Original scientific article UDK 316.62:[324:342.511](4974)"1997/2012" Mitja Hafner-Fink, Simona Kustec Lipicer, Danica Fink-Hafner, Gregor Čehovin DETERMINANTS OF VOTER BEHAVIOUR AT PRESIDENTIAL ELECTIONS IN A POST-SOCIALIST CONTEXT: THE CASE OF SLOVENIA ABSTRACT: This article attempts to fill the gap in the literature regarding the determinants ofvoter behaviour at presidential elections in a post-socialist context, taking into account the changing patterns of competition both within the party system (parliamentary elections) and at presidential elections. Slovenia is taken as a case study of predictors of voting behaviour at presidential elections (1997, 2002, 2007 and 2012) based on bivariate and multivariate methods using Politbarometer public opinion survey data. The keyfindings are that the predictors of voting behaviour change over time and that party identification has recently been declining in importance as a predictor of voting behaviour at presidential elections. KEY words: voting behaviour, presidential elections, parliamentary elections, logistic regression model, Slovenia Določnice obnašanja volivcev na predsedniških volitvah v posocialističnem kontekstu: Primer Slovenije izvleček-, članek prispeva k zapolnjevanju primanjkljaja v literaturi o določnicah volilnega obnašanja na predsedniških volitvah v posocialističnem kontekstu ob upoštevanju vzorcev tekmovanja na parlamentarnih in predsedniških volitvah. Študija primera Slovenije zajema dejavnike volilnega obnašanja na predsedniških volitvah (1997, 2002, 2007 in 2012) na podlagi analize podatkov splošne javnomnenjske anketne raziskave Politbarometer z bivariatnimi in multivariatnimi metodami. Ključne ugotovitve so, da se dejavniki volilnega obnašanja spreminjajo skozi čas in da strankarska identifikacija v zadnjem času izgublja pomen pri volilnem obnašanju na predsedniških volitvah. ključne BEsEDE: volilno obnašanje, predsedniške volitve, parlamentarne volitve, logistični regresijski model, slovenija 1. Introduction There is plenty of political science literature that examines voter behaviour as one of the central characteristic of any liberal-democratic political system (for an overview see Bartolini and Mair 1990; Niemi and Weisberg 2001; Dalton and Wattenberg 2000; Dalton and Klingemann 2007). For more than half of a century, different explanations as well as constitutive elements of voting behaviour have been evolving as a result of two interconnected processes: a) the marked growth in the collection and analysis of data on voter behaviour; and b) differences in types of political systems, elections, maturity of democratic electoral experience all over the world globe (Dalton and Klingemann 2007: 3). Shortly after the introduction of the democratisation processes and the first democratic elections held in Europe's former socialist systems, selected aspects of voting behaviour have also became popular research filed in these new democratic regimes (see Kitschelt and Hellemans 1990; Kitschelt et al. 1999; Fuchs and Klingemann 2002; Wessels and Klingemann 2006; Van Der Brug et al. 2008).1 Among the most popular topics in the field have been the role played by the various types of ideological division and the intervention of institutional engineering in the political sphere. The main focus of our paper is the voting dynamic at a presidential election in a post-socialist parliamentary system. In this context we are primarily interested in the influence of a left-right (party) identification of voters. At the same time, we will also take into account the possible interplay between the patterns of competition in parliamentary and in presidential elections. By focusing on Slovenia as a case study, we aim to address a gap in the current literature - namely, a lack of research into voter behaviour in a parliamentary constitutional democracy with a weak president and a fragmented party system. Our hypothesis is this: since 2004, the shift from a tripolar (Left, Centre, and Right) system of party competition in parliamentary elections toward a bipolar party system (Centre-left vs. Centre-right) in Slovenia has also impacted on the structure of the competition at presidential elections. More precisely, we will test the public perception that, in times of a tri-polar competition in the party system, voters select the president within the framework of a bipolar competition (Centre-left vs. Centre-right), while in circumstances of a bipolar party system competition and when selecting a presidential candidate in the second round of presidential elections (effectively selecting between the left and centre oriented presidential candidate, the right candidate having been eliminated) voters predominantly opt for the candidate who is positioned in the centre. We will test the hypothesis by comparatively analysing the determinants of voter self-identification, both social and political (including the left-right party identification), 1. As Kubicek (2000: 297) has observed, 28.8 % of all articles on post-socialist politics and society have focused on political institutions and parties, 23.8 % on political culture, 15 % on transition issues, 12.5 % on political economy, 8.8 % on political/voting behaviour, 5 % on nationalism or regional politics, and 6.3 % on other topics. at the 1997, 2002, 2007 and 2012 presidential elections. We will use bivariate and multivariate methods for analysing the Politbarometer survey data.2 In the following sections, we will first present an overview of the existing literature. This will be followed by the case study of Slovenia, including the contextualisation, the description of the methodology and the results of the quantitative analysis of the Politbarometer data. Our main findings are summarised in the conclusion. 2. Determinants of voter behaviour and the gaps therein In the Western context, since the mid-1950s, the study of voting behaviour has proliferated within three predominant schools. The first among them is the socio-economic or Columbian School that emphasises the relevance of determinants like gender, age, class status, occupation, ethnicity, later also stacked social and value divisions on voting behaviour and choice (Berelson et al. 1954; Lazarsfeld et al. 1944; Lipset and Rokkan 1967; Inglehart and Flanagan 1987). As another school of voting behaviour the psychological theory of party identification of the Michigan School emphasises the crucial role of the determinants of various political relevant determinants to have an impact on voter choice. Among the exposed the analysis of past government activities, election campaign and preferences, candidate selection processes and other issues relevant for the election processes can be recognised (Campbell et al. 1960; Bartels 2000). The rational choice theory as the third of the most applied voting behaviour theories that in a way evolved out of the Michigan school understands political engagement as a consequence of parties and/or citizens' rational assessments of their mutual ideological closeness as it is based on the parties' ideologies and stances