalence in the measurement of results.



Development of a Telephone Survey

within the Public Opinion and Mass

Communication Research Centre –

6 Comparative Aspects of the Method

6.1 Introduction

The developmental trajectory of computer assisted telephone interviewing (CATI) is a specific one. Its early, relatively steep rise subsided in its later history, and recently, the frequency of this kind of research even slightly decreased in some environments, or was partly replaced by other approaches. The dynamics of the development of telephone research was mainly determined by the development of information communication technology that has seen substantial changes over the past few decades. Its importance for the affirmation of this research technique is twofold. On one hand, the operation and possibilities of this technique are enabled by technology, which, on the other hand also presents an obstacle. A short presentation of the development of this survey approach in comparison with some other approaches will indicate its basic characteristics, its methodological reach, and its key comparative advantages that placed it first among the research approaches in terms of its frequency of use.

An important reason for presenting telephone research also involves the existence of some typical studies that were conducted within the Public Opinion and Mass Communication Research Centre (POMCRC) during the past decade, and that belong in its basic research programme. The total number of studies conducted is high, amounting to over 200 studies in total. A presentation of a set of some of the more relevant among them can be found in the second part of the book Values in Transition IV (Toš 2009).

6.2 History

Telephone research is a form of surveying in which data is collected through telephone communication. Data acquisition and recording, and a major part of the sampling procedures are based on similar principles (probability sampling) as in case of other surveying modes. Also the goals that are followed

– namely, to realise the maximum share of the sample design and to acquire the maximum quality data – are similar or equal. The specific nature of this kind of data collection is mainly determined by the development and nature of telephone communication. The history of telephone research is relatively short. Its emergence can be traced back to the 1970s in the USA, where it was affirmed only in the 1980s or later. The initial operationalisation (Dillman 1978; Blankenship 1977; Groves 1988) by which the process of data acquisition over the telephone was both grounded and methodologically justified also reaches back to this period. Before this, practically any data acquisition within social-scientific empirical studies was carried out through personal face-to-face

130

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism interviewing. Although general public knowledge mainly perceived telephone surveys as limited to inquiring about voting intentions, or more generally to the research of political self-determination and market research, telephone research actually became affirmed by the need to conduct applied research whose goals would be broader in contents and social-scientifically more relevant. Current media reports further affirm this impression, namely that first, telephone research mainly involves measurements of political climate with the emphasis on pre-election surveys that the media themselves often (erroneously) understand as the attempt to publish tomorrow’s news today, or simply to predict the winner of each new up-coming election; and second, that the use of the telephone for research purposes is directly or indirectly related to marketing. Researchers notice this as well, when we communicate with people during telephone surveys. Notably, a substantial part of the professional public is also unfamiliar with the broad range of contents and issues that our research centre has dealt with over several decades, since themes that the media find more “attractive” are always more resounding.

The aforementioned data bases (Toš 2009) represent a series of relevant social-scientific empirical studies that come to be less known in one part of the public (also the professional one), although during a specific period, some of them became part of the standard research repertory (such as Opinions about Corruption, Opinions about the Police etc.).

Often, the public’s ideas about and its understanding of the role of telephone research are not in line with the actual reach and scale of this kind of research.

Interestingly, even the first attempts at telephone research in the United States were closer in their contents to applied research projects than to those that the general public today normally links to telephone surveying (marketing, pre-election surveys etc.). These first attempts included, for example, themes involving the social position of specific social groups, issues concerning the development of agriculture, health culture issues etc. The emergence of expert literature – that apart from defining the methodological characteristics, also provided quite practical step-by-step instructions of how to practically conduct a survey (Groves and Khan 1979) – gave impetus to the emergence of countless research projects, prompting, at the same time, the creation of a large number of agencies that started to use the telephone to acquire data. In the 1980s the USA saw a large growth of this research approach. The idea was realised with great expectations and left the impression that telephone surveying would replace most of the personal surveying in the field. We could say that small (local) research organisations were also given the opportunity to expand.

The scope of small and medium sized research organisations that depended on personal contact or mail interviewing notably increased with the possibility of managing large and dispersed national samples. The problem of large countries

Development of a Telephone Survey within the Public Opinion 131

with a highly dispersed population (such as the USA) has always been how to define and enable practical access to respondents. The development of telephone research facilitated the conduct of surveys across a wider regional or national space for an essentially lower cost.

The popularisation of telephone research was initially influenced by the development of telephone sampling. It was largely fostered by two elements: high telephone coverage in most developed countries (90–95% coverage) and access to or the possibility of the numerical reconstruction of the telephone register. A clearly defined structure of a ten-digit telephone number (in the USA at the time) enabled the development of the RDD – the random digit dial ing technique – that facilitated random sample designs within the population of telephone subscribers. Considering the large telephone coverage, this population was practically identical to the general population in a society. Moreover, the understanding of the structure of the telephone number also enabled a relatively detailed spatial allocation of respondents. Based on their telephone number, the respondents’ home address, region and certain other (administrative) spatial data can be determined, such as electoral district, electoral unit etc. Similar reasons influenced the expansion of telephone surveying in Slovenia. Telephone coverage was over 90%, while the accessibility to the telephone register in Slovenia is even simpler. Through spatial manageability and with, until recently, only one supplier of land-line telephone connections in the country, the register of telephone subscribers (physical subjects) existed all in one place – the general telephone register, that under certain conditions was also accessible through the computer. All this made sample designing much simpler compared for example to the RDD technique.

6.3 Technique

The two above mentioned elements triggered the idea about telephone research that evolved into a useful methodological practice. It was further facilitated by the development of computer technology. In parallel with the development of telephony, the use of the computer helped overcome any remaining technological deficiencies with a final affirmation of telephone research as the dominant research methodology in many fields. The use of the computer was not only welcome as an efficient analytical tool, which it undeniably was in the entire segment of data analyses, both in telephone and other forms of surveying, but was also an extremely efficient tool for the process of data acquisition and recording. In this process the role of the computer is distinctly multi-layered, beyond being simply the device that replaces paper and pencil. The computer allows the interviewers to read the

132

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism questionnaires and other instructions they need to conduct an interview, while at the same time recording (saving) the respondents’ answers. Furthermore, it allows for the adjustment of the interview’s course according to the logical sequence of questions that depend on previous answers. With the use of telephone interviewing software, the routing can be very dynamical and changeable, and also complex, but, completely manageable through the predefined structure of the questionnaire that is implemented in the mentioned software, with all the necessary rules, filters and leaps. In these circumstances, telephone interviewing has a distinct advantage over, for example, usual personal and mail interviewing. Being based on an electronic medium, telephone interviewing has another specific advantage, namely, it facilitates the random change of the routing of questions and received answers in closed and semi-closed questions. This allows the avoidance of possible bias in interviewing, due to the serial position effect (such as the primacy or recency effect). Moreover, it enables the determination of a specific set of possible questions or answers, when we deal with the needs of sample randomisation ( e.g. in oversized samples).

Perhaps the most important role of the computer in the data collecting process, provided the availability of the appropriate software, lies in its ability to control the entire client network managed by the software for the CATI management – call management. Call management involves the management and processing of the rules regarding access to respondents. Normally, these rules are predefined, but they can also be dynamically changed during the course of interviewing. The first level of call management happens at the entrance level.

In the computer network, surveys are assigned to the first free interviewer based on the priority level of the unit (the telephone number or target customer at the telephone number). The unit can be without priorities (default, pending), and is randomly selected during interviewing after the priority units are automatically distributed among the surveyors: such as appointments or assigned requirements for immediate calls that can be defined by the research co-ordinator during the course of surveying. Prior to or during the interviewing, the co-ordinator can also define quotas in accordance with the required sample characteristics. When a quota for a sub-sample ( e.g. the number of units from a certain region) is realised, the programme for the allocation of the waiting telephone numbers no longer allocates these, but gives priority to those units from the sub-samples whose quotas have not yet been achieved. The programme control of the Computer-assisted telephone interviewing enables a large number of detailed settings (CATI specifications) that determine how and under what conditions the programme will assign telephone numbers to the interviewers.

Two of such specifications that deserve to be mentioned are: the definition of the interviewers shifts, and determination of the redirections of the subsequent cal , which basically means that considering the type of interviewer (his or her

Development of a Telephone Survey within the Public Opinion 133

knowledge, experience and skill) ‘more demanding’ respondents can be assigned to more competent interviewers.

Besides the sample units access management, which is determined by survey co-ordinators and depends on the research design, key element for the successful sample realisation is also appointment management. Unlike the aforementioned CATI settings, appointments are managed by the interviewers. While co-ordinators determine the conditions for the conduct of an appointment, along with time intervals and the type of required information, interviewers record the most suitable times for repeated calls according to the respondents’ preferences.

