Explaining Electoral Support for the Radical Right


1 Introduction: Voting for the Radical Right

Within the larger field of Radical Right studies, the question of why people vote for Radical Right Parties (RRPs) has attracted a large (perhaps disproportionally so) chunk of scholarly attention. There are at least three reasons for this. First, the early (and rather humble) electoral successes of the Radical Right in Western Europe during the early 1980s stirred memories of the 1920 and 1930s, when parties such as the Italian Fascists or the German Nazis rose from obscurity to overturn democracy (Prowe, 1994). Given these traumatic experiences, scholars were understandably eager to analyse the motives behind such potentially fatal electoral choices.

Second, when it became increasingly clear that the most electorally successful of these RRPs were not just clones of the old fascist right of the inter war years but rather belonged to a new party family (Mudde, 1996), researchers wanted to understand the social forces that brought about the rise of this largely unexpected phenomenon. After all, even non-extremist RRPs are still widely seen as problematic, because they promote a political ideal that has been dubbed “illiberal democracy” (Mudde, 2007), and often disrupt the political process.

Third, support for the Radical Right displays an unusual degree of variation across time and space. In Southern Europe, Cyprus (until 2016), Malta, Portugal and Spain never had a relevant RRP, whereas RRPs have been more or less consistently successful in Austria, Denmark, France, Italy, and Norway. Electoral support for the Radical Right has been volatile in Germany, Greece, Sweden and the UK. In the Netherlands, which featured extremist but tiny right-wing parties in the 1980s and 1990s, modern RRPs only emerged in the early 2000s. As of 2016, the radical right PVV is the country’s largest political party in terms of voting intentions. Belgium provides perhaps the most striking example of variability: While the Walloon National Front always remained at the margins in Wallonia, the Vlaams Blok/Vlaams Belang went from strength to strength in the Flemish part of the country during the 1990s and early 2000s, but lost roughly three quarters of their support between 2004 and 2014. To summarise, there is ample reason for treating support for the Radical Right as an unusual and potentially even dangerous phenomenon.

The most obvious way to study Radical Right voting would be to apply the standard tools of electoral research. Modern election studies usually rely on an eclectic blend of variables and alleged mechanisms, but at the core, there is usually the assumption that voters respond to short-term factors (candidates and political issues) on the one hand, and long-to-medium forces (party loyalties, value orientations, ideological convictions and group memberships) on the other. Almost sixty years ago, Angus Campbell and his associates (Campbell, 1960) have proposed a conceptual framework that encompasses these and other variables: In their “funnel of causality” metaphor, the proximate determinants of a given electoral choice are causally linked to more distant antecedents, forming a “funnel” that gets wider as more and more stable attitudes and earlier events are being considered. Decades of criticism not withstanding, this framework still explicitly or implicitly undergirds most empirical research into voting behaviour.

In the subfield of Radical Right voting, however, researchers habitually seem to ignore most of what constitutes the “normal science” (Kuhn, 1962) of electoral research, either because they are unaware of it, or because they are chiefly interested in “deeper” explanations that are located towards the far side of the funnel. Nonetheless, the funnel metaphor still provides a useful template for organising and comparing competing and complementary explanations for Radical Right electoral support.

However, the distinction between “supply side” and “demand side” factors, which can be traced back to an early article by Klaus von Beyme (Beyme, 1988), proved to be a much more popular schema for structuring potential explanations. Unfortunately, it is not entirely clear what is meant by “supply” and “demand” in this context and whether these two exhaust the full set of relevant factors, although the dichotomy has a certain heuristic value: The notion of a “supply side” usually refers to all variables pertaining to the RRP itself. This includes, but is not limited to, the stylistic and substantive content of the party manifesto and other texts, speeches or statements produced by the party, the party’s organisational structure and resources, and the presence or absence of a “charismatic leader”. The “demand side”, on the other hand, encompasses traits, experiences and attitudes that may predispose voters to support an RRP.

A number of other relevant factors, however, do not sit easily within the confines of this dichotomy. The ideological positions of mainstream right parties, for instance, could be considered part of the “supply” in a wider sense, but the same is not true for institutional variables such as the electoral system or the degree of decentralisation. These features of the wider political system may explain why would-be political entrepreneurs decide to enter the political arena to provide a RRP supply, or why a given demand for RRP policies may help or hurt the mainstream right parties. Put differently, many institutional factors should be seen as mediators of supply and demand rather than as members of either category. Other system-level variables – most prominently unemployment and immigration – are best understood as distal causes of demand, or as an incentives to provide supply.

Therefore, it seems more fruitful to distinguish between variables on the micro, meso, and macro level, and the remainder of this chapter will proceed accordingly. Most approaches, however, more or less explicitly follow the logic of a multi-level explanation (Coleman, 1994), requiring occasional cross-references between the sections.

The literature on this topic is already vast and keeps on growing quickly. My self-consciously eclectic bibliography on the Radical Right in Europe (http://www.kai-_arzheimer.com/extreme-_right-_western-_europe-_bibliography), which is nowhere near complete, currently stands at more than 600 titles. The literature review in this chapter is therefore by necessity highly selective and idiosyncratic: I will focus on (Western) Europe, and on a small number of contributions that I consider landmarks. Although comparative multi-level analyses are now something like the gold standard in the field, I will also consider single-country case studies where they present results that (probably) generalise beyond the polity in question, or designs that are of a more general interest. Moreover, while there is always the danger of aggregation bias lurking in the background, I will frequently discuss findings from field-defining aggregate studies, without re-iterating the usual warnings about the ecological fallacy (Robinson, 1950) time and again. Consider yourself trigger-warned.

2 Micro-level Factors

2.1 Party Identification

Party identification is arguably the most important factor when it comes to explaining voting decisions, but it is conspicuously underrepresented in the literature on the Radical Right. One possible explanation for this is the fact that party identification is supposed to be acquired through years, if not decades of political socialisation. As many RRPs only rose to prominence in the 1980s and 1990s, identification with them could hardly be a major factor behind their ascendancy. A a consequence, most early studies completely ignored party identification, and one of those few assessing its effect (based on data from the mid-1990s) concluded that “the identification motive is clearly significantly under-represented among VB [Vlaams Blok] voters” (Swyngedouw, 2001, p. 228).

A more modern approach highlights the negative effect of identifications with other parties. Building on the notion (derived from the older literature, e.g. Kitschelt (1995) and Ignazi (2003)) that the rise of the Radical Right only became possible once there was a sufficiently large pool of voters that were no longer attached to any of the established parties, Arzheimer and Carter (2009a) focus on (the lack of) identifications with mainstream right-wing parties. Using data from the 2002/03 wave of the European Social Survey, they demonstrate that voters who are still attached to a Christian Democratic or Conservative party almost never vote for a Radical Right party. Put differently, they see the absence of other identifications as a necessary (if insufficient) pre-condition for Radical Right-wing voting. However, some of the most successful RRPs (e.g. the French National Front, the Austrian Freedom Party, the Danish People’s Party or the Norwegian Progress Party) have been electorally relevant for two decades or more now, so the impact of identifying with the RRP should be modeled, too, but very few authors (e.g. Arzheimer, 2009b) account for this potential positive effect of party identification.