on various issues (Downs 1957; Tsebelis 1990; Dowding 1994; Brennan and Hamlin 1998). While the approaches of the presented schools assumed that voters are to at least some extent responsible and predictable in their voting behaviour, contemporary critics of those schools emphasise the on-going relevance of short-term and highly subjective determinants, like the role of charismatic leaders, political affairs or the role of the media are (see Popkin 1994; Lupia and McCubbins 1998; Norris et al. 1999). Parallel to all the exposed also the so called institutional theories of voting behaviour are on the other hand of crucial importance, pointing to the decisive role of the institutional relevant factors, such as the prevailing political and institutional order, the diverse institutional settings, and the electoral systems of the various elected bodies in a particular country (or in cross-country or cross-electoral comparisons) (Lipset and Rokkan 1967, 1990; Linz 1994; Cheibub et al. 1996; LeDuc et al. 2002; Hellwig and Samuels 2007). When it comes to the post-socialist context, the stock of studies on voting behaviour is smaller. While most research has been built on conceptualisations developed in the older democracies (Kitschelt and Hellemans 1990; Mishler and Rose 1994; Tucker 2002; Gunther et al. 2006; Klingemann et al. 2006), the research does reveal certain 2. The Politbarometer is conducted by the Public Opinion and Mass Communication Research Centre at the Faculty of Social Sciences, University of Ljubljana. post-socialist idiosyncrasies. There appear to be a number of determinants of voting behaviour, including the following: psychological variables (Inglehart and Catterberg 2003); the idiosyncrasies of party partisanship and idiosyncratic political divisions (Kitschelt et al. 1999; Van Der Brug et al. 2008; Tavits and Letki 2009; Lewis and Markowski 2011); candidate image and issue orientations; mass media influences; the emergence of strategic voting (Dalton and Klingemann 2007; Rupnik and Zielonka 2012); as well as institutional variables (Lijphart 1992; Birch 2003; Dalton et al. 2003). Additionally, the institutional, political and social idiosyncrasies of individual countries allow for interesting cross-country research (Karp and Banducci 2007; Tavits 2009). Since the time of the French Revolution, ideological opinions have often been understood in terms of a single and enduring political left-right ideological divide (Fuchs and Klingemann 1990; Huber and Inglehart, 1995; Budge et al. 2001; Jost et al. 2009) that has been amended since the 1970s by the emerging relevance of other types of non-political left-right identities. For instance, religious, value, and even educational divisions (Inglehart and Flanagan 1987). At least in the case of Western democracies, the studies that are based on socio-economic left-right policy positions still prevail and can be found in various research approaches that have been conducted all over the world. These also include the post-socialist region (e.g. Whitefield and Rohrschneider 2009) - although the characteristics of the left-right divisions in post-socialist environments and in Western democracies have proved to be different (Marks et al. 2006; Rovny and Edwards 2012). Furthermore, left-right divisions in post-socialist countries have been found to be based primarily on symbolic issues and identities (Kitschelt and Hellemans 1990; Evans and Whitefield 1998; Kitschelt et. al 1999; Fuchs and Klingemann 2002; Klingemann et al. 2006, Wessels and Klingemann 2006; Enyedi 2008; Tavits and Letki 2009). In this article, we reveal how the choice of candidates at a presidential election is affected by voters' left-right party affiliation. 3. The determinants of voter behaviour at presidential elections in Slovenia 3.1 Contextualisation By constitutional choice, Slovenia is a parliamentary republic in which the President of the Republic plays only a ceremonial role. Although Slovenia's president is the weakest of the post-socialist countries in the EU (Dubrovnik 2013), the role nevertheless has held a great deal of political symbolism since Slovenia's transition to democracy (see more in Fink-Hafner 2010). The fact that the president is elected directly based on a majoritarian electoral system encourages a rather plebiscitary selection of the 'father' of the nation. Although it is the individual candidates who compete, the contest comes down to political party competition due to the endorsement that individual presidential candidates receive from individual parties or party-blocks (as created in the party system arena). Since parliamentary elections have - since democratic transition - been based on a proportional electoral system, the party system has remained rather fragmented while the major political divisions have evolved in such a way that particular ideological clusters of parties have been created within the framework of party competition. So far, party-system competition in parliamentary elections has shifted from bipolar during the transition period (the 1990 elections) to tripolar competition (1992, 1996, 2000 elections), and back to bipolar competition (2004, 2008, 2011) (see more in Lewis and Markowski 2011) while an analysis of the determinants of political participation has revealed the increasing detachment of citizens from party identification (Hafner-Fink et al. 2011). Although a bipolar pattern of electoral competition has only prevailed in Slovenia since 2004, left-right divisions have persisted since Slovenia's transition to democracy. First of all, these divisions are constructed at the symbolic level, such as the clerical-liberal division over religious matters; as well as the communism-anti-communism division over the interpretation of recent national history (Fink-Hafner 2010). In fact, the enduring historical divide between conservatism and liberalism from previous centuries has now been overlaid and further enhanced by the communist-anti-communist division. Nevertheless, consistently inefficient governments that have been characterised by constant ideological conflict have lately brought about a sharp decrease in the public's trust in political parties in general, regardless of their ideological orientation. This factor, along with the mismanagement of the financial and economic crisis in Slovenia, led to both the 2011 preterm elections, as well as created a diverse wave of new parties that have been critical of the established core parties. Indicatively, political (party) divisions, in terms of support for the presidential candidates, have so far been bipolar (1992, 1997, 2002, and 2007) with the only exception being the shift to the tripolar competition during the 2012 presidential elections. So far, none of the candidates from the right or right-center has been elected. Of the elections included in the analysis, only the 1997 presidential elections were decided by the first round.3 The winner (from among eight candidates) was the incumbent Milan Kučan (with 55.6 percent of the vote), who had held his first term of office 1992-1997. Although proposed as an independent candidate, as former leader of Slovenian communists, who had left the Yugoslav League of Communists just in the time before the final disintegration of Yugoslavia and the independence of Slovenia, he was indisputably the political left's main candidate. Three main centre-right parties proposed two candidates to challenge Kučan: Janez Podobnik and Jožef Bernik, who between them won 28.6 percent of votes. The remaining five candidates gathered a little less than 17 percent.4 The 'left-right' division was evidence at the 2002 elections at which nine candidates competed. The two main rivals, Janez Drnovšek, as the foremost centre-left candidate, and Barbara Brezigar, as the centre-right candidate, were both nominated by political parties. Janez Drnovšek entered the election campaign as the then Prime Minister and President of the dominant party in the ruling centre-left coalition, the Liberal Democrats 3. There was no comparable survey data available for the presidential elections of 1992. 