The quality of survey realisation is primarily linked to achieving an optimum response from the respondents, which is why appointment management has an important function: on one hand it allows for the adjustment of a subsequent communication to respondent’s time preferences, and on the other it somehow binds the respondent to co-operate in at least one of the future contacts.

The whole managing procedure of the telephone research takes place within computer software and allows the determination of numerous settings and elements of control from the very beginning, which enable a higher quality of data capture. Important in this is mainly continuous monitoring and control of the dynamics and quality of the interviewers’ work (that partly also depend on the research plan and type of research). The control takes place at the level of the selection of the target households, at the level of the target persons selection procedures, in the process of the subsequent cal ing of target persons, and over the process of the conduct of the interview itself.

Monitoring at the level of the dynamics of interviewing involves the measuring of the total number of calls, the number of units according to their priority (such as, the number of the still adequate and accessible units), recorded time intervals of each operation before, between and after the survey, and the recording of the number and reasons of potential refusals.

6.4 The Practice and Characteristics of Telephone Surveying within the

Public Opinion and Mass Communication Research Centre (POMCRC) Research Programme

When speaking of the CATI systems, a necessary condition to maintain good quality is adequate technical equipment of the call centre or telephone studio, in which telephone interviews take place. In 1994 POMCRC established a telephone studio that contained, for that time, a modern computer, software and telephone equipment. In the years that followed the studio was technically upgraded in keeping with the development of information technology. Recently,

134

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism the telephone studio has included 32 telephone lines, 32 computerized work stations, a dedicated server and devices for the studio’s own network support including software for the conduct of telephone surveys.

Within its research an important part of POMCRC’s activities focused on the development of social-scientific methods in the framework of applied research.

It could be said that in this way it also developed original approaches and research methods mainly in examining public opinion about the functioning of the central state institutions. Important circumstances that functioned as salient motivators for the development of applied research at POMCRC

included the design and acquisition of the target research programme entitled Public Opinion Surveys on the Attitude of the Public towards Current Affairs and Developments in Slovenia, which took place from 1996 to 2014 (Toš 2001; Kurdija 2014). This project was designed quite extensively, while in its concrete realisation it was mainly based on the Politbarometer research that was frequently referred to by the media. With the study being developed in the period of intense social transition and in times of the assertion of political pluralism, its research data often played an important part in the preparation of strategic government decisions and measures. The study recorded the dynamics and temperature of the political, social and economic climate in Slovenian society, along with the citizens’ living conditions. This was the time when the monitoring of public reactions seemed relevant. The research of the standpoints of the citizens in transition was particularly topical after the year 2000, when Slovenia was accessing and getting ready to be integrated into the European Union and NATO.

The above mentioned set of studies (Toš 2009) includes some important research projects around which the methods of the public opinion research were shaped regarding citizens’ opinions about the following issues: topical social issues (such as obstacles in Slovenians’ employment mobility in EU countries), social and state institutions (such as Slovenian public opinion regarding the Police), and to sensitive social processes during transition (opinions about corruption).

In the context of this edition we will intentionally leave aside a large set of the Politbarometer studies (including over one hundred measurements realised), because it would need its own space in a separate publication. The methodological approach, shaped through a series of the above mentioned telephone surveys, both followed the established standards and was at the same time still specific enough to be, as it were, adapted to the Slovenian situation, in terms of either their sampling method or communicative approach, selection and training of the interviewers, as well as the design of the measuring instruments, i.e., types and kinds of questions that became part of the standard repertory in many studies conducted by the Centre. Methods and approaches grew together with the needs

Development of a Telephone Survey within the Public Opinion 135

and research goals. In time, the major part of telephone surveys became largely marked by certain regularities that were shaped through the practical methods approach, and the confirmed success of concrete research design routines.

Most studies were conducted by using the standardised questionnaire adjusted for the telephone survey. Unless it was otherwise determined in research aims and contents, the target population consisted of adult residents of Slovenia. The target persons selection procedure took place in two phases. First, the shares that reflect the geographic grid of Slovenia in terms of the size of individual inhabited localities and regional belonging were randomly selected from the list of all telephone subscribers that are ‘physical persons’. In the second phase random individuals belonging to the surveyed population (most often the legal age population of Slovenia) were selected within households. The whole procedure is based on randomness and provides the equal probability of selection (within the accessible population). If the target person was inaccessible, they could be called again during the days the survey was taking place (they could be called consecutively up to eight times). The target households selection procedure, their subsequent cal ing, and the procedure of control and management of telephone calls, appointments and interviewing are carried out with the dedicated telephone interviewing programme. The surveys were mainly conducted in the afternoon and evening, between 16.00 and 21.00, and on certain days with a reduced intensity, also in the morning. It usually took three working days per interview, and in the target population that most often included 900 to 1000

completed interviews, with a questionnaire that took between 10 to 15 minutes to answer.

The basic content design and a rough idea regarding the size of the study were often acquired early on during the call for applications or direct demand on the part of the institution who commissioned the survey (client). Often, additional co-ordination was needed that looked for an optimum relationship between the scope, the sampling design, the complexity of the contents and the given material possibilities for the conduct of the research. Where rather elaborated elements of the questionnaire’s contents were already given, such as concrete forms of questions and answers, they were always subject to a methodological reconsideration within the research group that assessed the adequacy of the proposed instrument. After the methodological revision a final proposal was shaped that was subject to re coordination with the client. Usually, the client is presented with the conditions for successful and good quality conduct of measurement (mainly in terms of total time range needed to carry out the research project), conditions of quality control, and the size and kind of research report with the presentation of the achieved results. All elements of the research cycle were mainly carried out by the research group of the POMCRC with

136

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism additional researchers participating in specific projects, in accordance with the requirements and usually within the Faculty of Social Sciences of the University of Ljubljana.

After the conclusion of interviewing, first, the coherence of the structure of the realised sample and the structure of the population was usually examined, and, if necessary, adequate ponderation and weighting were carried out. This lessens any bias created both due to the share of the population categories that did not have telephone coverage, as well as any bias due to unbalanced sample realisation in different population categories. The system of sampling weights reduces bias and result errors that can ensue from sampling, or any other errors related to the conduct of the survey. The weighting procedure belongs among the usual tools within the methodology of survey result presentation, with its employment depending on the specificities of each individual study (Kurdija, S., Štebe, J. 1997).

6.5 Quality

In its keeping with the principles of adjusting to methodological standards and based on its past experience with telephone interviewing, in the late 1990s the procedure that additionally corrects survey results was designed at POMCR.

Basically, it is the weighting procedure, recommendable, whenever, for diverse reasons, the sample characteristics do not accurately reflect the characteristics of the target population. This kind of data correction belongs among the usual tools within the methods of survey result presentation, equally applying to telephone surveying as well as other research approaches. It is requisite, wherever the access to the population is limited in any way. In this respect, telephone surveying seems particularly vulnerable (especially in recent years, with the growth of digital telephony).

Due to surveys being conducted through samples, a sampling error in the estimate of population parameters can occur, roughly for two reasons. The first is a random sampling error or random variance of a sample estimate, whose range is known from statistical sampling theory and the ensuing principles of drawing conclusions about a population from the random sample. The second reason lies in the non-sampling error or a series of potential biases that result from subjective or objective factors during the process of conducting the survey.

Some factors that can cause bias in the result of telephone surveying include the following: a random sampling error, that has already been mentioned, is the result of the probability effect in the sampling process. Usually, this deviation is expressed by the confidence interval. In a survey with a random sample of

Development of a Telephone Survey within the Public Opinion 137

approximately 1000 units, the deviation estimate moves by approximately 3 %.