2.2 Candidates: The (ir)relevance of charismatic leaders

While party identifications have been more or less neglected as a key explanatory variable for RRP support, candidates and more specifically “charismatic” party leaders have attracted a great deal of attention (e.g. Taggart, 1995). There are two reasons for this: First, many observers mistook the rise of the RRPs in the 1980s for a “Return of the Führers” of the 1920s (Prowe, 1994). Second, many RRPs appeared to be personal parties, especially during the break-through phase (Eatwell, 2005, p. 106). Third, agency is always more attractive than structure.

However, what is meant by “charisma” is not usually clear. There are serious doubts that Weberian “charisma” – a personal bond between the (party) leader and his followers – was in any way relevant for the rise of the Radical Right (Eatwell, 2005), and even those two parties most commonly associated with their “charismatic” leaders – Joerg Haider’s Austrian Freedom Party and Jean-Marie Le Pen’s French National Front – underwent a process of “institutionalisation” (Pedahzur and Brichta, 2002). Even more importantly for the question of electoral behaviour, Brug and Mughan (2007) demonstrate that RRPs benefit from candidate effects in exactly the same way as established parties: While having an appealing candidate is certainly linked to greater electoral support, the magnitude of this effect is not larger than it is for other parties.

2.3 Issues, Ideology and Value Orientations

2.3.1 Pure Protest Voting, Anti-Immigrant Sentiment, and Unemployment (threat)

When it comes to explaining Radical Right support, the notion of a “pure protest vote” is still prominent. In its most extreme guise, the pure protest thesis claims that Radical Right support is driven by feelings of alienation from the political elites and the political system that are completely unrelated to policies or values and hence have nothing to do with the Radical Right’s political agenda (Eatwell, 2000). A more realistic variety of the protest thesis suggests that voters do indeed care about policies but hold less extreme preferences than the Radical Right manifestos would suggest. In this scenario, voters instrumentally support the Radical Right in the hope that mainstream right parties will reconsider their position and move somewhat closer to the Radical Right without copying all of their policies. Once the mainstream right has made this adjustment, Radical Right support would collapse. This logic is akin to directional voting (Merrill and Grofman, 1999) but puts more emphasis on emotions.

Empirically, pure protest voting remains elusive. Starting with Billiet and Witte’s (1995) study of Vlaams Blok support in the 1991 General Election in Belgium, a host of single-country and comparative studies have demonstrated time and again that anti-immigrant sentiment is the single most important driver of the Radical Right vote (Mayer and Perrineau, 1992; Brug, Fennema, and Tillie, 2000; Brug and Fennema, 2003; Norris, 2005; Mughan and Paxton, 2006; Arzheimer, 2009b; Ford, Goodwin, and Cutts, 2011). That does not mean that the prototypical voter of the Radical Right is not alienated from the political elites and susceptible to the populist rhetoric of many RRPs. But the vast majority of their voters support the Radical Right because of their anti-immigrant claims and demands, and their sense of frustration and distrust may very well result from their political preferences on immigration not being heeded by the mainstream parties.

Anti-immigrant sentiment is a handy but slightly awkward catch-all term for negative attitudes towards immigrants, immigration, and immigration policies. In a seminal contribution, Rydgren (2008) distinguishes between “immigration sceptics”, “xenophobes”, and “racists”. For Rydgren (2008, pp. 741-744), xenophobes have a latent disposition to react with fear and aversion to outsiders, but this only becomes an issue if the number of outsiders is too high by some subjective standard, or if the outsiders otherwise seem to pose a threat to in-group. Racists always hold outsiders in contempt irrespective of any exposure to “strangers”, with “classic” racism being based on notions of biological hierarchies, whereas “modern” or “cultural” racism subscribes to the idea of incompatible but (nominally) coequal cultures.1 Finally, immigration sceptics want to reduce the number of immigrants in their native country (Rydgren, 2008, p. 738), but not necessarily because they hold racist or xenophobic attitudes. As Rydgren (2008, p. 740) suggests, the most plausible structure for these attitudes is a nested one, where xenophobes form a subgroup of the immigration sceptics and racists form a subgroup of the xenophobes.

The distinction between immigration sceptics, xenophobes, and racists is particularly useful because not all Radical Right voters are full-blown racists. Moreover, many of the approaches that are discussed in the literature may help to explain deep-seated, stable racism but not necessarily a more specific and volatile scepticism regarding current immigration policies.

“Deep” explanations for Radical Right support have been developed since at least the 1930s. The monographs and articles on the roots of rightist political views fill several libraries by now and any attempt to classify them is crude by necessity. Nonetheless, it makes sense to distinguish between three very broad groups.

A first class of explanations focuses on personality traits2, with authoritarianism being the most prominent amongst them. Authoritarianism as a concept is most closely associated with the (controversial) Berkeley Study (Adorno et al., 1950) but has more recently been modernised and promoted by Bob Altemeyer (1981; 1996). For Altemeyer, Right-Wing Authoritarianism (RWA) consists of three key elements: a desire to submit to established and legitimate authorities (authoritarian submission), a hostility towards deviants and other out-groups (authoritarian aggression), and an exaggerated respect for traditions and social norms (conventionalism).

Authoritarianism and similar concepts such as dogmatism (Rokeach, 1960) or tough-mindedness (Eysenck, 1954) go a long way towards explaining the relevance of xenophobia and the appeal of other right-wing ideas and movements to some voters, but there are a few important caveats. First, compared to classic right-wing extremist groups, authoritarianism is much less important for the ideology of the modern populist Radical Right (Mudde, 2007). Unlike the Fascists or the Nazis of the interwar period, the most successful of these parties do not seek to replace democracy by some authoritarian type of regime but rather promote a narrow, “illiberal” concept of democracy. Second, support for the Radical Right has surged (and sometimes declined) over relatively short periods, whereas personality traits are by definition stable. They may thus help us to explain why there is potential for authoritarian parties in the first place. The exploitation of this potential by political entrepreneurs and the channeling of this general hostility towards out-groups into a more specific anti-immigrant sentiment, however, are political processes that must be understood by means of different concepts.

Theories of group conflict and deprivation form a second and more immediately relevant cluster of explanations. This cluster can be subdivided in four broad categories

  1. Theories of “realistic group conflict” (RGCT) and “ethnic competition” (EC)
  2. Theories of “status politics” and “symbolic racism”
  3. Theories of “social identity”
  4. Theories of “scapegoating”

The ordering is deliberate: From the top to the bottom, these approaches put less and less emphasis on material conflicts and conscious mental processes and instead focus on the importance of visceral hostility (which might still be induced by political entrepreneurs) towards members of the out-group.

Both for proponents of RGCT (see Jackson, 1993 for a review) and EC (e.g. Bélanger and Pinard, 1991), tensions between (ethnic) groups are rooted in conflicts over the distribution of material resources in a society, which is often perceived as unfair. The main difference between both approaches is that RGCT is more interested in the micro-dynamics of group psychology whereas EC is primarily concerned with the societal level. Either way, the distributional conflict is couched in collective terms, even if the resource in question is a personal good (e.g. a secure job). Both strands of the literature as well as the other approaches discussed in this section are therefore closely related to classic theories of collective relative deprivation (Runciman, 1966, pp. 33-34, see also Ellemers, 2002 and Taylor, 2002). While students of electoral behaviour rarely investigate the lengthy and complex causal chains that link social change, group dynamics, and inter-ethnic contacts to psychological processes, feelings of material threat that is allegedly posed by immigrants have become a staple explanatory variable for analysing anti-immigrant sentiment, and by implication the Radical Right’s electoral support. On the contextual level, (potential) exposure to material threats if often captured by incorporating macro-economic variables in statistical models of Radical Right voting (see below).