4. Data source: Državna volilna komisija 1997; RTV SLO 2012. of Slovenia (LDS) (once elected president, he resigned from both positions). In the first round, Drnovšek won 44.4 percent of votes and Brezigar won 30.8 percent, while each of the remaining seven candidates garnered less than ten percent of the votes. In the second round, Drnovšek won with 56.5% of the vote.5 Two rounds were also required at the 2007 national presidential elections. Seven candidates competed in the first round. The main centre-right parties did not nominate a candidate due to their failure to reach an agreement on a joint candidate. Therefore, Lojze Peterle, as the main representative of the right political pole, proceeded to announce his candidacy, endorsed by a non-party list of voters. He campaigned from his position as a member of the European Parliament, where he had been elected on the list of New Slovenia (NSi), which was a member of the European People's Party6. Due to disagreements, the centre-left parties nominated two main candidates, Danilo Türk and Mitja Gaspari. In the first round, Lojze Peterle came first with 28.7 percent, followed by the two centre-left candidates, Türk on 24.5 percent and Gaspari on 24.1 percent. The right-wing (nationalist) candidate, President of the Slovenian National Party, Zmago Jelinčič Plemeniti also won a large share of the vote (19.2 percent). In the second round, the centre-left candidate, Danilo Türk, defeated Peterle with a clear majority (68 percent against 32 percent).7 The 2012presidential elections appear to depart from the previous bipolar patterns. Only three candidates competed in the first round. Milan Zver stood as the centre-right candidate, nominated by two centre-right parties: the Slovenian Democratic Party and New Slovenia. On the other side, there were (seemingly) two left candidates: (a) Danilo Türk, the incumbent president, nominated by a group of voters and supported by the new centre-left party, Pozitivna Slovenija (PS), and (b) Borut Pahor, the former president of the Social Democrats (SD) and previously Prime Minister until the preterm elections of 2011. After losing his position as party leader by a narrow margin in 2012, Pahor declared his candidacy for the presidential elections; later 'his' party was 'forced' to support him as their candidate. During the campaign, Pahor adopted a stance between the left and right poles and thus competed with both the centre-left Türk and the centre-right Zver. With Pahor (39.9 percent) and Türk (35.9 percent) going through to the second round, the final campaign was fought as a 'centre versus centre-left' battle. By successfully appealing to voters on the centre-right, Pahor won the elections by a considerable margin (67.4 percent against Türk's 32.6 percent).8 Despite these shifting patterns of political competition, it would appear that Slovenian voters primarily follow their party identification and support whichever presidential candidate receives the endorsement of their favoured political party. In order to test the validity of this thesis, we conducted an analysis of the determinants of voter behaviour at presidential elections (voter's decision on the dichotomy 'left versus 5. Data source: Državna volilna komisija 2002; RTV SLO 2012 6. As the President of Slovenian Christian Democrats (later transformed into NSi), Peterle was the Prime Minister of the first Slovenian Government of 1992. 7. Data sources: Državna volilna komisija 2007; DELO 2007 8. Data source: Državna volilna komisija 2012. non-left candidate') with two main predictors (independent variables): (a) left-right self-placement and (b) political party affiliation (the voter's affinity to a political party). 3.2 Description of the Data The analyses are based on data from the Politbarometer public opinion survey9 on the public attitude towards the present conditions and developments in Slovenia, where respondents from a representative sample of adult telephone subscribers in Slovenia are interviewed via telephone according to a standardised questionnaire. We selected data from four Politbarometer surveys for the purposes of analysing voting choice at the Slovenian presidential elections. The four surveys were carried out in December 1997 (N=955) (Toš et al. 1997), December 2002 (N=941) (Toš et al. 2006), December 2007 (N=883) (Toš et al. 2009) and January 2013 (N=964) (Toš et al. 2013). Voting behaviour at presidential elections in 1997, 2002, 2007 and 2012 served as our dependent variable. We used political party affiliation ('affinity to a centre-left or centre-right political party') as our main independent variable (predictor)10, which was appended with another political orientation variable: the left-right self-positioning. In order to minimise the potential impact of a smaller sample size, political parties were grouped into three categories (left, right and other) and by grouping presidential candidates into two categories (left and other) (for a detailed description of these variables see Appendix 1). In order to test the hypothesis on the (relative) importance of party identification for determining voting behaviour at presidential elections, we also observed the influence of five background (demographic) variables: gender, age, education, place of residence, and religiosity. 3.3 Analysis and Results The analysis of the determinants of voting preferences at presidential elections in Slovenia was conducted in two steps. First, in order to investigate the differences in voting behaviour between groups by each of the seven independent variables, we used a bivariate analysis (see Table 1). The effect of independent variables was then tested using a logistics regression analysis. Three regression models were applied. First, all independent variables (two variables of political orientation and five demographic variables) were included in the model for an initial insight into the statistical significance and importance of all predictor variables - Model 1 (see Appendix 2 for results). Based on this information we began the process of formulating Model 2, in which we gradually remove statistically non-significant variables and variables that are very close to the p>0.05 limit. During this process we assessed the impact on the goodness-of-fit model, based on one hand on the Hosmer and Lemeshow test, and on the other hand on the predictive strength of the logistic regression model based on the Nagelkerke R-squared statistic. The purpose of Model 2 is to include statistically 9. For more, see http://www.cjm.si/?q=PolitBarometer. 