This deviation can be reduced by increasing the final sample size. The confidence interval could be reduced to 1 %, if under the conditions of random sampling the survey would reach around 10,000 units. In practice, the question of compromise between cost and quality results in a reasonable sample size allowing for around 2 to 3 % of sampling error, which generally means a fairly solid sample at a reasonable cost for the survey. It needs to be pointed out, however, that this error does not express other potential errors, such as measurement errors. Related to the sampling error, although they are not the same thing, is the error in the structure of the sampling frame. It occurs due to differences, such as, between the population and population coverage. In Slovenia, telephone coverage is not completely equal in all regions, with smaller and more remote places still having less coverage (when speaking of the landline coverage), and some other (economic as well as non-economic) factors potentially influencing the possession of a landline. While in the 1990s and during transition to the new millennium, the coverage as such was no longer considered a problem, the issue re-emerged as a serious coverage-related problem with the development of digital telephony and other forms of communication that are mainly enabled by the Internet and mobile telephone network. In many places landlines were largely replaced by mobile telephones with the extent of coverage, like in IP-telephony as well, not being totally known or being largely covert. All this reveals new specific problems and, primarily, limitations in sample determination. Unlike 20 or more years ago, the development of telecommunications and the popularisation of telephone surveying find themselves in opposing developmental tendencies. While at the beginning, technological development was considered to be enabling the growth of telephone surveying, recent technological trends have raised obstacles to an efficient access to the target population. Caller identification and display, call barring, monitoring and recording of calls – which are all enabled by a common digital telephone –, largely enable the respondent to avoid survey communication. A potential respondent that, still quite recently, has been

‘forced’ to answer the ringing telephone, becomes the subject of control in their telephone communication. Moreover, the past few years have seen the negative effects of mass telephone selling and direct marketing that act as very de-motivational factors for potential respondents. This set of reasons also includes

‘abuses’ of the called individuals’ personal data that are stored in large data bases or registers forwarded to third-party clients for further commercial purposes.

The next important cause of possible biases is non-cooperation: the larger the degree of non-cooperation the larger the possibilities for sample incongruence.

The problem is even more pronounced, if due to the nature of the survey – and also in general – co-operation is typically declined by specific population categories.

138

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism This kind of selection can also contribute to changes in the characteristics of the final results. For example, the systematic nature of non-co-operation is shown in the fact that some unit types are more prone to non-cooperation than others, e.g. experience shows that men, young people and middle-aged generations, people who are economically active, and those with a higher socio-economic status, belong to the categories that prove more difficult to be accessed for surveying. Politically less informed individuals and those with more extreme political standpoints (due to the social undesirability of their standpoints) show a lower degree of willingness to co-operate. If the survey questions do not differentiate the population, there are no larger deviations; if, however, the questions differentiate the population in terms of the accessibility of specific population categories and their willingness to co-operate, then we have to think about mechanisms to alleviate the existing biases. Clearly, these errors cannot be eliminated by increasing the sample, but only through mechanisms that generally increase the degree of co-operation, and partly through weighting.

FIGURE 6.1 Response rate in scope of comparable telephone surveys



conducted in POMCR

Response rate in scope of C in the period fr comparable telephone surv om 2002 to 2013

eys

conducted in POMCRC in the period from 2002 to 2013

60%

50%

40%

30%

20%

10%

0%

2-3

2-5

2-6

2-9

-10

-12

3-2

3-4

3-6

-10

-11

-12

-03

-05

-06

-09

-10

-12

-02

-04

-06

-09

-11

-01

-04

-10

-03

-09

-12

-04

-12

-03

-01

-10

-03

-10

-01

-09

-06

200

200

200

200

02

02

03

03

03

04

04

04

04

04

04

05

05

05

05

05

06

06

06

07

07

07

08

08

09

10

10

11

11

12

12

13

20

20

200

200

200

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

20

Response rate

Linear (Response rate )

The level of co-operation can be improved by increasing the number of contacts (calls) during surveying. Since the increase in the number of calls is also related to the costs of the survey, differences in quality among research organisations can largely be explained by their willingness to invest in the element of persistence during the conduct of the research. The degree of non-cooperation related to refusals can also be reduced by training the interviewers to be even more

Development of a Telephone Survey within the Public Opinion 139

persuasive, while still being appropriate and tactful in their communication with the respondent. A better co-operation is also influenced by how they adequately present the contents of the survey, and, last but not least, by the reputation of the contractor, i.e. the research organisation. The academic environment, and scientific and expert research aims are often seen as better motivation for the respondents to co-operate than commercial ones, in which the boundary between research work and market communication or even direct telephone selling is not always clear. In this sense, the scientific and research environment at POMCRC has certain advantages, which is seen both from the respondents’

responses as well as the shares of sample realisation, when compared to certain other, commercially oriented research organisations and agencies. Regardless of all this, it needs to be pointed out that for diverse (above mentioned) reasons, researchers face a gradual but insistent overall decrease of sample realisation that is the result of, problems with access to the population on one hand, and its diminishing willingness to participate in the surveys on the other.

Sample stratification is an important mechanism for reducing bias resulting from the above mentioned circumstances. There are two modes of stratification, namely preliminary stratification and post-stratification. Preliminary stratification is optimal because it has no side effects on the increase of the variance of sample estimates. Shares of individual population categories that we wish to capture in a sample are predetermined, while the selection of individuals still remains random. This means that according to criteria number of units fit (as much as possible) the population shares in advance. The limitation is in that for the enactment of preliminary stratification we need information about the size of the selected category within the sample frame. In sampling for telephone surveys this procedure was often applied in case of the size of settlement. Preliminary stratification procedure is grounded in the expectation that the size categories of settlements in Slovenia are never accurately reflected in the respective category sizes in the telephone subscriber samples. After this correction, the sample better reflects the population, which increases its representativeness. Therefore, stratification can potentially help alleviate biases, both due to differences in the population coverage as well as the disproportion in the extent of non-cooperation within different categories that are the subject of preliminary stratification.

Following graphs (Figure 6.2) illustrate weighting correction on the two of the referential variables during and after the survey. It is evident, what was the size of the needed adjustment.





140

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism FIGURE 6.2 Illustration of weighting correction within POMCRC



telephone surveys

The Size of the Settlement

An example of the Politbarometer Survey October 2005

40

35

30

25

1. day batch (n=295)

20

%

2. day batch (n=354)

15



3. day batch (n=173

10

TOTAL - non-weighted

5

TOTAL - weighted

0

up to 500.

500-4000.

4000-50000.

over 50000

The number of inhabitants

Age

An example of the Politbarometer Survey October 2005

40

35

30

25

1. day batch (n=295)

20

%

2. day batch (n=354)

15



3. day batch (n=173

10

TOTAL - non-weighted

5

TOTAL - weighted

0

up to 30

31 - 45

46 - 60

over 61

Age group



However, categories for which there are no data within the sampling frame as such, can present problems. In our case each unit from the sampling frame is only defined by region and the size of the settlement. However, most characteristics, that can also be the subject of weighting correction, are collected subsequently through the survey itself. This is when post-stratification adjustment is applied.

The process of post-stratification adjustment of the sample is technically carried out by calculating the weights for each unit in the sample based on its belonging to a certain category. It seems best to include as many variables as possible as criteria for post-stratification. But If there are several variables in the combination of weights, the number of cells that need to be considered in determining the weights is multiplied. If too many variables have been included the number of control cells could be largely increased, and the estimates within the cells could become unreliable. This is why in practice not more than three

Development of a Telephone Survey within the Public Opinion 141

to four different variables are involved. Deficiencies due to a large number of variables are avoidable by using the raking method, where weights are corrected iteratively through a specific algorithm for several stratification variables at the same time, simultaneously retaining the highest degree possible of the relation of marginal frequencies of each of the variables. (Kurdija, S., Štebe, J. 1997) 6.6 Comparison of the Approaches

The logic of empirical research through telephone surveying is a specific one. It brings with it certain problems, but also has some advantages (if we leave aside some of the before mentioned accessibility-related reasons). The comparison of four approaches (see Table 6.1) that are the most relevant for communication with the respondent (namely: personal ‘face-to-face’ interviewing, mail interviewing, web interviewing and telephone interviewing) shows that only personal and telephone surveying can, more or less, be considered candidates for a survey result that can be acceptable in terms of quality.1 Both approaches enable a relatively good monitoring and control of the course of the study in its different phases, and also ensure an acceptable level of sample realisation.

Sometimes, the differences between both approaches in terms of quality are not that big, when compared, on the other hand, to big differences in the price of the survey. Researchers have expressed a fear that the absence of the respondent’s visual control of the questionnaire would influence their bias in responding, but their fear was fully eliminated by practice. Since the early 1990s (in developed countries) or the mid-1990s (in Slovenia), the method has become the most widespread form of data collection in terms of frequency of use, in various social-scientific, mainly applied studies.2

The advantages of computer assisted telephone interviewing are many: the possibility of making widely dispersed one-stage (probability) samples; rapidity and responsiveness – usually, such surveys are completed in a few days; adaptability –

administration and the course of the interviewing go on live and simultaneously, which enables a continuous adjustment of the elements in the study; anonymization – in some cases telephone communication has an advantage over personal communication in questions with delicate topics, because it gives the respondent the feeling of anonymity; good control – over the quality of conduct of the survey interview; still acceptable sample realisations – under the conditions 1 We speak about comparable surveys of the general population that in terms of scope and topic, are feasible with all four approaches.