Similarly, proponents of the “status politics” approach (e.g. Hofstadter, 2002b) argue that (recent) immigrants are perceived as a collective threat by members of the in-group. Here, the collective good in question is not a material one but rather the collective social status of the in-group, or the cultural hegemony of their values, norms, and social practices (Hofstadter, 2002a) – ideas which in turn bear some resemblance with the idea of “symbolic racism” (Kinder and Sears, 1981; see Walker, 2001 for a critical review of this and some related concepts). Again, psephologists usually take the alleged causal mechanisms for granted and focus on the effect of perceived cultural threats on anti-immigrant sentiment and the Radical Right vote.

(Modern) theories of social identity provide another approach for explaining anti-immigrant sentiment. “Social Identity Theory” (SIT) and its successor, “Self-Categorisation Theory” (SCT), were developed in response to an empirical puzzle: Even in a “minimal effects” experimental setting where subjects were randomly assigned to socially meaningless groups, where there was no interaction whatsoever between subjects, and no material incentive to put members of the out-group at a disadvantage, a large proportion of subjects was willing to discriminate against the outsiders. Tajfel and Turner (1986) interpret this unexpected finding as the result of a cognitive process during which one’s social identity becomes the yardstick for assessing a given situation, whereas the importance of one’s personal identity declines. As a corrolary, members of the out-group are subject to a process of stereotyping. In combination with an innate desire for positive distinctiveness, stereotyping and self-stereotyping can bring about discrimination and prejudice against out-group members, because they represent one avenue towards a more positive self-image. However, whether discrimination actually occurs depends on a number of conditions (Reynolds and Turner, 2001, p. 166). Crucially, these mechanisms are independent of any material or cultural threat that the out-group may seem to pose to the members of the in-group.

Once more, psephologists have mostly ignored the details and instead focused on the impact of a single variable (identity) on Radical Right voting intentions, and even this alleged mechanism is often problematic, because most items available in representative surveys do not capture the complexity of the concept. Nonetheless, SIT/SCT has the potential to make a crucial contribution to a fuller explanation of the Radical Right vote: While most group dynamic processes must remain under the radar of mass surveys, SIT/SCT informs experimental and observational research on the conditions under which stereotypes and prejudices that may result in anti-immigrant sentiment become activated. It also provides a useful framework for the analysis of party documents and social and mass media content, which play an ever more important role in the study of Radical Right electoral support.

Finally, theories of “scapegoating” need to be addressed. These hark back to the late 1930s (Dollard et al., 1939) and have even older roots in the Sumner’s early work on ethnocentrism (Sumner, 1906), maintain that members of the ethnic majority who experience feelings of frustration and deprivation that are objectively unrelated to the presence of other ethnic groups nonetheless turn towards immigrants simply because those provide a conveniently defenceless target for the in-group members’ aggression. Due to the “cognitive turn” in social psychology, theories of scapegoating have somewhat fallen out of fashion, and for the applied psephologist relying on secondary data analysis, the result of simple scapegoating will often be indistinguishable from the more complex stereotyping processes.

All theories of group conflict are complemented by the “contact hypothesis”, which maintains that under certain favourable conditions, inter-ethnic contacts (which often presuppose immigration) can reduce prejudice (Pettigrew and Tropp, 2008) and hence anti-immigrant sentiment. Some of the newer research aims at incorporating the contact hypothesis by either using micro-level information on inter-ethnic contact or by deriving the probability of such contacts from small-area data on the spatial distribution of ethnic groups. Unfortunately, both approaches are subject to endogeneity bias, because voters who are less prejudiced are more likely to seek inter-ethnic contacts.

2.3.2 Anti Post-Materialism and Other Social Attitudes

A Silent Counter-Revolution? Immigration emerged as the core issue of the Radical Right in Western Europe and Australia in the mid-1980s, making anti-immigrant sentiment the single most important attitudinal driver of Radical Right support. In Central and Eastern Europe (CEE), hostility towards ethnic minorities seems to act as the functional equivalent. But very few RRPs have ever been single-issue parties (Mudde, 1999). Many of them have a broader right-wing agenda, and Radical Right support has been linked to a host of other attitudes than anti-immigrant sentiment.

The Rise of the RRP family in the 1980s and early 1990s has therefore been interpreted as a reaction to large-scale social change.3 In a seminal article, Ignazi (1992) claims that these new right-wing parties embody the backlash against post-materialism and the New Left politics which it has inspired: a “silent counter-revolution”. Similarly, Kitschelt (1995) has argued that globalisation has created a new class of authoritarian private-sector workers, who combine market-liberal preferences with an authoritarian outlook on society and find their political representation in the Radical Right. While the market-liberalism of the Radical Right’s electorate remains elusive (Kitschelt and McGann, 2003; Arzheimer, 2009b; Mayer, 2013), it has become ever more evident that non-traditional working-class voters form the Radical Right’s core electoral base (see the contributions in Rydgren, 2013).

Moral conservatism, homophobia and more generally anti-postmaterialism may have played a role, too (and probably are still relevant for party members and activists), but they seem to be much less important than they were for the classic Extreme Right, at least in some countries. As early as 1988, the French FN voters were slightly “more permissive in sexual matters” than the voters of the mainstream right (Mayer and Perrineau, 1992, p. 130). 25 years later, the FN is lead by a single mother of three, twice divorced (Mayer, 2013, p. 175), whose attendance at homophobic rallies seems to be more a matter of strategy than of convictions. Even more strikingly, the Lijst Pim Fortuyn, the Netherland’s first successful RRP, was founded and led by an openly gay libertine (Akkerman, 2005), and its de facto successor, the PVV, claims that defending the freedom of the LGBT community is part of their commitment to Dutch values. But even in the Netherlands, culturally progressive values are not an important driver of the RRP vote, at least not when anti-immigrant sentiment is controlled for (De Koster et al., 2014). One way or the other, for many RRP voters in Western Europe, homophobia and social conservatism do not seem to matter too much any more.

Religion The Extreme Right of the interwar years could be roughly divided in two groups (Camus, 2007): In some cases (most prominently Portugal and Spain), they aligned themselves with the most authoritarian and reactionary elements of the (Catholic) church. In other instances (e.g. Germany and Austria after the “Anschluss”), the Extreme Right distanced itself from Christianity and/or relied on the traditional loyalty of the (Protestant) church to the political leadership.

Today’s RRPs have inherited some of this historical baggage. While religious conservatism may inspire some of their members and voters (see the previous section), church leaders have often spoken out against the Radical Right’s anti-immigrant policies. To complicate matters further, the Radical Right is now often couching their anti-immigrant message in terms of a clash between “Western Values” and “Islam”. In a sense, criticising Islam abroad and at home has become the socially acceptable alternative to more openly xenophobic statements (Zúquete, 2008).

In a bid to disentangle this relationship, Arzheimer and Carter (2009a) estimate a Structural Equation Model of religiosity, anti-immigrant sentiment, party identification with mainstream right parties, and Radical Right voting intentions in seven West European countries. Their results show that in the early 2000s, religiosity had no significantly positive or negative effect on either anti-immigrant sentiment or RRP voting intentions. Religious people are, however, much more likely to identify with a mainstream right party, which in turn massively reduces the likelihood of an RRP vote. Using a slightly different model and data collected in 2008, Immerzeel, Jaspers, and Lubbers (2013) arrive at very similar conclusions.