10. We were only able to measure party identification on a nominal level - the Politbarometer survey does not include questions that would assess the intensity of party identification. significant variables that contribute most to the model's predictive strength while at the same time attempting to ensure a good model fit. By way of reference, we also include Model 3 which contains only two variables of political orientation (party affiliation and the left-right self-placement) and obtains good results in terms of the statistical significance of the Hosmer and Lemeshow test; however it lacks the predictive strength of Model 2 if we compare the Hosmer and Lemeshow test chi-square values to the degrees of freedom and take into account the Nagelkerke R-squared values. (See the synthetic results for the three models in Table 2.) At the bivariate level of analysis (presented in Table 1) we can already observe some regularity in the association between voter choice in presidential elections on the one side and their political orientations and demographic characteristics on the other. As we expected, those who feel more aligned with leftist parties and place themselves on the political left are substantially more inclined to vote for a leftist presidential candidate: Cramer's V for party affiliation in all years from 1997 to 2007 exceeds the value of 0.5 and is only lower in 2012 (0.344). We can see the strongest association between political orientation variables and voting behaviour at the 2007 elections. A similar pattern is evident for religiosity, which, in Slovenia, is closely related to the left-right political divisions. Other demographic variables exhibit weaker and very often statistically insignificant associations. The level of education is a special case: only for the 2012 elections we can prove a statistically significant association (Cramer's V = 0.194). The results for all four elections enable us to identify a certain trend in the period from 1997 to 2012: the association between political orientation variables and voting preferences at the presidential elections significantly drops in the case of 2012 elections, while the opposite appears to be the case with the demographic variables. Table 1: Factors of voting choice in Slovenian presidential elections in 1997, 2002, 2007 and 2012 (bivariate analysis results) 1997 2002 2007 2012 % X2 % X2 % X2 % X2 left Cramer V left Cramer V left Cramer V left Cramer V closest party **160.065 **217.380 **199.608 **82.243 - left 94.2 0.505 90.9 0.568 98.0 0.590 48.0 0.344 - right 44.3 26.9 43.1 8.7 - other, don't know 84.2 53.5 82.5 35.7 self-orientation **52.246 **69.888 **158.719 **136.712 - rather left 90.6 0.351 83.1 0.389 97.2 0.575 59.4 0.486 - centre 66.4 57.4 79.2 17.7 - rather right 56.0 41.2 40.9 5.1 age *10.077 *10.617 1.393 **18.733 - up to 30 years 67.4 0.126 61.7 0.125 75.0 0.049 37.5 0.164 - 31 to 45 years 76.1 57.8 77.1 19.2 - 46 to 60 years 76.1 68.4 81.3 24.0 - over 60 years 83.7 72.7 79.1 38.2 education 1.332 3.396 5.835 **26.157 - primary 73.7 0.046 62.1 0.071 72.1 0.101 14.8 0.194 - vocational secondary 73.9 59.3 79.7 24.0 - secondary 78.1 66.6 77.1 32.9 - college 76.2 69.1 84.3 42.9 religiosity **30.900 **47.711 **97.251 **90.266 - religious 66.0 0.222 52.7 0.267 61.2 0.414 14.7 0.361 - not religious 86.9 81.2 98.0 42.5 - somewhere 81.1 68.1 87.4 33.0 in between gender **11.966 2.414 *6.014 0.073 - male 69.4 0.137 61.4 0.060 73.6 0.102 31.7 0.010 - female 81.2 67.3 82.1 32.6 place of residence **19.518 **13.642 *10.089 **38.294 - rural 65.6 0.176 57.0 0.142 72.3 0.133 20.8 0.234 - smaller town 80.8 66.5 79.9 33.0 - larger city 82.7 78.2 79.2 38.5 - Ljubljana, Maribor 81.6 67.6 87.7 52.2 ** p < 0.01 * p < 0.05 Source: authors' analyses based on the Politbarometer survey data: December 1997, December 2002, December 2007 and January 2013. The main idea behind the application of the logistic regression model was to test the importance of party identification (political orientation) for voting behaviour at the presidential elections over time and in comparison to other possible factors that might affect voting behaviour11. When we entered all seven predictors (independent variables) into a logistic regression model (Model 1), the results clearly confirmed the trend that was already at the level of bivariate analysis (see Table 1 and Appendix 2). If we examine the importance of statistically-significant logistic regression predictor variables based on odds ratios, the party affiliation variable is the most important in 1997, 2002 and 2007, but not in 2012, where, for the first time in the second round, voters were deciding between a centrist and a leftist candidate, as opposed to the traditional left-right choice of candidates, as had been the case in all previous elections. In 2012, the left-right self-placement is the most important statistically significant predictor variable, while the importance of party affiliation is on the level (or even below the level) of the two demographic variables of education and religiosity. When we compare three logistic regression models on the basis of the model fit (the Hosmer and Lemeshow Test) and the predictive power of the models (R2 Nagelkerke) at all four time points (see Table 2), we found the following: a) considering the 1997 elections, we can see that the model with only two political orientation variables as predictors (Model 3) has a much better fit than both the models in which other predictors are also included - namely, Model 1 and Model 2; b) the results for the 2012 elections show the opposite picture - Model 3 has a much worse fit than Model 1 and Model 2; c) when we compare the results for model fit between all four elections, we can see the best fit (for all three models) in 2007; d) the predictive power of the models seems to be worst in 2012. 