2 Recently, a large number of similar surveys have been transferred to the Web.

142

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism of persistence and the use of the call management system; the speed of feedback information – the client can acquire their first impressions about the survey even during its course; a relatively low price.

TABLE 6.1

Comparison of four survey approaches: advantages and disadvantages

Type of survey

Characteristics

mail surveys web surveys telephone face-to-face surveys

surveys

Administration

price

cheaper

the cheapest lower

high

speed

the slowest

fast

the fastest

slow-moderate

length of the

moderate

shorter-

shorter

long

questionnaire

moderate

sample realisation

low

low

medium

the highest

Control over Survey

testing of the

no

yes

yes

yes

questionnaire

respondent targeting

no

no

yes

yes

change of sequence

no

yes

yes

yes

respondent’s privacy

no

partly

yes

yes

visual control of

no

no

no

yes

communication

Usefulness of Different Types of Questions

visual support

limited

limited

no

yes

open questions

limited

limited

limited

yes

conditioned questions

limited

yes

yes

yes

complex questions

limited

limited

limited

yes

delicate questions

some

some

some

some

Potential Bias

socially desirable

no

no

partly

possibly

responding

influence of the

no

no

partly

possibly

interviewer

respondent’s reading

yes

yes

no

no

skills

(Source: Neuman, 2003, updated and adapted by authors) Last but not least, when speaking of quality in telephone surveying the meaning of experience should be pointed out. We speak about experience in terms of

Development of a Telephone Survey within the Public Opinion 143

what is effective in a certain environment, and where more effort should be applied to obtain quality survey data. The many years of research practice in telephone surveying undoubtedly show that quality assurance is not dependant on the technical capacities and implementation of certain methodological standards alone. The import of specific approaches and tools that can be found in other (prominent) research is not always effective in every environment.3

The success (and therefore quality) of a survey, and the usefulness of the final results often largely depend also on the experience of the organisation that carries out the survey. The multi-layered nature of administrative mechanisms in survey management and design – which are enabled by telephone surveying as compared to other approaches4 – and rich experiences in topics covered that are enabled by the academic environment with decades of experiences, continuously confirm the relevance of accumulated knowledge in this type of work.

6.7 Conclusion

In the wider research field, especially within the sphere characterised by scholarship, telephone surveying is sometimes being reproached as unsatisfactory in relation to the established procedures in academic driven social scientific survey research that in its most reliable method still leans on face-to-face surveying.

Much has been written about the differences between, and the characteristics of, different approaches (Leeuw 2005). The failure of telephone surveying to meet the requirements in the domain of ‘big’ or ‘serious’ research have also been corroborated by studies (Martin 2011) conducted by the methodological expert group within the European Social Survey – ESS (in the framework of which POMCRC has been an active partner from the Study’s very beginning).

Understanding of the reach of telephone research is based on the aforementioned comparisons between the approaches and their methodological characteristics.

But in terms of placing telephone research in the wider field of approaches, the meaning of, and differences between the so called basic and applied research seem to be more important. Although this division has never been totally clear and indisputably definable, it is a systematic and often also formal differentiation 3 We can take the example of the use of diverse scales. In the experimental phase of the mid-1990s, when the instruments for a continuous measurement of public response to topical issues were still developed in Slovenia, it became obvious that in the Slovenian environment the effectiveness of -5 to +5 scale was not as expected, unlike the later introduced 1–5 scale that resembled the familiar school marks which everybody could understand.

4 Recently, web surveying is very much following the same pattern in this respect.

144

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism type of research. The first type mainly serves to acquire wider, general knowledge relevant to a certain scientific discipline, and the second type involves research that looks for solutions and answers to specific, problematic social issues. The first type of studies explains the basic level of social reality and basic mechanisms of social dynamics, while the second responds to concrete, practical questions.

These can include specific questions encountered by relevant state institutions (the police, the military, the health and social care institutions, educational institutions, public administration, political organisations, governmental and non-governmental organisations etc.) in their policy-making and management of different projects. These can also be concrete, topical social questions that are the subject of interest of a specific research group that conducts research within the basic research programme. There are also projects in which policy makers as well as researchers are brought together through research. This indicates the sometimes unclear differentiation in the determination or division of the basic and applied research.

In Slovenia, the interest in applied research has markedly grown during the period of its accession to the European Union, involving requests for the harmonisation of information in a large number of services that needed feedback about their work; the harmonisation of standards for writing reports, and the requisite of grounding the planned measures insisted on the increase in the number of applied research topics; state services needed relatively fast and clear answers to concrete questions that resulted in medium- and smal -scale studies. Data obtained in this way enabled practical solutions mainly for short-term projects.

For projects that measure the effects of various measures for a longer period, the clients usually formulated their demand for long-term longitudinal research.

The question that presents itself in such studies is that of the appropriate use of data by the client and especially the appropriate and proper use of data for a potential implementation of measures. The research organisation’s important role in this respect needs to be pointed out. Its duty is to offer full methodological support to enable understanding of the acquired results, and to provide, through documentation and otherwise, clarification of the specifics of the expertise of the project – which, for example define the methodology of research – to the client. The contractor’s duty is to also call attention to the limitations and scope of a concrete study, and prevent possible misinterpretations. In their hastiness and desire to obtain “clean” and indisputable solutions, clients tend to jump to too quick and the wrong conclusions or wittingly or unwittingly instrumentalise research results to achieve their ‘internal-political’ goals. As much as possible, the contractor should aim to prevent this kind of data use.

Another characteristic of telephone research worth mentioning is its information aspect. The aim of POMCRC as the contractor of a wide programme of applied

Development of a Telephone Survey within the Public Opinion 145

studies (for the needs of diverse clients) has always included the accessibility of research and data. Applied research has often been known to be aimed at a narrower circle of users, i.e. mainly the client’s analytical services that use the results to realise the aims of their own projects, while access to data by the potential wider interested public is more difficult or has to deal with certain limitations. POMCRC provides the continuing practice of informing about its research work, with a large majority of projects being taken care of in terms of the proper presentation of data and informing the public.5 Special care is devoted to the appropriate data and content storage supplied by the Social Science Data Archives (ADP) at the Faculty of Social Sciences.6 This is a transparent arrangement and storage of databases suitable for further expert and wider public use.

There is an indisputable connection between the POMCRC and the opening up of the space for telephone research in Slovenia. Experience and practice of this kind of research in connection with the network of partnerships related to POMCRC have contributed to this approach having been well established in Slovenia during a certain time in the field of important applied research.

Initial reservations in regard to the use of this approach soon yielded with the development of methods and technical opportunities in the 1990s, with some state institutions responding positively and employing the possibilities offered by this approach. Most frequently, this involved the need to obtain feedback from the population that was intended for the improvement of work within diverse programmes, or the desire to identify the response of people (or special social groups) to the concrete measures of the government or sectorial services.

The assertion of the use of telephone research has been going on during the past 20 years and through over one hundred surveys carried out during this time.

A series of diverse but important empirical studies have emerged and been realised (for the list of some of the most typical see the attachment). For a large number of projects to be carried out entails a very manifold treatment of the research subject in terms of content and methodology, and at the same time much communication with diverse and specific clients. In this frequency of work many experiences have been fortified that, together with the requirements at the executive level, largely form the basis to establish the standards in social-scientific applied research in Slovenia.

5 In 2013 the Centre received the Prometheus of Science Award for Excellence in Communication awarded by the Slovenian Scientific Foundation.

6 Social Science Data Archives at Faculty of Social Sciences, Ljubljana http://www.adp.fdv.uni-lj.si/en/

146

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism Attachment – the list of selected telephone survey studies conducted at POMCRC

1. ‘Politbarometer’ – Target research programme entitled as Public Opinion Surveys on the Attitude of the Public towards Current Affairs and Developments in Slovenia which took place from 1996 to 2014. Conducted more that hundred telephone surveys. Data files are arranged and deposited in Social Science Data Archives (ADP) at Faculty of Social Sciences in Ljubljana.

2. The public image of the Petrol company (biggest petroleum company in Slovenia) (December 1994)

3. Public view and assessment of the Slovenian Police Force (December 1995), five annual surveys.

4. Slovene Red Cross, Survey on Slovenes attitudes on humanitarian activities (May 1996)

5. A survey among shareholders of Mercator, biggest Slovenian retail chain (December 1996)

6. Public attitudes to Slovenian forests, Department of Forestry and Renewable Forest Resources (March 1999)

7. An opinion survey about the biggest Slovenian retail chain Mercator (February 2000)

8. Employee satisfaction survey in the biggest Slovenian retail chain (2001) 9. Observations on Sunday shopping (Chamber of Commerce) (2002) 10. Views on corruption in Slovenia (Commission for the Prevention of Corruption) (2002), four annual surveys.