Crime Law and order politics is traditionally the domain of both the mainstream and the Radical Right (Bale, 2003), with some authors going as far as saying that the Radical Right “owns” the crime issue (Smith, 2010). At any rate, talking about crime and immigration is a core frame of Radical Right discourses (Rydgren, 2008). Data from the European Social Survey clearly show that many West Europeans associate immigration with crime, and panel data from Germany suggest that that worries about crime have a substantial effect on anti-immigrant sentiment (Fitzgerald, Curtis, and Corliss, 2012). Many authors subsume such immigration-related crime fears into the larger complex of subjective threat that immigration poses to susceptible voters. Others model the effect of objective crime figures on the Radical Right vote (see below).

Euroscepticism Mudde (2007) has convincingly argued that nativism, i.e. the desire for an ethnically homogeneous nation state, forms the core of the Radical Right’s ideology. Accordingly, RRPs reject the European Union as a general rule, although Vasilopoulou (2011) has demonstrated that opposition to the European projects is by no means uniform within the Radical Right camp. Unsurprisingly, individual eurosceptic attitudes come up as predictors of Radical Right voting intentions in some studies (e.g. Arzheimer, 2009a; Brug, Fennema, and Tillie, 2005), although anti-immigrant and even general dissatisfaction with the elites exert a stronger effect (Werts, Scheepers, and Lubbers, 2013). Given that at least some countries feature leftist eurosceptic parties whose voters hold opinions which differ markedly from those of the RRP voters (Evans, 2000; Elsas and Brug, 2015), it seems safe to assume that euroscepticism per se does not predispose voters to support the Radical Right but needs to be linked to more general nativist beliefs.

3 Meso-level Factors

3.1 Party Strength

It is more than plausible that organisational assets and other party resources including leadership should be important pre-conditions for RRP success, but in applied research, they are often overlooked, because they are difficult to measure and tend not to vary too much over time. Carter (2005) is one of the very few studies that systematically incorporates party strength into a quantitative model of Radical Right support. Distinguishing between “(1) weakly organised, poorly led and divided parties, (2) weakly organised, poorly led but united parties, and (3) strongly organised, well-led but factionalised parties” she finds that the latter group performs substantially better than the former two (Carter, 2005, pp. 98-99).

David Art’s qualitative study of Radical Right party organisations in twelve West European countries (Art, 2011) provides an important complement to this finding. Taking a longitudinal perspective, Art shows that prospective RRPs need to attract ideologically moderate, high-status activists early in the process to build sustainable party structures and become electorally viable. Otherwise, there is a high probability that they will be subject to factionalism and extremism, which renders them unattractive for most voters.

While Art and Carter compare parties and countries, it is also possible to incorporate information on organisational strength in a within-country model of Radical Right voting. Erlingsson, Loxbo, and Öhrvall (2012) identify a positive effect of “local organisational presence” on the vote of the Sweden Democrats in the 2006 and 2010 elections. One the one hand, this modelling strategy is advantageous, because it maximises the number of cases and can avoid aggregation bias. On the other hand, the validity of Erlingsson, Loxbo, and Öhrvall’s findings is threatened by endogeneity: parties will be more inclined to invest resources and prospective activists will be more inclined to create and join a local organisation if there is a prospect of success in the first place.

3.2 Party Ideology

As a general rule, RRPs take political positions that are in some ways more radical than what the mainstream right is offering, but the ideological heterogeneity of the RRPs is sometimes baffling. It took therefore more than a decade to establish some sort of consensus that these parties do indeed form a party family (Mudde, 1996), and twenty years down the line, scholars still find it difficult to agree on a name for this family, although “Radical Right” is arguably the most popular label at the moment. There are various attempts to distinguish between subgroups within this large cluster. Mudde (2007) identifies a small number of parties that he classifies as “Extreme Right”, i.e. aiming at replacing democracy with some authoritarian system. Similarly, Golder (2003b) draws a line between “populist” and “neo-fascist” parties. Summarising electoral data from Western Europe for the 1970-2000 period, Golder (2003b, p. 444) notes that support for the “neo-fascist” group was very limited in the first place and further declined over time, whereas the appeal of the “populist” parties has grown enormously since they emerged in the 1980s. By and large, this finding still holds today: In Western Europe, where democracy has become “the only game in town”, the vast majority of voters deems openly non-democratic parties unelectable.4 In other European countries where democracy is newer, however, even overtly extremist parties may be electorally successful (see Ellinas 2013; Ellinas 2015 for Greece, Mudde, 2005 and Mareš and Havlík, 2016 for Central and Eastern Europe after 1990, and Stojarová, 2012 for former Yugoslavia).

A different classification, which is not based on the fundamental question of support for democracy but rather on policy positions, was developed by Herbert Kitschelt in his seminal monograph (Kitschelt, 1995). Kitschelt aims at locating RRPs in a policy space that is spanned by two dimensions: a purely economic left-right axis (state vs market) and a more complex dimension that encompasses issues of citizenhood (“group”, see Kitschelt, 2013) on the one hand and individual and collective decision making (“grid”) on the other. Originally, Kitschelt claimed that the then unusual blend of market-liberalism and authoritarian social conservatism represented an “electoral winning formula”. While this may still hold in the US, RRP voters in Western Europe are no longer interested in market liberalism (Lange, 2007; Arzheimer, 2009b), if they ever were. Moreover, electorally successful RRPs have recently de-emphasised their positions on the “grid” (authoritarian) dimension (Kitschelt, 2013, see also section 2.3.2).

3.3 Party System Factors

RRPs do not operate in a vacuum. While they may have a degree of control over their leadership/candidates, their organisational structure, and their ideology, they are but one part of the larger party system, and the words and actions of other parties may have as big an impact on the Radical Right’s electoral fortunes as anything that the RRP themselves do. Presumably, there are two major and partly competing mechanisms at work: From a Downsian logic, it follows that a successful RRP will eventually emerge if there is a demand for more restrictive (migration) policies, which is not satisfied by the existing parties in general and the mainstream right in particular. In this view, a mainstream right party that is soft on immigration and/or the existence of a formal “Grand Coalition” between centre-left and centre-right parties will have a positive impact on the Radical Right vote.

The psychological counter-argument is that political demands are rarely fixed, and that an elite consensus to de-emphasise immigration as a political issue (Zaller, 1992) and to impose a cordon sanitaire might rob the Radical Right of its potential support. Whether this latter strategy is politically feasible is quite a different question. Centre-right parties may have strong incentives to shore up the Radical Right in a bid to strengthen the rightist bloc (Bale, 2003). Centre left parties may want to split the right-wing vote: Mitterand’s decision to hold the 1986 legislative election under PR and Kreisky’s kind words for Haider are cases in point.