11. We observe the direct effect of each independent variable (predictor), which is controlled for effects of the other independent variables in the model. Table 2: Factors of voting choice in Slovenian presidential elections in 1997, 2002, 2007 and 2012. The results of binary logistic regression of election choice in presidential elections (1 - left; 0 - other) - only statistics for whole models are presented (method used: Enter) (see details in Appendix 2 and Appendix 3) 1997 2002 2007 2012 MODEL 3 (N = 421) closest party and self-orientation MODEL 3 (N = 460) closest party and self-orientation MODEL 3 (N = 479) closest party and self-orientation MODEL 3 (N = 574) closest party and self-orientation R2 Nagelkerke = 0.448 Omnibus tests of model coefficients: Chi-square = 152.964; df = 4; sig. = 0.000 Hosmer and Lemeshow test: Chi-square = 0.852, df = 7, sig. = 0.997 R2 Nagelkerke = 0.528 Omnibus tests of model coefficients: Chi-square = 221.003; df = 4; sig. = 0.000 Hosmer and Lemeshow test: Chi-square = 2.616, df = 5, sig. = 0.759 R2 Nagelkerke = 0.611 Omnibus tests of model coefficients: Chi-square = 241.081; df = 4; sig. = 0.000 Hosmer and Lemeshow test: Chi-square = 0.424, df = 4, sig. = 0.980 R2 Nagelkerke = 0.349 Omnibus tests of model coefficients: Chi-square = 166.343; df = 4; sig. = 0.000 Hosmer and Lemeshow test: Chi-square = 9.699, df = 5, sig. = 0.084 MODEL 2 (N = 419) closest party, self-orientation, gender, religiosity MODEL 2 (N = 665) closest party, religiosity, place of residence type MODEL 2 (N = 476) closest party, self-orientation, religiosity MODEL 2 (N = 569) closest party, self-orientation, religiosity, place of residence type, education, age R2 Nagelkerke = 0.489 Omnibus tests of model coefficients: Chi-square = 169.961; df = 7; sig. = 0.000 Hosmer and Lemeshow test: Chi-square = 17.145, df = 8, sig. = 0.029 R2 Nagelkerke = 0.427 Omnibus tests of model coefficients: Chi-square = 247.075; df = 7; sig. = 0.000 Hosmer and Lemeshow test: Chi-square = 2.436, df = 8, sig. = 0.965 R2 Nagelkerke = 0.664 Omnibus tests of model coefficients: Chi-square = 267.889; df = 6; sig. = 0.000 Hosmer and Lemeshow test: Chi-square = 2.461, df = 7, sig. = 0.930 R2 Nagelkerke = 0.425 Omnibus tests of model coefficients: Chi-square = 208.124; df = 15; sig. = 0.000 Hosmer and Lemeshow test: Chi-square = 9.285, df = 8, sig. = 0.319 MODEL 1 (N = 416) all independent variables MODEL 1 (N = 452) all independent variables MODEL 1 (N = 475) all independent variables MODEL 1 (N = 569) all independent variables R2 Nagelkerke = 0.506 Omnibus tests of model coefficients: Chi-square = 176.044; df = 16; sig. = 0.000 Hosmer and Lemeshow test: Chi-square = 17.494, df = 8, sig. = 0.025 R2 Nagelkerke = 0.560 Omnibus tests of model coefficients: Chi-square = 234.476; df = 16; sig. = 0.000 Hosmer and Lemeshow test: Chi-square = 4.761, df = 8, sig. = 0.783 R2 Nagelkerke = 0.674 Omnibus tests of model coefficients: Chi-square = 271.942; df = 16; sig. = 0.000 Hosmer and Lemeshow test: Chi-square = 0.923, df = 8, sig. = 0.999 R2 Nagelkerke = 0.426 Omnibus tests of model coefficients: Chi-square = 209.104; df = 16; sig. = 0.000 Hosmer and Lemeshow test: Chi-square = 6.511, df = 8, sig. = 0.590 Source: authors' analyses based on the Politbarometer survey data: December 1997, December 2002, December 2007 and January 201 In addition to the more general trend described above, we can identify religiosity as a particularly important factor in the 2007 election choice. The odds ratio value for the not religious category in 2007 is many times greater in terms of importance than for any other year (see Appendix 2). In addition, 2007 is the only year in which the second most important variable in terms of odds ratios is religiosity - for other observed years, party affiliation and/or self-placement are always the two most important. Even in the 2012 elections, when, as mentioned, the demographic variables gain considerable importance, religiosity is not as essential as it was in the 2007 elections; in 2012, the odds ratios identify education as the second most important while religiosity follows in third place. 3.4 Discussion Based on these findings, we could argue that the importance of party identification (and political orientation in general) for determining voter choice at presidential elections is unstable and in fact rather volatile (Table 3). However, this does not mean that party identification is significantly diminishing in its importance but rather that it no longer ranks head and shoulders above other possible factors (such as left-right self-placement, religiosity, and education). Table 3: The main competitors and factors of voting choice in the Slovenian presidential elections in 1997, 2002, 2007 and 2012 (summary of binary logistic regression analysis) 1997 2002 2007 2012 Kučan vs. 'others' (L vs. R) Drnovšek vs. Brezigar (L vs. R) Türk vs. Peterle (L vs. R) Türk vs. Pahor (L vs C.) Variables: - closest party (left), - self-orientation (rather left), - gender (binary, male = 0), - religiosity (religious) Variables: - closest party (left), - religiosity (religious), - place of residence (rural) Variables: - closest party (left), - self-orientation (rather left), - religiosity (religious) Variables: - closest party (left), - self-orientation (rather left), - religiosity (religious), - place of residence (rural), - education (primary), - age (up to 30 years) Significant categories: Significant categories: Significant categories: Significant categories: - closest party: right - self-orientation: rather right - closest party: other - gender: female - self-orientation: centre - religiosity: not religious - closest party: right - closest party: other - place of residence: larger city - religiosity: not religious - closest party: right - religiosity: not religious - self-orientation: rather right - closest party: other - self-orientation: centre - religiosity: in between - self-orientation: rather right - self-orientation: centre - education: college - education: secondary - religiosity: not religious - closest party: right - place of residence: Ljubljana, Maribor - closest party: other Source: authors' analyses based on the Politbarometer survey data: December 1997, December 2002, December 2007 and January 2013. Note: The reference category for binary logistic regression is indicated in parentheses next to the name of the variable.This list includes only statistically significant variable categories ordered by importance (based on odds ratios) (see details in Appendix 3). Furthermore, factors other than 'closest party' and 'left-right self-orientation' interfere with the voting decision. While voter religiosity has featured among the factors at all elections investigated, gender, type and place of residence, education and age have emerged only occasionally. When the competition among candidates moves from a 'left versus right' to a 'left versus centre' pattern (as was the case in the 2012 elections), almost all non-political factors become statistically significant in co-determining voter behaviour. Religiosity was most significant when Lojze Peterle from the New Slovenia - Christian-People's Party competed at the 2007 elections as the centre-right candidate, albeit without the support of the centre-right party bloc. Taking into account the main rivals at each election, we can also consider another pattern: when rivals are substantially uneven regarding their political reputation, voters' party affiliation (or political orientation in general) pales in comparison to other factors; but when there is no substantial difference in the political reputations of the rival candidates, other factors become more important. The 1997 elections are an example of competition between uneven candidates - the winner Milan Kučan was at the time the most reputable politician by far12. The 2012 elections are an example of competition between candidates with a similar reputational standing: almost all public opinion surveys prior to the election ranked both the centre-left candidates (Türk and Pahor) in the first two places, and centre-right candidate (Zver) marginally behind13. Comparing both elections in terms of factors of voting choice, we can see that voters' political affiliation is much more important at elections where 'uneven' candidates are fielded (e.g. 1997) than at elections in which candidates with similar levels of reputation compete (e.g. 2012) (Table 3, Appendix 2). 