11. Survey about the usefulness of weather forecast information (Slovenian Environment Agency, Ministry of the Environment and Spatial Planning) (2003), three surveys.

12. Opinions about wind power plants in the municipality of Ilirska Bistrica (2003)

13. Habits and the use of medication (The Health Insurance Institute of Slovenia) (2004), three surveys

14. The attitude towards genetically modified food (National institute of Public Health) (2004)

15. Survey: Parents between work and family (Ministry of Labour, Family, Social Affairs) (2004), three surveys

16. The attitude of Slovenians towards help for developing countries (Ministry of Foreign Affairs) (2005)

17. Mobility of Slovenian workers: How to find employment in the EU

(Employment service of Slovenia) (2006)

Development of a Telephone Survey within the Public Opinion 147

18. A survey on smoking and passive smoking among adults in Slovenia (2006) 19. Opinions about Students’ organisation (SOU) (2006) 20. Survey On health, mood and well-being (National institute of Public Health) (2006)

21. International research on religions – the effects of religious pluralism (NORFACE 2008)

22. Survey on drug abuse (2008)

23. Survey on Domestic Help (The Peace Institute) (2009) 24. Survey on Slovenian Railways (Slovenian Railways) 25. Customer Satisfaction study among unemployed persons in Slovenia (Employment Service of Slovenia) (2009), six surveys – mostly mail, also telephone survey.

26. Confidence in the institution of the Ombudsman (the Ombudsman) (2010)

27. Study on fertility (Social Protection Institute) (2010) 28. Survey on transition to the digital TV signal (The Information Society Directorate, Ministry of Education, Science and Sport) (2010) 29. Views about/on Pension Reform (Government Communication Office) (2011)

30. Views about Mini Jobs (Government Communication Office) (2011) 31. The views toward genetically modified organisms (Institute of Social Sciences) (2012)

32. First aid as a form of solidarity in contemporary society (Red Cross of Slovenia) (2012)

33. A survey on the impact of unemployment on the health of workers in Pomurje, University Medical Centre (2012), two surveys.

34. Public attitudes towards used and surplus products and their potential re-use (Interseroh – company for the handling of raw materials) (2014) 35. Research on Innovative living environments for older people in Slovenia, Institute for Urban Planning (2015)

References

Adam, F., et al. (2012), Kvalitativno raziskovanje v interdisciplinarni perspektivi.

Ljubljana:Inštitut za razvojne in strateške analize Al ardt, E. (1990), “Challenges for Comparative Social Research”, Acta Sociologica, vol.33, no.3, pp. 183–193

Alvesson, M., Sköldberg, K. (2000), Reflexive Methodology. New Vistas for Qualitative Research. London, New Delhi:Sage Publications Anderson, A.J. (2012), Media Research Methods. Understanding Metric and Interpretive Approaches. Sage Publications:Los Angeles Arts, W., Halman, L. (1999), “New directions in quantitative comparative sociology”, International Journal of Comparative Sociology. Vol. XL, No.1, pp. 1–12

Bang, H.P. (2004), Everyday Makers and Expert Citizens. Building Political not Social Capital. ANU Political Science Program, RSSS. At http://hdl.

handle.net/1885/42117, 19.10.2016

Barnes, S.H. (2006), “The Changing Political Participation of Post-communist Citizens”, in International Journal of Sociology, vol. 36, no. 2, pp. 76–98

Beck, U., Beck-Gernsheim, E. (2002), Individualization. London, Thousand Oaks, New Delhi:Sage

Ben-Baruch, E. (1980), “Classification, Ranking, and Measurements in the Social Sciences”, Philosophica, vol. 25, no. 1, pp. 97–116

Berger, P.L., Luckmann, T. (1988), Družbena konstrukcija realnosti. Razprava iz sociologije znanja [The Social Construction of Reality. A Treatise in the Sociology of Knowledge]. Ljubljana:Cankarjeva založba Berg-Schlosser, D., De Meur, G., Rihoux, B., Ragin, C.C. (2009), “Qualitative Comparative Analysis (QCA) as an Approach”, in Benoît Rihoux, and Charles C. Ragin (eds.), Configurational Comparative Methods: Qualitative Comparative Analysis (QCA) and Related Techniques, pp. 1–18. Los Angeles, London, New Delhi, Singapore:Sage.

Best, S., Kellner, D. (1991), Postmodern Theory. Critical Interro-gations. New York:Guilford Press

Billet, J., Koch, A., Philippens, M. (2007), “Understanding and imporving response rates”, in: Roger Jowell, Caroline Roberts, Rory Fitzgerald, and Gillian Eva (eds.), Measuring Attitudes Cross-Nationally. Lessons from the European Social Survey. Los Angeles, London, New Delhi, Singapore:Sage, pp. 113–137

Blankenship, A.B. (1977), Professional Telephone Surveys. New York:McGraw-Hil

150

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism Bourdieu, P., et al. (2002), The Weight of the World – Social Suffering in Contemporary Society. Cambridge:Polity Press Brewer, J., Hunter, A. (2006), Foundations of Multi-method Research: Synthesising Styles. Thousand Oaks, London, New Delhi:Sage Bryant, A., Charmaz, K. (2007), The Sage Handbook of Grounded Theory.

Advancing Ethnographic Research through Grounded Theory Practice. p.

493–510. DOI: 10.4135/978184607941

Charmaz, K. (2006), Constructing Grounded Theory. A Practical Guide through Qualitative Analysis. London:Sage Publications Cooper, B., Glaesser, J. (2012), “Creating Typologies: Comparing Fuzzy Qualitative Analysis with Fuzzy Cluster Analysis”, in: B. Cooper, J.

Glaesser, R. Gomm and M. Hammersley, Challenging the Qualitative –

Quantitative Divide: Explorations in Case-Focused Causal Analysis, pp.

208–239. London, New York:Continuum

Cooper, B., Glaesser, J., Gomm, R., Hammersley, M. (2012), Challenging the Qualitative – Quantitative Divide: Explorations in Case-Focused Causal Analysis. London, New York:Continuum

Dalton, R.J. (1996), Democracy and its Citizens: Patterns of Political Change.

Centre for the Study of Democracy, UC Irvine. Athttp://escholarship.org/

uc/item/9pn25985, 19.10.2016

Dalton, R.J. (2008), “Citizenship Norms and the Expansion of Political Participation”, Political Studies, vol. 56, no. 1, pp. 76–98

de Leeuw, E. (2005), “To mix or not to mix data collection modes in surveys”.

Journal of official statistics, 21(2): 233–255

Della Porta, D., Keating, M. (eds.) (2008), Approaches and Methodologies in the Social Sciences. A Pluralist Perspective. Cambridge:Cambridge University Press

Derrida, J. (1988), Glas in fenomen [The Voice and the Phenomenon]. Ljubljana: Studia Humanitatis

Derrida, J. (1997), Izbrani spisi. [Selected Essays] Ljubljana: Krt 91

Deželan, T., Fink-Hafner, D., Hafner-Fink, M., Uhan, S. (2007), Državljanstvo brez meja? [Citizenship without borders?] Ljubljana:Založba FDV

Dillman, D.A. (1978), Mail and the Telephone surveys: The Total Design Method.

New York:Wiley-Interscience

Durkheim, E. (1982), The Rules of Sociological Methods. New York, London, Toronto; Sydney:The Free Press

Eppler, M.J., Mengis, J. (2011), “Drawing Distinction: The Visualization of Classification in Qualitative Research”, =mcm working paper, No. 2/2011, July 2011. St. Gallen: =mcm institute, University of St. Gallen. At http://

www.knowledge-communication.org, 8.8.2016

References

151

European Social Survey (ESS 1), ESS7 – 2014. Fieldwork Summary and deviations.

At: http://www.europeansocialsurvey.org/data/deviations_7.html, (17. 11.

2016)

European Social Survey (ESS 2), ESS Methodology. Data Collection. At: http://

www.europeansocialsurvey.org/methodology/ess_methodology/data_

collection.html, (17. 11. 2016)

European Social Survey (ESS 3), ESS Methodology. Modes of Data Collection.

At: http://www.europeansocialsurvey.org/methodology/methodological_

research/modes_of_data_collection/, (17. 11. 2016) Fink-Hafner, D., Kropivnik, S. (2006), “Politična udeležba v posocializmu: med deformirano modernostjo, novo modernizacijo in post-modernostjo.

Družboslovne razprave”, Vol. 22, no. 51, pp. 55–72

Francois, P. (2009), Beda ideologije. [The Misery of Ideology] Le Monde diplomatique. November, 051.