The empirical evidence is somewhat mixed. Arzheimer and Carter (2006) find no statistical effect of the mainstream right’s ideological position, or of ideological convergence between the centre left and centre right, but note a substantial positive impact of Grand Coalitions. This result, however, may be shaped by the inclusion of respondents from Austria, which features a long and almost unique history of Grand Coalitions and a consistently strong RRP. On the other hand, Lubbers, Gijsberts, and Scheepers (2002) report that a restrictive “immigration climate” (operationalised as the vote-share weighed average of the other parties positions on immigration) increases the likelihood of a Radical Right vote. Using a slightly different approach that is derived from Zaller’s work, Arzheimer (2009a) notes that the Radical Right benefits from an increasing salience of their issue, regardless of the direction of the statements, and Dahlstroem and Sundell (2012) find a positive effect of anti-immigrant positions held by local politicians from other parties. Again, endogeneity could potentially be a problem in these studies, although this seems less likely in the case of data based on an expert survey (Lubbers, Gijsberts, and Scheepers, 2002) or party manifestos (Arzheimer and Carter, 2006; Arzheimer, 2009a).

3.4 Social Capital

In line with classic theories of the “mass society” (Kornhauser, 1960; Bell, 2002), the rise of the Radical Right has sometimes been linked to widespread feelings of isolation and Anomia. If this relationship holds, higher levels of Social Capital (Putnam, 1993) should curb support for the Radical Right.

Once more, the empirical evidence is limited and contradictory. In a series of case studies in Western and Eastern Europe, Rydgren (2009; 2011) finds that membership in civic organisations does not reduce the probability of casting a vote for the Radical Right. But this does not necessarily disconfirm the Social Capital hypothesis, because Social Capital is not an individual-level but rather a meso-level concept. Coffé, Heyndels, and Vermeir (2007), on the other hand, demonstrate in their model of RRP voting in Flanders that the Vlaams Blok performs significantly worse in municipalities with higher levels of associational life, ceteris paribus, but this finding might be the result of aggregation bias as the authors rely exclusively on census data and electoral counts. Finally, Fitzgerald and Lawrence (2011) combine micro and meso data to estimate a multi-level model of support for the Swiss People’s Party. Even after controlling for a host of variables at the person and at the “commune” level, they find that a municipality’s “social cohesion index” has a substantial positive effect on the probability of a vote for the Radical Right. But while their research design and statistical model are close to ideal, it is not quite clear what they actually measure. Their index includes the proportion of the working population who are not commuters, the proportion of residents who speak the most common language in a given municipality, and the percent of residencies inhabited by their owners. These variables may relate to “bonding” Social Capital, which could explain the positive effect on the RRP vote, but further research is clearly needed.

4 Macro-level Factors

4.1 Institutional Factors

The impact of institutional factors – most prominently, features of the electoral system, decentralisation, and welfare state protection – are very difficult to assess, because they change very slowly or not at all over time and are hence highly correlated with any idiosyncratic unit (=country) effects. Somewhat unsurprisingly, empirical findings are mostly contradictory and inconclusive. As regards electoral systems, Jackman and Volpert (1996) claim that the Radical Right benefits from lower electoral thresholds, but Golder (2003a) argues that this conclusion is based on an erroneous interpretation of an interaction effect and a somewhat idiosyncratic data collection effort. In the same vain, Carter (2002) reports that electoral support for the Radical Right is unrelated to the type of electoral system that is in place in a given election, whereas Arzheimer and Carter (2006) find a positive effect of more disproportional systems but maintain that this might be an artefact.

As regards features of the welfare state, Swank and Betz (2003) find that higher level of welfare state protection seem to reduce the appeal of the Radical Right. However, their analysis is based exclusively on macro data. Using a more specific indicator (generosity of unemployment benefits) and micro data, Arzheimer (2009a) finds that more generous benefits, which may cause “welfare chauvinism”, are linked to higher levels of support but only if levels of immigration are below average (see also next section).

4.2 Immigration and Unemployment

For obvious reasons, the two macro-level variables whose effects have been most extensively studied are immigration, unemployment, and their interaction: a high immigration / high unemployment situation represents perhaps the most clear-cut scenario for ethnic competition for scarce jobs. Nonetheless, the findings are far from conclusive, as can be seen by looking at two of the first comprehensive comparative studies: While Jackman and Volpert (1996) find a substantial positiveeffect of aggregate unemployment on the Radical Right vote, Knigge (1998), who uses a design that is quite similar, reports a negative effect. So do Arzheimer and Carter (2006). Lubbers, Gijsberts, and Scheepers (2002), in their first multi-level model of Radical Right voting in Western Europe, find no significant relationship between the unemployment rate the Radical Right voting intentions, whereas Golder (2003b), whose analysis is once more based on aggregate data, reports a positive (main) effect as well as a positive interaction between unemployment and immigration. Finally, Arzheimer’s (2009a) results from a rather complex multi-level model of Radical Right voting suggest that unemployment may have a positive effect under some scenarios when unemployment benefits are minimal and contributing factors (both individual and contextual) are already favourable.

Although measures for immigration are hardly ideal and differ across studies, results for the effect of immigration are less equivocal: Knigge (1998), Lubbers, Gijsberts, and Scheepers (2002) , Golder (2003b), Swank and Betz (2003), and Arzheimer and Carter (2006) all find a positive effect of (national) immigration figures on the likelihood of a Radical Right vote. Arzheimer (2009a) by and large confirms this, although with an important qualification: In his study, the interaction between unemployment and immigration is negative so that a high levels of both variables, their effects do not reinforce each other any more but rather hit a ceiling. Moreover, generous unemployment benefits reduce the effect of immigration.

4.3 Crime

Like immigration and unemployment, high crime rates are supposed to benefit the Radical Right, but there is not much empirical evidence to back up this claim. Coffé, Heyndels, and Vermeir (2007) conducted one of the first studies that tests the alleged relationship. In an aggregate model of Vlaams Blok support in Flemish municipalities, they find that high crime rates increase the likelihood of the Vlaams Blok contesting an election, presumably because the party anticipates higher levels of support. However, once this selection mechanism is accounted for, crime has no positive effect on the Vlaams Blok’s result.

The study by Coffé, Heyndels, and Vermeir has three distinct advantages: It models the decision to compete in an election and the results of that decision separately, it is built on a large number of cases, and the level of aggregation is low. But unfortunately, their design does not allow for comparisons across time or political systems. In a sense, the article by Smith (2010) provides the complement to their work: Smith studies the relationship between support for the Radical Right and crime rates at the highest possible level of aggregation by analysing 182 national parliamentary elections that were held in 19 Western European countries between 1970 and 2005. Controlling for unemployment, inflation, immigration, and various interactions, he finds that higher crime rates are associated with stronger support for the Radical Right. This relationship becomes stronger if immigration rates are higher.

Finally, the contribution by Dinas and Spanje (2011) specify a multi-level model of Radical Right voting in the Netherlands in 2002. Like in the case of Coffé, Heyndels, and Vermeir (2007), their results are confined to one election in a single country. As they combine individual and contextual data, there is no aggregation bias, and they can even tease apart the effects of objective crime rates and subjective attitudes towards crime. Their results suggest that the effects of crime and immigration do not operate across the board but rather only affect those citizens who perceive a link between the two.

4.4 Media

One final variable at the macro level that attracts considerable interest is the media coverage of the Radical Right’s issues. While voters will be exposed to crime, immigration and unemployment to one degree or another, media reports may have a stronger effect than personal experiences or non-experiences via two alleged mechanisms: Theories of agenda setting claim that the media, by focusing on certain topics, select a handful of politically relevant issues from a much larger pool of problems. Those issues on the agenda then serve as yardsticks for evaluating parties, an effect known as priming (Scheufele and Tewksbury, 2007). In extreme cases, an issue may become so closely associated with a party that this party “owns” the issue (Petrocik, 1996) and will almost automatically benefit whenever it achieves a high rank on the agenda. Green parties and the environment are an oft-cited example, but the Radical Right and immigration have become a close second in the eyes of many observers (Meguid, 2005).