4. Conclusions In this article we examined the variables determining voting behaviour at presidential elections in circumstances of changing patterns of party competition in the parliamentary arena. Looking longitudinally, we expected tripolar competition in the party system to be amended by bipolar political competition at the presidential elections and vice-versa. This expectation was based on the presumption that the communist-anti-communist division still mattered in the post-socialist context and that voters would strategically allow this division to be awarded only in one political arena (parliamentary or presidential) and counter-balanced in the other. Our longitudinal analysis in the case of Slovenia shows that party identification and political orientation in general continue to matter, although their importance has decreased. Meanwhile other social determinants and thus more social relevant identities have gained in importance in which candidates have not enjoyed the clear support of a particular party block anymore, particularly when candidates have initially been self-nominated, as was the case for Lojze Peterle and Borut Pahor who only garnered political backing with a delay. Indeed, at the 2012 presidential elections, there were many relevant non-political factors, among them especially religiosity, place / type of residence, education and age. Additionally, we were unable to trace any direct impact of concurrent elections and the political ideological identification of voting behaviour in the electoral cases in Slovenia analysed so far. Besides the reasons already stated, this result could be explained by the specifics of institutional rules and political circumstances which enable non-party nominations for the presidential elections (as was the case in the concurrent elections of 1992), and which enabled the calling of the 2011 preterm parliamentary elections. The preterm elections limited the potential 12. For example, the results of the Slovenian Public Opinion Survey prior to the election reveals that nearly 54 percent of respondents ranked Milan Kučan as the most reputable politician, followed by Janez Drnovsek with 15 percent, while Kučan's main rival ranked fourth with only a five percent rating (Toš et al. 1997). 13. The public opinion poll in the main national broadsheet, DELO, shows the following ranking (on a scale of 1 to 5): Türk in the first place with scores of 3.44, followed by Pahor in the second place (3.23), and Zver in the seventh place (2.93) (DELO 2012). advantages of concurrent election in 2012, when, after two decades of democracy, both the presidential and parliamentary elections had been planned to be held in the same year for the first time since 1992. Nevertheless, a sharp decrease in trust in all political parties, as well as voters' extreme dissatisfaction with the ideological struggles between the governmental and oppositional parties at a time when all parties were expected to manage the financial and economic crisis efficiently, may have added to the intervening 'crisis' variable. This leads us to the broader contextual determinants of voting behaviour. As for the broader post-socialist context, it is clear that, in investigating the relationships between voter behaviour in different elections (as in other political analysis), we still need to take into account the idiosyncratic ideological-political divisions, along with the persistent problem of citizens being alienated from political parties. The latter predictably feeds the phenomena of individual 'political personalities' seeking their own political promotion, as well as the growing relevance of wider social identities. Finally, some of the preliminary findings that relate to the issues of strategic voting and political supply call for further in-depth research. Appendix 1: Description of the Variables Table 4: Description of variables voting choice at presidential election and party closeness Description Starting categories Final categories (included in analysis) voting choice in the first round* of the 1997 presidential elections Milan Kučan Bogomir Kovač Tone Peršak left Marjan Poljšak Janez Podobnik Franc Miklavčič Marjan Cerar Jože Bernik other voting choice in the second round of the 2002 presidential elections Janez Drnovšek Barbara Brezigar left other voting choice in the second round of the 2007 presidential elections Danilo Türk Lojze Peterle left other voting choice in the second round of the 2012 presidential elections Danilo Türk Borut Pahor left other closest party Democratic party of Pensioners of Slovenia (DeSUS) Liberal democracy of Slovenia (LDS) Social Democrats (SD) ZARES - New Politics (ZARES) Sustainable Development of Slovenia (TRS) Positive Slovenia (PS) Democrats of Slovenia (DS) left New Slovenia - People's Christian Party (NSi) Slovenian Democratic Party (SDS) Slovenian people's Party (SLS) Slovenian National Party (SNS) Citizens' List of Gregor Virant (DL) Slovenian Christian Democrats (SKD) right other parties don't know other, don't know *Only data for the first round wa:i used as the 1997 Slovenian presidential elections did not have a second round. The candidates are grouped into two categories (left and other) as is the case for all other observed presidential elections. Appendix 2: Factors of Voting Choice in Slovenian Presidential Elections in 1997, 2002, 2007 and 2012 (binary logistic regression - Model 1) Table 5: Results of the binary logistic regression of voting choice in presidential elections (1 - left; 0 - other) (Model 1) (method: Enter). 1997 (N = 416) 2002 (N = 452) 2007 (N = 475) 2012 (N = 569) B p Exp B p Exp B p Exp B p Exp (B) (B) (B) (B) closest party 0.000 0.000 0.000 0.001 - left (ref. categ.) 0.000 0.000 0.000 0.000 - right -2.850 0.000 0.058 -3.655 0.000 0.026 -3.114 0.000 0.044 -0.805 0.023 0.447 - other, don't know -1.496 0.001 0.224 -2.323 0.000 0.098 -1.592 0.012 0.204 0.550 0.035 1.734 self-placement 0.000 0.166 0.000 0.000 - rather left (ref. categ.) 0.000 0.000 0.000 0.000 - centre -1.141 0.005 0.319 -0.284 0.450 0.753 -1.444 0.007 0.236 -1.484 0.000 0.227 - rather right -1.950 0.000 0.142 -0.663 0.060 0.515 -2.620 0.000 0.073 -2.471 0.000 0.084 Age 0.410 0.292 0.882 0.017 - up to 30 years (ref. categ.) 0.000 0.000 0.000 0.000 - 31 to 45 years 0.686 0.100 1.985 0.105 0.808 1.111 -0.399 0.512 0.671 -0.564 0.418 0.569 - 46 to 60 years 0.361 0.383 1.435 0.660 0.128 1.935 -0.359 0.546 0.698 -0.259 0.680 0.772 - more than 60 years 0.506 0.261 1.659 0.563 0.226 1.756 -0.144 0.807 0.866 0.408 0.497 1.504 religiosity 0.143 0.090 0.000 0.007 - religious (ref. categ.) 