Glaser, B.G., Strauss, L.A. (1967), The discovery of Grounded Theory: Strategies for Qualitative Research. New York:Aldine de Gruyter Groves, R.M., Kahn R.L. (1979), Survey by Telephone Methodology. New York: Wiley-Interscience

Groves, R.M. et al. (1988), Telephone Survey Methodology. New York:Wiley-Interscience

Hafner Fink, M. (1994), Sociološka razsežja razpada Jugoslavije (Znanstvena knjižnica, 9). Ljubljana:Faculty of Social Sciences Hafner Fink, M., Oblak Črnič, T. (2014), “Digital Citizenship as Multiple Political Participation? Predictors of Digital Political Participation in Slovenia”, Teorija in praksa, vol. 51, no. 6, pp. 1284–1303

Hafner-Fink, M., Uhan, S. (2013), “The Bipolarity and/or Duality of Social Survey Measurement Scales and the Question-order Effect”. Quality & Quantity 47 (2): 839–852.

Hair, M. (2005), “Using an Individual Differences Perspective to Investigate Context Effects in Surveys: a Test of the Belief-Sampling Model”.

International Journal of Public Opinion Research 17: 213–225

Harkness, J.A. (2007), “Improving the comparability of translations”, in: R. Jowell, C. Roberts, R. Fitzgerald, and Gillian E. (eds.), Measuring Attitudes Cross-Nationally. Lessons from the European Social Survey. Los Angeles, London, New Delhi, Singapore:Sage, pp. 79–93

Hewstone, M., Young, L. (1988), “Expectancy-value Models of Attitude: Measurement and Combination of Evaluations and Beliefs”. Journal of Applied Social Psychology 18: 958–971

152

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism Hippler, J.H. (1989), Response Order Effects in Dichotomous Questions.

The Impact of Administration Mode. Paper prepared for the meeting of American Association for Public Opinion Research.

International Social Survey Programme (ISSP 1), General information. History of the ISSP. At: http://www.issp.org/ (17. 11 2016) International Social Survey Programme (ISSP 2), General information.

Organisational structure and working principles. At: http://www.issp.org/

(PDF: 17. 11. 2016)

International Social Survey Programme (ISSP 3), General information. Publications (overview of publication trends). At: http://www.issp.org/ (17. 11. 2016) ISSP Research Group (2012), International Social Survey Programme 2004: Citizenship (ISSP 2004). GESIS Data Archive, Cologne. ZA3950 Data file Version 1.3.0, doi:10.4232/1.11372

ISSP Research Group (2016), International Social Survey Programme: Citizenship II – ISSP 2014. GESIS Data Archive, Cologne. ZA6670 Data file Version 2.0.0, doi:10.4232/1.12590

Jöreskog, K.G., Sörbom, D. (2001), LISREL 8.51. Scientific Software International, Inc

Johnson, T.P. (1998), “Approaches to Equivalence in Cross-Cultural and Cross-National Survey Research”, in: J. Harkness (ed.), Cross-Cultural Survey Equivalence (ZUMA-Nachrichten Spezial Band 3), pp.1–40. Mannheim: ZUMA. At: http://www.gesis.org/fileadmin/upload/forschung/publikationen/zeitschriften/zuma_nachrichten_spezial/znspezial3.pdf, 16.11.2016

Johnson, T.P., O’Rourke, D., Severns, E. (2009), “Effects of Question Context and Response Order on Attitude Questions”, in: 1998 Proceedings of the Section on Survey Research Methods pp: 857–860. American Statistical Association, Alexandria, VA. Dostopno preko http://www.amstat.org/

sections/srms/proceedings/. 14. 11. 2009

Jowell, R., Kaase, M., Fitzgerald, R., Gillian, E. (2007), “The European Social Survey as a measurement model”, in: R. Jowell, C. Roberts, R. Fitzgerald, and Gillian E. (eds.), Measuring Attitudes Cross-Nationally. Lessons from the European Social Survey. Los Angeles, London, New Delhi, Singapore:Sage, pp. 1–31

Joy, D., Sapin, M. (eds.) (2016), ISSP 2014 – Citizenship II. Study Monitoring Report.

Lausane: Swiss Centre of Expertise in the Social Sciences – FORS, University of Lausane. A: http://www.gesis.org/issp/modules/issp-modules-by-topic/

citizenship/2014/, (PDF: 17. 11. 2016)

Kembel, D. (2005), Lažljivčeva priča. Istorija neistine. [Campbell, Jeremy. The Liar’s Tale: A History of Falsehood] Beograd:Biblioteka XX vek

References

153

Koch, A., Halbherr, V., Stoop, I.A.L., Kappelhof, J.W.S. (2014), Assessing ESS

sample quality by using external and internal criteria. Mannheim: European Social Survey, GESIS. At: http://www.europeansocialsurvey.org/ (PDF: 17.11.2016)

Kohn, M.L. (1989), “Introduction”, in: M.L. Kohn (ed.), Cross-national research in sociology. Newbury Park, London, New Delhi:Sage, pp. 17–31

Kripendorff, K. (2005), “The Social Construction of Public Opinion”, in: Kommunikation über Kommunikation. Theorie, Methoden und Praxis.

Festschrift für Klaus Merten, 129–149. Retrieved from http://repository.

upenn.edu/asc_papers/75 (28.6.2017)

Krosnick, A.J. (1992), “The Impact of Cognitive Sophistication and Attitude Importance on Response Order and Question-Order Effects”, in: Schwartz, N., Sudman, S. (eds.), Context Effects in Social and Psychological Research pp: 203–218. New York:Springer Verlag

Kurdija, S., Štebe, J. (1997), Oblikovanje vzorčnih uteži za telefonsko anketo Politbarometer, Fakulteta za družbene vede, Ljubljana Kurdija, S., et al. (2014), Politbarometer 06/14, Slovenija: junij 2014. Ljubljana: Fakulteta za družbene vede, Arhiv družboslovnih podatkov, 2014

Lazarsfeld, P. (1958), “Evidence and Inference in Social Research”, Daedalus, Vol. 4, no. 4, pp. 99–130

Leeuw, E.D. (2005), “To Mix or Not to Mix Data Collection Modes in Surveys”, Journal of Official Statistics, Vol. 21, No. 2, 2005, pp. 233–255

Li, Y., Marsh, D. (2008), “New Forms of Political Participation: Searching for Expert Citizens and Everyday Makers”, British Journal of Political Science, vol. 38, no. 2, pp. 247–272

Li, T., Wehr, K. (2007), The Validity of the Factor Structure of the General Social Survey Environmentalism Scales across Gender and Ethnicity in the United States. Organ. Environ. doi:10.1177/1086026607306461

Linssen, R., Schmeets, H., Scheepers, P., de Grotenhuis, M. (2015), “Trends in Conventional and Unconventional Political Participation in Europe, 1981–2008”, in: C. Eder, I.C. Mochmann and M. Quandt (eds.), Political Trust and Disenchantment with Politics, pp. 31–58, Leiden, Boston: Brill MacIntosh, R. (1998), “A Confirmatory Factor Analysis of the Affect Balance Scale in 38 Nations: a research note”. Soc. Psychol. Q. 61: 83–91

Margolis, J. (1989), “Postscript on Modernism and Postmodernism, Both”.

Theory, Culture & Society. 6 (1): 5−30. London:Sage Marradi, A. (1990), “Classification, Typology, Taxonomy“, Quality & Quantity, 24(2): 129–157.

Marsh, D. (2011), “Late Modernity and the Changing Nature of Politics: Two Cheers for Henri Bang”, Critical Political Studies, vol. 5, no. 1, pp. 73–89

154

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism Marsh, W.H., Hocevar, D. (1985), “Application of Confirmatory Factor Analysis to the Study of Self-Concept: First- and higher order factor models and their invariance across groups”. Psychol. Bull 97: 562–582

Martin, P. (2011), What makes a good mix? Chances and challenges of mixed mode data collection in the ESS. Centre for Comparative Social Surveys Working Paper Series no. 02. London:CCSS, City University Marvasti, B.A. (2004), Qualitative Research in Sociology. London, New Delhi: Sage Publications

Milbrath, L. (1965), Political Participation. How and Why Do People Get Involved in Politics? Chicago:Rand McNally College Publishing Company Newton, K., Montero, J.R. (2007), “Patterns of Political and Social Participation in Europe”, in: R. Jowell, C. Roberts, R. Fitzgerald, Gillian E. (eds.), Measuring Attitudes Cross-Nationally. Lessons from the European Social Survey, pp. 205–237. Los Angeles, London, New Delhi, Singapore:Sage Neuman, W.L. (2003), Social Research Methods: Qualitative and Quantitative Approaches. Boston:Pearson Education, Inc Neumann, E.N. (1993), The Spiral of Silence. Chicago and London:The Univerity of Chicago Press

Nowak, S. (1989), “Comparative studies and social theory”, in: M.L. Kohn (ed.), Cross-natioanl research in sociology. Newbury Park, London New Delhi:Sage, pp. 34–56

Ostrom, M.T., Betz, A.M., Skowronski, J.J. (1992), “Cognitive Representation of Bipolar Survey Items”, in: N. Schwarz and S. Sudman (eds.), Context Effects in Social and Psychological Research pp: 297–315. New York:Springer-Verlag Øyen, E. (1990), “The imperfection of Comparisons”, in: V.E. Øyen (ed.), Comparative Methodology. Theory and Practice in International Social Research. London, Newbury Park, New Delhi:Sage, pp. 1–18

Pawson, R. (1989), A Measure for Measures: a Manifesto for Empirical Sociology.