Notwithstanding the importance of the alleged nexus between media coverage and Radical Right support, the evidence is limited once more. The main reason for this is that data on media content are difficult to come by and expensive to produce in the first place. This is slowly changing now, with automated coding methods and open data bases such as GDELT providing new avenues for research, but even so, matching media with micro-level data is next to impossible, because mass opinion surveys do not normally collect detailed (i.e. per item) information on media consumption. Most of the existing research is therefore based on aggregated (i.e. time-series) data.

In their pioneering study, Boomgaarden and Vliegenthart (2007) find a positive relationship between salience of immigration in Dutch media and aggregate support for Radical Right parties during the 1990-2002 period, net of any changes that can be ascribed to the unemployment and immigration rates and their interaction. This article is complemented by Koopmans and Muis (2009), who focus on the end of that period (i.e. Pim Fortuyn’s 2002 campaign) and aim to identify a number of “discursive opportunities” that facilitated Fortuyn’s breakthrough. In another study that resembles their 2007 piece (Boomgaarden and Vliegenthart, 2009), Boomgaarden and Vliegenthart can further demonstrate a link between news content and anti-immigrant sentiment in Germany for the 1993-2005 period.

Finally, in a bid to overcome the dearth of micro-level data on media consumption from mass surveys as well as the limits of the ex-post-facto design, interest in in experimental studies has grown considerably over the last decade. One such study is that by Sheets, Bos, and Boomgaarden (2015), who exposed members of an online-access panel to an synthetic news article. Some small parts of this article were systematically varied to provide “cues” that would prime the issues of immigration, anti-politics, and the RRP itself. While Sheets, Bos, and Boomgaarden can demonstrate some effects of these cues on anti-immigrant attitudes, political cynicism, and ultimately on PVV support, some question marks remain. First, the effects on anti-immigrant attitudes are weak compared to those on political cynicism. Second, like with any experimental intervention, it is not clear if effects of a similar magnitude occur “in the wild”, and if so, how long they persist. Third, the experiment was designed in a way that means that the immigration and anti-politics cues were always combined with an RRP cue, which will in all likelihood bias the estimates for their respective effects either upwards or downwards. Clearly, further (cross-national) research is needed.

5 Small Area Studies

By now it should be clear that nearly all authors in the field treat support for the Radical Right as a multi-faceted phenomenon that must be explained at multiple levels, with unemployment, immigration, and political factors and media cues being the most prominent contextual variables. Most studies measure these variables at the national level, but living conditions in European states vary considerably across regions, so designs that compare provinces, districts or even neighbourhoods within countries are becoming more and more prominent. One of the first of these studies was conducted by Bowyer (2008), who looks at electoral returns for the British National Party (BNP) in several thousand wards in the 2002/2003 local elections in England. He finds that the BNP was strongest in predominantly white neighbourhoods that are embedded within districts which are characterised by the presence of large ethnic minorities, a pattern that has been described as the “halo effect” (Perrineau, 1985). Economic deprivation (though not necessarily unemployment) also played a role. Similarly, Rydgren and Ruth (2011), who analyse support for the Sweden Democrats in the 2010 election across the country’s 5668 voting districts, show that the party did better in poorer districts with bigger social problems. Once these factors are controlled for, there is also some evidence for the existence of a “halo effect”.

Other studies have focused on units that are larger but politically more meaningful than census districts or electoral wards, e.g. departements, provinces, or sub-national states (Kestilä and Söderlund, 2007; Jesuit, Paradowski, and Mahler, 2009), accepting possible aggregation bias in exchange for the ability to include political and/or media variables in the model. The former study reports positive effects of unemployment and some institutional variables but no effect of immigration, whereas the latter identifies some complex interactions that link immigration and unemployment to Radical Right support via an increase in inequality and a lack of social capital.

Studies in small(ish) areas are currently one of the most promising avenues of research into the Radical Right vote, be it on the level of subnational political units or in even smaller tracts. Either way, researchers need to account for the fact that an increasing number of voters are either immigrants or the offspring of immigrants, who will be disinclined to support the Radical Right. Estimates from small area studies that are based on aggregate data will therefore be biased downward (Arzheimer and Carter, 2009b). Hence, multi-level analyses that combine micro data with information on local living conditions are the way forward in this particular branch of research.

6 Conclusions

Over the last three decades, Radical Right parties have become a permanent feature of most European polities. Their rise, persistence, and decline can be quite well explained by the usual apparatus of electoral studies. On the micro level, the most important factors are value orientations, attitudes towards social groups, candidates and political issues as well as (the lack of) party identifications. At the macro level, social change (broadly defined) undoubtedly plays an important role, while parties, the media and all other sorts collective actors operate at the meso-level in between.

Because RRPs are often perceived as divisive, disruptive, or outright dangerous, a great deal of intellectual energy has been spent looking for “deeper” explanations. And indeed, there can be very little doubt that the presence or absence of immigrants and immigration, the frequency and nature of contacts between the immigrants and the native population, and the way immigration is framed by other political actors and the media is a major contributing factor to Radical Right support. However, given that immigration, ethnic tensions, and RRP actors are almost ubiquitous in Western societies, their success is not a major surprise. Ultimately, trying to understand why they are not successful in some cases might be more rewarding, both politically and intellectually.


Adorno, Theodor W. et al. (1950). The Authoritarian Personality. New York: Harper.

Akkerman, Tjitske (2005). “Anti-immigration parties and the defence of liberal values. The exceptional case of the List Pim Fortuyn”. In: Journal of Political Ideologies 10.3, pp. 337–354. DOI: 10 . 1080 / 13569310500244354.

Altemeyer, Bob (1981). Right-Wing Authoritarianism. Winnipeg: The University of Manitoba Press.

— (1996). The Authoritarian Specter. Cambridge: Harvard University Press.

Art, David (2011). Inside the Radical Right. The Development of Anti-Immigrant Parties in Western Europe. Cambridge: Cambridge University Press.

Arzheimer, Kai (2009a). “Contextual Factors and the Extreme Right Vote in Western Europe, 1980-2002”. In: American Journal of Political Science 53.2, pp. 259–275. DOI: 10.1111/j.1540-_5907.2009.00369.x.

— (2009b). “Protest, Neo-Liberalism or Anti-Immigrant Sentiment: What Motivates the Voters of the Extreme Right in Western Europe?” In: Comparative Governance and Politics 2, pp. 173–197. DOI: 10.1007/ s12286-_008-_0011-_4.

Arzheimer, Kai and Elisabeth Carter (2006). “Political Opportunity Structures and Right-Wing Extremist Party Success”. In: European Journal of Political Research 45, pp. 419–443. DOI: 10.1111/j.1475- _6765.2006.00304.x.

— (2009a). “Christian Religiosity and Voting for West European Radical Right Parties”. In: West European Politics 32.5, pp. 985–1011. DOI: 10. 1080/01402380903065058.