0.000 0.000 0.000 0.000 - not religious 0.773 0.050 2.166 0.712 0.043 2.038 3.151 0.000 23.354 0.892 0.002 2.440 - somewhere in between 0.152 0.685 1.164 -0.086 0.826 0.918 1.160 0.026 3.188 0.601 0.063 1.824 education 0.245 0.501 0.699 0.069 - primary (ref. categ.) 0.000 0.000 0.000 0.000 - vocational secondary -0.442 0.340 0.643 0.028 0.956 1.028 0.726 0.244 2.067 0.663 0.208 1.941 - secondary 0.012 0.977 1.012 0.266 0.521 1.304 0.452 0.376 1.572 0.963 0.033 2.619 - college -0.817 0.116 0.442 -0.314 0.533 0.730 0.480 0.432 1.616 1.217 0.011 3.377 place / type of residence 0.679 0.166 0.576 0.139 - rural (ref. categ.) 0.000 0.000 0.000 0.000 - smaller town 0.306 0.421 1.358 0.067 0.844 1.069 0.178 0.651 1.195 0.139 0.620 1.150 - larger city 0.312 0.516 1.366 1.023 0.030 2.781 -0.596 0.309 0.551 0.231 0.512 1.260 - Ljubljana, Maribor 0.523 0.240 1.688 0.313 0.444 1.368 0.247 0.689 1.281 0.710 0.025 2.033 Gender (dichotomised) (1 = female) 1.058 0.001 2.879 0.007 0.980 1.007 0.589 0.115 1.803 -0.233 0.323 0.792 Constant 2.531 0.000 12.569 1.979 0.002 7.236 3.677 0.000 39.509 -1.588 0.041 0.204 R2 Nagelkerke = R2 Nagelkerke = R2 Nagelkerke = R2 Nagelkerke = 0.506 0.560 0.674 0.426 Omnibus tests Omnibus tests Omnibus tests Omnibus tests of model: of model : of model: of model: c2=176.044, df =16, c2=234.476; df=16; c2=271.942; df =16; c2=209.104; df =16; sig.=0.000 sig.=0.000 sig.=0.000 sig.=0.000 Hosmer and Hosmer and Hosmer and Hosmer and Lemeshow test: Lemeshow test: Lemeshow test: Lemeshow test: c2=17.494, df =8, c2=4.761, df =8, c2=0.923, df =8 c2=6.511, df =8 sig.=0.025 sig.=0.783 sig.=0.999 sig.=0.590 Source: authors' analyses based on the Politbarometer survey December 2007 and January 2013. data: December 1997, December 2002, Bibliography Bartels, Larry M. (2000): Partisanship and Voting Behaviour, 1952-1996. American Journal of Political Science, 44 (1): 35-50. Bartolini, Stefano, and Mair, Peter (1990): Identity, Competition, and Electoral Availability. New York: Cambridge University Press. Berelson, Bernard R., Lazarsfeld, Paul F., and McPhee, William N. (1954): Voting: A Study of Opinion Formation in a Presidential Campaign. Chicago: University of Chicago Press. Birch, Sarah (2003): Electoral Systems and Political Transformation in Post-communist Europe. New York: Palgrave MacMillan. Brennan, Geoffrey, and Hamlin, Alan (1998): Expressive voting and electoral equilibrium. Public Choice, 95: 149-175. Van der Brug, Wouter, Franklin, Mark, and Tokad, Gabor (2008): One electorate or many? Differences in party preference formation between new and established European democracies. Electoral Studies, 27 (4): 589-600. Budge, Ian, Klingemann, Hans-Dieter, Volkens, Andrea, Bara, Judith, and Tanenbaum, Eric (2001): Mapping Policy Preferences: Estimates for Parties, Electors, and Governments 1945-1998. New York: Oxford University Press. Campbell, Angus, Converse, Philip E., Miller, Warren E., and Stokes, Donald E. (1960): The American Voter. Chicago: University Of Chicago Press. Cheibub, Jose Antonio, Przeworski, Adam, Neto, Fernando Papaterra Limongi, and Alvarez, Michael M. (1996): What Makes Democracies Endure? Journal of Democracy, 7 (1): 39-55. Dalton, Russell J. (1984). Cognitive Mobilization and Partisan Dealignment in Advanced Industrial Democracies. The Journal of Politics, 46 (1): 264-284. Dalton, Russell J., and Klingemann, Hans-Dieter (2007): The Oxford Handbook of Political Behavior. New York: Oxford University Press. Dalton, Russell J., Scarrow, Susan E, and Cain, Bruce E. (2003): New Forms of Democracy? Reform and Transformation of Democratic Institutions. In Cain, Bruce E., Dalton, Russel J., and Scarrow, Susan E. (eds.): Democracy Transformed? Expanding Political Opportunities in Advanced Industrial Democracies: 1-22. New York: Oxford University Press. Dalton, Russell J., and Wattenberg, Martin P. (2000): Parties without Partisans: Political Change in Advanced Industrial Democracies. New York: Oxford University Press. Dowding, Keith (1994): The Compatibility of Behaviouralism, Rational Choice and 'New In-stitutionalism'. Journal of Theoretical Politics, 6 (1): 105-117. Downs, Anthony (1957): An Economic Theory of Democracy. New York: Harper and Row. Dubrovnik, Tadej (2013): Položaj predsednika republike na izvršilnem in zakonodajnem področju v novih demokracijah EU. Master thesis. Ljubljana: University of Ljubljana, Faculty of Law. Enyedi, Zsolt (2008): The Social and Attitudinal Basis of Political Parties: Cleavage Politics Revisited. European Review, 16 (3): 287-304. Evans, Geoffrey, and Whitefield, Stephen (1998): The Structuring of Political Cleavages in Post-Communist Societies: The Case of the Czech Republic and Slovakia. Political Studies, 46 (1): 115-139. Fink-Hafner, Danica (2010): Slovenia since 1989. In Ramet, Sabrina P. (ed.): Central and Southeast European Politics since 1989: 235-257. Cambridge, New York: Cambridge University Press. Fuchs, Dieter, and Klingemann, Hans-Dieter (2002): Eastward Enlargement of the European Union and the Identity of Europe. West European Politics, 25 (2): 19-54. Gunther, Richard, Montero, Jose Ramon, and Puhle, Hans-Jürgen (eds.) (2006): Democracy, Intermediation, and Voting on Four Continents. Oxford: Oxford University Press. Hafner-Fink, Mitja, Fink-Hafner, Danica, and Krašovec, Alenka (2011): Changing Patterns of Political Participation. In Lewis, Paul, and Markowski, Radoslaw. (eds.): Europeanising Party Politics: Comparative Perspectives on Central and Eastern Europe: 69-96. Manchester, New York, Vancouver: Manchester University Press. Hellwig, Timothy, and Samuels, David (2007): Electoral Accountability and the Variety of Democratic Regimes. British Journal of Political Science, 38 (1): 65-90. Huber, John D., and Inglehart Ronald (1995): Expert interpretations of party space and party locations in 42 societies. Party Politics, 1: 73-111. Inglehart, Ronald, and Catterberg, Gabriela (2003): Trends in Political Action: The Developmental Trend and the Post-Honeymoon Decline. International Journal of Comparative Sociology, 43 (3-5): 300-316. Inglehart, Ronald, and Flanagan, Scott C. (1987): Value Change in Industrial Societies. American Political Science Review, 81 (4): 1289-1319. Jost, John T., Federico, Christopher M., and Napier, Jaime L. (2009): Political Ideology: Its Structure, Functions, and Elective Affinities. Annual Review of Psychology, 60: 307-337. Karp, Jeffrey A., and Banducci, Susan A. (2007): Party Mobilization and Political Participation in New and Old Democracies. Party Politics, 13 (2): 217-234. Kitschelt, Herbert, and Hellemans, Staf (1990): The Left-Right Semantics and the New Politics Cleavage. Comparative Political Studies, 23 (2): 210-238. Kitschelt, Herbert, Mansfeldova, Zdenka, Markowski, Radoslaw, and Toka, Gabor (1999): Post-Communist Party Systems: Competition, Representation, and Inter-Party Cooperation. Cambridge: Cambridge University