London and New York:Routledge

Pennings, P., Keman, H., Kleinnijenhuis, J. (2006), Doing Research in Political Science. London, Thousand Oaks, New Delhi:Sage Popper, R.K. (1998), Logika znanstvenega odkritja [The Logic of Scientific Discovery]. Ljubljana:Studia Humanitatis

Przeworski, A., Teune, H. (1970), The logic of comparative social inquiry. New York:John Wiley & Sons (Reprint: Malabar, Florida: Robert E. Krieger Publishing Company, 1985)

Ragin, C.C. (1987), The Comparative Method: Moving Beyond Qualitative and Quantitative Strategies. Berkeley, Los Angeles, London:University of California Press

References

155

Ragin, C.C., Amoroso, L.M. (2011), Constructing Social Research: The Unity and Diversity of Method, (2nd ed.), Thousand Oaks, London, New Delhi, Singapore:Pine Forge Press, Sage

Rihoux, B., Ragin, C.C. (eds.) (2009), Configurational Comparative Methods.

Qualitative Comparative Analysis (QCA) and Related Techniques. Los Angeles, London, New Delhi, Singapore:Sage

Root, M. (1993), Philosophy of Social Science: The Methods, Ideals, and Politics of Social Inquiry. Oxford UK and Cambridge USA:Blackwell Sabucedo, J.M., Arce, C.(1991), “Types of Political Participation: A Multidimensional Analysis”, European Journal of Political Research, vol.20, no. 1, pp. 93–102

Scheuch, E.K. (1968), “The cross-cultural use of sample surveys: problems of comparability”, in: Stein Rokkan (ed.), Comparative research across Cultures and Nations. Paris, The Hague:Mouton, pp. 176–209

Scheuch, E.K. (1989), “Theoretical Implications of Comparative Survey Research: Why the Wheel of Cross-cultural Methodology Keeps on Being Reinvented”, International sociology, Vol. 4, No. 2: 147–167

Scheuch, E.K. (1993), “The Cross-cultural Use of Sample Surveys: Problems of Comparability”, in: Historical Social Research, vol. 18,no. 2, pp. 104–138

Scholz, E., Heller, M. (2009), ISSP Study Monitoring 2007. Report to the ISSP General Assembly on monitoring work undertaken for the ISSP (GESIS-Technical Reports 2009|05). Mannheim: GESIS – Leibniz-Institut für Sozialwissenschaften.

At: http://www.gesis.org/issp/modules/issp-modules-by-topic/leisure-time-sports/ (PDF: 17. 11. 2016)

Schuman, H., Presser, S. (1981), Questions and Answers in Attitude Surveys: Experiments in Question Form, Wording, and Context. New York:Academic Press

Schwartz, N., Sudman, S. (eds.) (1992), Context Effects in Social and Psychological Research. New York:Springer-Verlag

Silverman, D. (1993), Interprating Qualitative Data. Methods for Analysing Talk, Text and Interaction. London:Sage Publications Smith, T.W. (1988), Ballot Position: An Analysis of Context Effects Related to Rotation Design (GSS Methodological Report No. 55). Chicago:NORC

Smith, T.W. (2009), “The ISSP: History, Organization and Members, Working Principles and Outcomes. An historical-sociological account”, in: M.

Haller, R. Jowell, and T.W. Smith (eds.) (2009), The International Social Survey Programme, 1984–2009. Charting the Globe. London and New York:Routledge, pp. 2–27

Social Science Data Archives, Faculty of Social Sciences, (ADP), http://www.adp.

fdv.uni-lj.si/en/

156

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism Stempel, G., Westley, B. (1989), Research methods in mass communication. Englewood Cliffs:Prentice Hal

Stoop, I. (2007), “If it bleeds, it leads: the impact of media-reported events”, in: R. Jowell, C. Roberts, R. Fitzgerald, and Gillian E. (eds.), Measuring Attitudes Cross-Nationally. Lessons from the European Social Survey. Los Angeles, London, New Delhi, Singapore:Sage, pp. 95–111

Strauss, A. (1987), Qualitative Analysis for Social Scientists. Cambridge:Cambridge University Press

Strauss, A., Corbin, J. (1998), Basics of Qualitative Research. Techniques and Procedures for Developing Grounded Theory, second edition. Thousand Oaks:Sage Publications

Sudman, S., Bradbourn, N.N., Schwartz, N. (1996), Thinking about Answers: The Application of Cognitive Processes to Survey Methodology. San Francisco:Jossey-Bass Publishers

Tashakkori, A., Teddlie, C. (1998), Mixed Methodology: Combining Qualitative and Quantitative Approach. London, New Delhi:Sage Teddlie, C., Tashakkori, A. (2009), Foundations of Mixed Methods Research.

Thousand Oaks:Sage

Teune, H. (1990), “Comparing countries: Lessons learned”, in: E. Øyen (ed.), Comparative Methodology. Theory and Practice in International Social Research. London, Newbury Park, New Delhi:Sage, pp. 38–62

Toš, N., et al. (2001), Politbarometer PB9/01, Slovenija: september 2001. Ljubljana: Fakulteta za družbene vede, Arhiv družboslovnih podatkov, 2006

Toš, N., et al. (2004a), Slovenian Public Opinion Survey 2003/3: National and International Security; Military Profession [codebook file]. Ljubljana: University of Ljubljana, Faculty of Social Sciences, Public Opinion and Mass Communication Research Centre [production]. Ljubljana:University of Ljubljana, Social Science Data Archive [distribution]. URL: http://www.

adp.fdv.uni-lj.si/opisi/sjm034.xml (accessed: June 2005) Toš, N., et al. (2004b), Slovenian Public Opinion Survey 2003/4: Citizenship /

International Survey ISSP [codebook file]. Ljubljana:University of Ljubljana, Faculty of Social Sciences, Public Opinion and Mass Communication Research Centre [production]. Ljubljana:University of Ljubljana, Social Science Data Archive [distribution]. URL: http://www.adp.fdv.uni-lj.si/opisi/sjm034.xml, (accessed: June 2005)

Toš, N., et al. (2005), Politbarometer 10/05, Slovenija: oktober 2005. Ljubljana: Fakulteta za družbene vede, Arhiv družboslovnih podatkov, 2005

Toš, N., et al. (2009), Vrednote v prehodu IV. Ljubljana:Fakulteta za družbene vede, IDV

References

157

Tourangeau, R., Rasinski, K. (1988), “Cognitive Processes Underlying Context Effects in Attitude Measurement”. Psychological Bulletin 103: 299–314

Tourangeau, R., Rips, J.L., Rasinski, K. (2000), The Psychology of Survey Response.

New York:Cambridge University Press

Tourangeau, R., Singer, E., Presser, S. (2003), “Context Effects in Attitude Surveys.

Effects on Remote Items and the Impact on Predictive Validity”. Sociological Methods & Research 31: 486–513

Uhan, S. (1998), Prava in neprava mnenja. [Correct and Incorrect Opinions]

Ljubljana:Znanstvena knjižnica (Scientific Library) Faculty of Social Sciences

Vattimo, G. (2004), Filozofska karta 20. stoletja: Tehnika in eksistenca

[Philosophical Map of the 20th Century: Technique and Existence].