— (2009b). “How (Not) to Operationalise Subnational Political Opportunity Structures: A Critique of Kestilä and Söderlund’s Study of Regional Elections”. In: European Journal of Political Research 48.3, pp. 335–358. DOI: 10.1111/j.1475-_6765.2009.00842.x.

Bale, Tim (2003). “Cinderella and Her Ugly Sisters: The Mainstream and Extreme Right in Europe’s Bipolarising Party Systems”. In: West European Politics 26, pp. 67–90.

Bélanger, Sarah and Maurice Pinard (1991). “Ethnic Movements and the Competition Model. Some Missing Links”. In: American Sociological Review 56, pp. 446–457.

Bell, Daniel, ed. (2002). The Radical Right. Third Edition. With a New Introduction by David Plotke. New Brunswick: Transaction Publishers.

Beyme, Klaus (1988). “Right-Wing Extremism in Post-War Europe”. In: Right-Wing Extremism in Western Europe. Ed. by Klaus Beyme. London: Frank Cass, pp. 1–18.

Billiet, Jaak and Hans Witte (1995). “Attitudinal Dispositions to Vote for a ’New’ Extreme Right-Wing Party: The Case of ’Vlaams Blok’”. In: European Journal of Political Research 27, pp. 181–202.

Boomgaarden, Hajo G. and Rens Vliegenthart (2007). “Explaining the Rise of Anti-Immigrant Parties: The Role of News Media Content”. In: Electoral Studies 26.2, pp. 404–417.

— (2009). “How News Content Influences Anti-Immigration Attitudes: Germany, 1993-2005”. In: European Journal of Political Research 48.4, pp. 516–542. DOI: 10.1111/j.1475-_6765.2009.01831.x. URL:http://www3.interscience.wiley.com/journal/122273548/abstract.

Bowyer, Benjamin (2008). “Local context and extreme right support in England. The British National Party in the 2002 and 2003 local elections”. In: Electoral Studies 27.4, pp. 611–620.

Brug, Wouter van der and Meindert Fennema (2003). “Protest or Mainstream? How the European Anti-Immigrant Parties Developed into two Separate Groups by 1999”. In: European Journal of Political Research 42, pp. 55–76.

Brug, Wouter van der, Meindert Fennema, and Jean Tillie (2000). “Anti-Immigrant Parties in Europe: Ideological or Protest Vote?” In: European Journal of Political Research 37.1, pp. 77–102.

— (2005). “Why Some Anti-Immigrant Parties Fail and Others Succeed. A Two-Step Model of Aggregate Electoral Support”. In: Comparative Political Studies 38, pp. 537–573.

Brug, Wouter van der and Anthony Mughan (2007). “Charisma, Leader Effects and Support for Right-Wing Populist Parties”. In: Party Politics 13.1, pp. 29–51. DOI: 10.1177/1354068806071260.

Campbell, Angus (1960). “Surge and Decline. A Study of Electoral Change”. In:

Camus, Jean-Yves (2007). “The European Extreme Right and Religious Extremism”. In: Central European Political Studies Review 9.4, pp. 263–279. URL: http://www.cepsr.com/clanek.php?ID=317.

Carter, Elisabeth (2002). “Proportional Representation and the Fortunes of Right-Wing Extremist Parties”. In: West European Politics 25, pp. 125–146.

— (2005). The Extreme Right in Western Europe. Manchester, New York: Manchester University Press.

Coffé, Hilde, Bruno Heyndels, and Jan Vermeir (2007). “Fertile grounds for extreme right-wing parties: Explaining the Vlaams Blok’s electoral success”. In: Electoral Studies 26.1, pp. 142–155. DOI: 10.1016/j. electstud.2006.01.005.

Coleman, James S. (1994). Foundations of Social Theory. Cambridge, London: The Belknap Press of Harvard University Press.

Dahlstroem, Carl and Anders Sundell (2012). “A Losing Gamble. How Mainstream Parties Facilitate Anti-Immigrant Party Success”. In: Electoral Studies 31.2, pp. 353–363. DOI: 10.1016/j.electstud.2012. 03.001.

De Koster, Willem et al. (2014). “Progressiveness and the New Right. The Electoral Relevance of Culturally Progressive Values in the Netherlands”. In: West European Politics 37.3, pp. 584–604. DOI: 10.1080/01402382. 2013.814963.

Dinas, Elias and Joost van Spanje (2011). “Crime Story. The Role of Crime and Immigration in the Anti-Immigration Vote”. In: Electoral Studies 30.4, pp. 658–671. DOI: 10.1016/j.electstud.2011.06.010.

Dollard, John et al. (1939). Frustration and Aggression. New Haven, London: Yale University Press.

Eatwell, Roger (2000). “The Rebirth of the ”Extreme right” in Western Europe?” In: Parliamentary Affairs 53, pp. 407–425.

— (2005). “Charisma and the Revival of the European Extreme Right”. In: Movements of Exclusion. Radical Right-wing Populism in the Western World. Ed. by Jens Rydgren. Hauppauge: Nova Science, pp. 101–120.

Ellemers, Naomi (2002). “Social Identiy and Relative Deprivation”. In: Relative Deprivation. Specification, Development, and Integration. Ed. by Iain Walker and Heather J. Smith. Cambridge: Cambridge University Press, pp. 239–264.

Ellinas, Antonis A. (2013). “The Rise of Golden Dawn. The New Face of the Far Right in Greece”. In: South European Society and Politics 18.4, pp. 543–565.

— (2015). “Neo-Nazism in an Established Democracy. The Persistence of Golden Dawn in Greece”. In: South European Society and Politics 20.1, pp. 1–20. DOI: 10.1080/13608746.2014.981379.

Elsas, Erika van and Wouter van der Brug (2015). “The changing relationship between left–right ideology and euroscepticism, 1973–2010”. In: European Union Politics 16.2, pp. 194–215. DOI: 10 . 1177 / 1465116514562918.

Erlingsson, Gissur Ó., Karl Loxbo, and Richard Öhrvall (2012). “Anti-Immigrant Parties, Local Presence and Electoral Success”. In: Local Government Studies 38.6, pp. 817–839. DOI: 10.1080/03003930.2012. 740411.

Evans, Jocelyn (2000). “Contrasting Attitudinal Bases to Euroscepticism amongst the French Electorate”. In: Electoral Studies 19, pp. 539–561.

Eysenck, Hans Jürgen (1954). The Psychology of Politics. London: Routledge, K. Paul.

Fitzgerald, Jennifer, K. Amber Curtis, and Catherine L. Corliss (2012). “Anxious Publics: Worries About Crime and Immigration”. In: Comparative Political Studies 45.4, pp. 477–506. DOI: 10.1177/ 0010414011421768.

Fitzgerald, Jennifer and Duncan Lawrence (2011). “Local cohesion and radical right support: The case of the Swiss People’s Party”. In: Electoral Studies 30.4, pp. 834–847. DOI: 10.1016/j.electstud.2011.08.004.

Ford, Robert, Matthew J. Goodwin, and David C. Cutts (2011). “Anti-Immigrant, Politically Disaffected or Still Racist After All? Examining the Attitudinal Drivers of Extreme Right Support in Britain in the 2009 European Elections”. In:European Journal of Political Research 50.3, pp. 418–440.

Golder, Matt (2003a). “Electoral Institutions, Unemployment and Extreme Right Parties. A Correction”. In: British Journal of Political Science 33, pp. 525–534.