Press. Klingemann, Hans-Dieter, Volkens, Andrea, Bara, Judith, Budge, Ian, and McDonald, Michael D. (eds.) (2006): Mapping Policy Preferences II: Estimates for Parties, Electors, and Governments in Eastern Europe, the European Union and the OECD, 1990-2003. Oxford: Oxford University Press. Kubicek, Paul (2000): Post-communist political studies: ten years later, twenty years behind? Communist and Post-Communist Studies, 33 (3): 295-309. Lazarsfeld, Paul F., Berelson, Bernard, and Gaudet, Hazel (1944): The People's Choice: How the Voter Makes Up His Mind in a Presidential Campaign. New York: Columbia University Press. LeDuc, Lawrence, Niemi, Richard G., and Norris, Pippa (eds.) (2002): Comparing Democracies 2: New Challenges in the Study of Elections and Voting. London: Sage Publications. Lewis, Paul, and Markowski, Radoslaw. (2011) (eds.): Europeanising Party Politics: Comparative Perspectives on Central and Eastern Europe. Manchester, New York, Vancouver: Manchester University Press. Lijphart, Arend (1992): Parliamentary versus presidential government. New York: Oxford University Press. Linz, Juan J. (1994): Presidential or Parliamentary Democracy: Does it make a difference? In Juan J. Linz, and Valenzuela, Arturo (Eds.): The failure of presidential democracy, Volume 1: 3-87. Baltimore: Johns Hopkins University Press. Lipset, Seymour Martin, and Rokkan, Stein (1967): Cleavage Structures, Party System and Voter Alignments: An Introduction. In Lipset, Seymour Martin, and Rokkan, Stein (eds.): Party Systems and Voter Alignments: Cross-National Perspectives: 1-64. New York: Free Press. Lipset, Seymour Martin, and Stein Rokkan. 1990. "Cleavage Structures, Party Systems, and Voter Alignments." Pp. 91-139 in The West European Party System. http://www.amazon. com/European-System-Readings-Politics-Government/dp/0198275838 Lupia, Arthur and McCubbins, Mathew D. (1998): The Democratic Dilemma: Can Citizens Learn What They Need to Know? New York: Cambridge University Press. Marks, Gary, Hooghe, Liesbet, Nelson, Moira, and Edwards, Erica. (2006): Party Competition and European Integration in the East and West, Comparative Political Studies, 39 (2): 155-175. Mishler, William, and Rose, Richard (1994): Support for Parliaments and Regimes in the Transition toward Democracy in Eastern Europe. Legislative Studies Quarterly, 19 (1): 5-32. Niemi, Richard G., and Weisberg, Herbert F. (eds.) (2001): Controversies in Voting Behavior, 4th ed. Washington, DC: Congressional Quarterly Press. Norris, Pippa, Curtice, John, Sanders, David, Scammell, Margaret, and Semetko, Holli A. (1999): On Message: Communicating the Campaign. London: Sage. Popkin, Samuel L. (1994): The Reasoning Voter: Communication and Persuasion in Presidential Campaigns, 2nd edition. Chicago: University of Chicago Press. Rovny, Jan, and Edwards, Erica E. (2012): Struggle over Dimensionality: Party Competition in Western and Eastern Europe. East European Politics & Societies 26 (1): 56-74. Rupnik, Jacques, and Zielonka, Jan (2012): Democracy in Central and Eastern Europe: The State of the Art. East European Politics & Societies, 27 (1): 3-25. Tavits, Margit (2009): Direct Presidential Elections and Turnout in Parliamentary Contests. Political Research Quarterly, 62 (1): 42-54. Tavits, Margit, and Letki, Natalia (2009): When Left Is Right: Party Ideology and Policy in Post-Communist Europe. American Political Science Review, 103 (4): 555-569. Tsebelis, George (1990): Nested Games: Rational Choice in Comparative Politics. Berkeley: University of California Press. Tucker, Joshua (2002): The First Decade of Post-Communist Elections and Voting: What Have We Studied, and How Have We Studied It? Annual Review of Political Science, 5: 271-304. Wessels, Bernhard, and Klingemann, Hans-Dieter (2006): Partisans and Voters - Representative Consolidation in Central and Eastern Europe? International Journal of Sociology, 36 (2): 11-44. Whitefield, Stephen, and Rohrschneider, Robert (2009): Representational Consistency: Stability and Change in Political Cleavages in Central and Eastern Europe. Politics Policy, 37 (4): 667-90. Sources DELO (2007): Med uradnimi predsedniškimi kandidati tudi Monika Piberl. Available on: http:// www.delo.si/clanek/o243083 (23. 2. 2014) DELO (2012): Anketa Dela: Zmaga stranke neopredeljenih. Available on: http://www.delo.si/ novice/politika/anketa-dela-zmaga-stranke-neopredeljenih.html (2.3.2014) Državna volilna komisija (1997): Volitve predsednika RS - leto 1997. Available on: http://www. dvk-rs.si/index.php/si/arhiv-predsednika-rs/leto-1997 (21. 2. 2014) Državna volilna komisija (2002): Volitve predsednika RS - leto 2002. Available on: http://www. dvk-rs.si/index.php/si/arhiv-predsednika-rs/leto-2002 (21. 2. 2014) Državna volilna komisija (2007): Volitve predsednika RS - leto 2007. Available on: http://www. dvk-rs.si/index.php/si/arhiv-predsednika-rs/leto-2007 (21. 2. 2014) Državna volilna komisija (2012): Volitve predsednika RS - leto 2012. Available on: http://www. dvk-rs.si/index.php/si/arhiv-predsednika-rs/volitve-predsednika-rs-leto-2012 (23. 2. 2014) RTV SLO (2012): Predsedniške volitve 1990 - 2007. Available on: http://www.rtvslo.si/strani/ zgodovina/4348 (4. 3. 2014) Toš, Niko, Malnar, Brina, Kurdija, Slavko, and Švara, Sergio (1997): Politbarometer PB12/97, Slovenia: December 1997. Data file. Ljubljana: University of Ljubljana, Public Opinion and Mass Communication Research Centre. Toš, Niko, Hafner-Fink, Mitja, and Malnar, Brina (1997): Slovensko javno mnenje 1997/1: Stališča Slovencev o pridruževanju Evropski Uniji [data file]. Ljubljana: Fakulteta za družbene vede, Center za raziskovanje javnega mnenja in množičnih komunikacij [producer]. Ljubljana: Univerza v Ljubljani, Arhiv družboslovnih podtakov [distributor], 2000. Toš, Niko, Kurdija, Slavko, Falle, Rebeka, Filej, Živa and Kecman, Ivana (2006): Politbarometer PB12/02, Slovenia: December 2002. Data file. Ljubljana: University of Ljubljana, Public Opinion and Mass Communication Research Centre. Toš, Niko, Kurdija, Slavko, Kovačič, Matej, Vovk, Tina, Bešter Falle, Rebeka, Broder, Živa, and Kecman, Ivi (2009): Politbarometer PB12/07, Slovenia: December 2007. Data file. Ljubljana: University of Ljubljana, Public Opinion and Mass Communication Research Centre. Toš, Niko, Kurdija, Slavko, Broder, Živa, and Vovk, Tina (2013): Politbarometer PB01/13, Slovenia: January 2013. Data file. Ljubljana: University of Ljubljana, Public Opinion and Mass Communication Research Centre. Authors' data: Dr. Mitja Hafner-Fink, Assistant Professor FDV, Kardeljeva ploščad 5, 1000 Ljubljana E-mail: mitja.hafner-fink@fdv. uni-lj. si Telephone: 01/ 5805-283 Dr. Simona Kustec Lipicer, Associate Professor FDV, Kardeljeva ploščad 5, 1000 Ljubljana E-mail: simona. kustec-lipicer@fdv. uni-lj. si Telephone: 01/ 5805-178 Dr. Danica Fink-Hafner, Professor FDV, Kardeljeva ploščad 5, 1000 Ljubljana E-mail: danica.fink-hafner@fdv.uni-lj.si Telephone: 01/ 5805-173 Gregor Čehovin, Senior administrative officer FDV, Kardeljeva ploščad 5, 1000 Ljubljana E-mail: gregor. cehovin@fdv. uni-lj. si Telephone: 01/ 5805-352