Ljubljana:Založba Sophia publishing house

van de Vijver, F.J.R. (1998), “Towards a Theory of Bias and Equivalence”, in: Janett Harkness (ed.), Cross-Cultural Survey Equivalence (ZUMA-Nachrichten Spezial Band 3), pp. 41–65. Mannheim: ZUMA. At: http://www.gesis.

org/fileadmin/upload/forschung/publikationen/zeitschriften/zuma_

nachrichten_spezial/znspezial3.pdf, 16.11.2016

van Deth, J.W. (2014), “A Conceptual Map of Political Participation”, ActaPolitica, vol. 49, no. 3, pp. 249–367

Vezovnik, A. (2008), Kritična diskurzivna analiza v kontekstu sodobnih diskurzivnih teorij [Critical Discursive Analysis in the Context of Contemporary Discursive Theories]. Družboslovne razprave Journal, XXV

(60): 27−45

Ward, J.H. Jr. (1963), “Hierarchical Grouping to Optimise an Objective Function”. Journal of the American Statistical Association, vol. 58, no. 301, pp. 236–244

Willis, G.B. (2015), Analysis of the cognitive interview in questionnaire design.

Oxford, New York:Oxford University Press

Zaller, J., Feldman, S. (1992), “A Simple Theory of the Survey Response: Answering Questions Versus Revealing Preferences”. American Journal of Political Science 36: 579–616

Index

coding 24, 25, 45, 62

A

Communication Research 35, 38

Adam, F. 15, 19, 29

Communication Research Centre 9,

Allardt, E. 75

109, 127, 129, 133

Alvesson, M. 15, 20, 21, 22, 23, 24, 25,

comparative approach 11, 55, 75

26, 27, 28, 32, 35, 36, 37, 38, 39,

complementarity 10, 19

41, 42, 43, 46, 48, 49, 50

Computer Assisted Telephone

Amoroso, L.M. 27, 43, 45, 59, 60, 104



Interviewing (CATI) 129, 141

Anderson, J.A. 37, 49, 51

constructivist paradigm 11

Argentina 90

context effect 113, 124

Asian Barometer 55

Corbin, J. 21

Austria 64, 65, 68, 70, 71, 72, 88, 89

cross-national comparison 10, 11, 55, 61,

64, 73

B

D

Beck-Gernsheim, E. 62

Beck, U. 62

Dalton 150

Belgium 89, 103

deconstruction 11, 38, 39, 40, 41, 42, 51

Berger, P.A. 9

Denmark 64, 65, 68, 70, 72, 74, 89

Berg-Schlosser, D. 59

Derrida, J. 11, 38, 39, 40, 41, 42, 43, 48

Bipolar Survey Scale 115

Deželan, T. 62, 102

Blankenship, A.B. 129

Dillman, D.A. 129

Bourdieu, P. 22, 23, 31, 40, 45, 46, 51

Bray, Z. 26, 27, 46

E

Brewer, J. 59

Eastern Europe 64

Bryant, A. 25, 28, 29, 30

Eastern Germany 104

Bulgaria 64, 65, 68, 70, 89

epistemological gap 23

equivalence 36, 55, 61, 64, 65, 66, 68, 73,

C

75, 77, 78, 87, 88, 90, 92, 93, 94,

Campbell, J. 38, 48

95, 97, 98, 99, 100, 101, 106, 126

care 22, 100, 144, 145

European Social Survey (ESS) 55, 76,

categories 23, 24, 25, 26, 30, 43, 46, 47,

88, 89, 97, 143

60, 61, 67, 68, 69, 71, 74, 76, 95,

European Union 134, 144

96, 116, 126, 136, 137, 138, 139,

European Values Study (EVS) 76, 88

140

Charmaz, K. 20, 25, 28, 29, 30

F

Chile 89, 90

Family 90, 146

China 89, 90

Feldman, S. 113

citizenship 62, 102, 168

Fink Hafner, D. 62

160

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism Flanders 103

L

Foucault, M. 38

lagged modernization 63

France 64, 65, 68, 70, 72, 74, 82, 89

Latin America 64

Francois, P. 49

Leeuw, E.D. 10, 143

G

level of generality 22

levels of research 59

gender roles 43

Li, Y. 62, 120

Glaser, B.G. 19, 20, 21, 22, 23, 24, 28, 29

logocentrism 41

Grounded Theory (GT) 9, 10, 15, 19,

Luckmann, T. 9

22, 28, 29, 31

Lyotard, J.F. 38, 42

Groves, R.M. 129, 130

M

H

Margolis, J. 43

Hafner Fink, M. 62, 104

Marradi, A. 60, 78

Hal , S. 51

Marsh, D. 62, 120

health 130, 144, 147

Martin, P. 143

Health and Health Care 90

Marvasti, B.A. 21

Heller, M. 97, 98

Mexico 89, 90

Hunter, A 59

Montero, J.R. 62

I

N

ideal type 59

National Identity 90

indexicality 31

NATO 134

India 89, 90

Newton, K. 62

inductive ethnography 10, 15, 19, 26

inductive logic 61, 63

O

International Social Survey Programme

Oblak Črnič, T. 62

(ISSP) 11, 55, 63, 88, 105

Ostrom, M.T. 115, 116, 117, 118, 119

interpretative approach 11

interpretative richness 32

P

J

Pennings, P. 59

phonocentrism 41

Johnson, T.P. 10, 61, 64, 93, 113

Poland 64, 65, 68, 70, 83, 89

K

Politbarometer 134, 146

political culture 72

Kembel, D. 42, 48

political participation 11, 43, 61, 62, 63,

Kohn, M. 104

64, 65, 66, 67, 68, 69, 70, 71, 73,

Krippendorff, K. 9

74, 75, 76, 77, 78, 79, 81, 102

Krosnick, A.J. 10, 113

Index

161

Popper, K.R. 29

Sköldberg, K. 15, 20, 21, 22, 23, 24, 25,

Portugal 64, 65, 68, 70, 89

26, 27, 28, 32, 35, 36, 37, 38, 39,

positivist paradigm 11

41, 42, 43, 46, 48, 49, 50

post-structuralism 36, 37, 38, 39, 40, 43,

Slovenia 64, 65, 68, 70, 89, 102, 103,

44, 46, 48, 50

118, 119, 131, 134, 135, 137,

power asymmetry 46

139, 141, 143, 144, 145, 146, 147

Presser, S. 10, 113, 114

Slovenian Public Opinion Survey (SJM)

protectionism 118, 119, 121

109, 117

Przeworski, A. 92

Smith, T.W. 55, 88, 94, 105, 113, 114,

Public Opinion and Mass 9, 109, 127,

125

129, 133

Social Inequalities 100, 90

Social Networks 89

Q

Social Science Data Archives (ADP)

question-order effect 113

145, 146

Social Survey 11, 55, 63, 76, 85, 88, 89,

R

97, 101, 105, 114, 143

Stempel, G. 35

Ragin, C.C. 10, 27, 43, 45, 55, 59, 60,

Strauss, A.L. 19, 20, 21, 22, 23, 24, 25,

104

28, 29, 38

Rasinski, K. 10, 113, 114

subjectivity 38, 42

reality 9, 19, 24, 29, 32, 35, 37, 39, 40,

Sudman, S. 10, 113

43, 44, 46, 47, 48, 49, 50, 144

Sweden 64, 65, 68, 70, 72, 74, 89

reduction 43, 45

reflexive methodology 15, 35, 36, 40,

T

44, 51

reflexivity 36, 37

Taiwan 89, 104

reliability 77, 92, 93, 95, 101

Tashakkori, A. 59

representation 44, 46, 47, 109, 113, 115

Teddlie, C. 59

Rips, J.L. 113, 114

Telephone Survey 135, 146, 147

Role of Government 89, 95, 96

Teune, H. 90, 92, 100

Toš, N. 118, 129, 130, 134

S

Tourangeau, R. 10, 113, 114, 125

sampling 25, 92, 97, 100, 103, 129, 131,

U

134, 135, 136, 137, 139, 140

Scheuch, E.K. 64, 94, 102, 103

Uhan, S. 10, 44, 113, 117, 125, 126

Scholz, E. 97, 98

Uruguay 90

Schuman, H. 10, 113, 114

USA 64, 65, 68, 69, 70, 71, 72, 74, 88,

Schwartz, N. 10

89, 98, 129, 130, 131

Silverman, D. 26, 31

162

Social Research: From Paradigmatic Divide to Pragmatic Eclecticism V

van de Vijver, F.J.R. 61, 64, 67, 94

Vattimo, G. 29, 38, 42

Venezuela 64, 65, 68, 70, 71, 89, 90

Vezovnik, A. 36

W

Wallonia 103

Ward, J.H. 70

Westley, B. 35

Willis, G.B. 9

Work Orientations 90

World Values Survey (WVS) 55, 76, 88

X

xenophobia 118, 119, 121

Y

Yugoslavia 103

Z

Zaller, J. 113





Document Outline


Blank Page