— (2003b). “Explaining Variation in the Success of Extreme Right Parties in Western Europe”.

Hofstadter, Richard (2002a). “Pseudo-Conservatism Revisited: A Postscript [1962]”. In: The Radical Right. Third Edition. With a New Introduction by David Plotke. Ed. by Daniel Bell. New Brunswick: Transaction Publishers, pp. 97–103.

— (2002b). “The Pseudo-Conservative Revolt [1955]”. In: The Radical Right. Third Edition. With a New Introduction by David Plotke. Ed. by Daniel Bell. New Brunswick: Transaction Publishers, pp. 75–95.

Ignazi, Piero (1992). “The Silent Counter-Revolution. Hypotheses on the Emergence of Extreme Right-Wing Parties in Europe”. In: European Journal of Political Research 22, pp. 3–34.

— (2003). Extreme Right Parties in Western Europe. Oxford u.a.: Oxford University Press.

Immerzeel, Tim, Eva Jaspers, and Marcel Lubbers (2013). “Religion as Catalyst or Restraint of Radical Right Voting?” In: West European Politics 36.5, pp. 946–968. DOI: 10.1080/01402382.2013.797235.

Jackman, Robert W. and Karin Volpert (1996). “Conditions Favouring Parties of the Extreme Right in Western Europe”. In: British Journal of Political Science 26, pp. 501–521.

Jackson, Jay W. (1993). “Realistic Group Conflict Theory. A Review and Evaluation of the Theoretical and Empirical Literature”. In: The Psychological Record 43, pp. 395–414.

Jesuit, David K., Piotr R. Paradowski, and Vincent a. Mahler (2009). “Electoral support for extreme right-wing parties: A sub-national analysis of western European elections”. In: Electoral Studies 28.2, pp. 279–290. DOI:10.1016/j.electstud.2009.01.009.

Kestilä, Elina and Peter Söderlund (2007). “Subnational Political Opportunity Structures and the Success of the Radical Right. Evidence from the March 2004 Regional Elections in France”. In: European Journal of Political Research 46, pp. 773–796. DOI: 10.1111/j.1475-_6765. 2007.00715.x.

Kinder, Donald R. and David O. Sears (1981). “Prejudice and Politics. Symbolic Racism vs. Racial Threats to the Good Life”. In: Journal of Personality and Social Psychology 40, pp. 414–431.

Kitschelt, Herbert (1995). The Radical Right in Western Europe. A Comparative Analysis. Ann Arbor: The University of Michigan Press.

— (2013). “Social Class and the Radical Right. Conceptualizing Political Preference Formation and Partisan Choice”. In: Class Politics and the Radical Right. Ed. by Jens Rydgren. London, New York: Routledge, pp. 224–251.

Kitschelt, Herbert and Anthony McGann (2003). “Die Dynamik der schweizerischen Neuen Rechten in komparativer Perspektive. Die Alpenrepubliken”. In: Schweizer Wahlen 1999 / Elections fédérales 1999. Ed. by Pascal Sciarini, Sybille Hardmeyer, and Adrian Vatter. Bern, Stuttgart, Wien: Haupt, pp. 186–216.

Knigge, Pia (1998). “The Ecological Correlates of Right-Wing Extremism in Western Europe”. In: European Journal of Political Research 34, pp. 249–279.

Koopmans, Ruud and Jasper Muis (2009). “The rise of right-wing populist Pim Fortuyn in the Netherlands. A discursive opportunity approach”. In: European Journal of Political Research 48.5, pp. 642–664.

Kornhauser, William (1960). The Politics of Mass Society. London: Routledge and Kegan Paul.

Kuhn, Thomas S. (1962). The Structure of Scientific Revolutions. Chicago: University of Chicago Press.

Lange, Sarah L. de (2007). “A New Winning Formula?: The Programmatic Appeal of the Radical Right”. In: Party Politics 13.4, pp. 411–435. DOI: 10.1177/1354068807075943.

Lubbers, Marcel, Mérove Gijsberts, and Peer Scheepers (2002). “Extreme Right-Wing Voting in Western Europe”. In: European Journal of Political Research 41, pp. 345–378.

Mareš, Miroslav and Vratislav Havlík (2016). “Jobbik’s Successes. An Analysis of its Success in the Comparative Context of the {V4} countries”. In: Communist and Post-Communist Studies, in press. ISSN: 0967-067X. DOI:10.1016/j.postcomstud.2016.08.003.

Mayer, Nonna (2013). “From Jean-Marie to Marine Le Pen: Electoral Change on the Far Right”. In: Parliamentary Affairs 66.1, pp. 160–178. DOI: 10.1093/pa/gss071.

Mayer, Nonna and Pascal Perrineau (1992). “Why Do They Vote for Le Pen?” In: European Journal of Political Research 22, pp. 123–141.

Meguid, Bonnie M. (2005). “Competition Between Unequals: The Role of Mainstream Party Strategy in Niche Party Success”. In: American Political Science Review 99.3, pp. 347–359. DOI: 10.1017/ S0003055405051701.

Merrill, Samuel and Bernard Grofman (1999). A Unified Theory of Voting. Cambridge, New York: Cambridge University Press.

Mudde, Cas (1996). “The War of Words. Defining the Extreme Right Party Family”. In: West European Politics 19, pp. 225–248.

— (1999). “The Single-Issue Party Thesis: Extreme Right Parties and the Immigration Issue”. In: West European Politics 22.3, pp. 182–197.

— (2005). “Racist Extremism in Central and Eastern Europe”. In: East European Politics and Societies 19, pp. 161–184. URL

— (2007). Populist Radical Right Parties in Europe. Cambridge: Cambridge University Press.

Mughan, Anthony and Pamela Paxton (2006). “Anti-immigrant sentiment, policy preferences and populist party voting in Australia”. In: British Journal of Political Science 36, pp. 341–358.

Norris, Pippa (2005). Radical Right. Voters and Parties in the Regulated Market. Cambridge, New York: Cambridge University Press.

Pedahzur, Ami and Avraham Brichta (2002). “The Instituionalization of Extreme Right-Wing Charismatic Parties: A Paradox?” In: Party Politics 8, pp. 31–49.

Perrineau, Pascal (1985). “Le Front National. Un Électorat Autoritaire”. In: Revue Politique et Parlementaire 87.918, pp. 24–31.

Petrocik, John R. (1996). “Issue Ownership in Presidential Elections, with a 1980 Case Study”. In: American Journal of Political Science 40, pp. 825–850.

Pettigrew, Thomas F. and Linda R. Tropp (2008). “How Does Intergroup Contact Reduce Prejudice? Meta-Analytic Tests of Three Mediators”. In: European Journal of Social Psychology 38.6, pp. 922–934. DOI: 10.1002/ ejsp.504.

Prowe, Diethelm (1994). “”Classic” Fascism and the New Radical Right in Western Europe: Comparisons and Contrasts”. In: Contemporary European History 3, pp. 289–313.

Putnam, Robert D. (1993). Making democracy work. Civic traditions in modern Italy. Princeton, NJ: Princeton University Press. 258 pp. ISBN: 0691078890.

Reynolds, Katherine J. and John C. Turner (2001). “The Changing Nature of Racism: From Old to New?” In: Understanding Prejudice, Racism, and Social Conflict. Ed. by Martha Augoustinos and Katherine J. Reynolds. London, Thousand Oaks: Sage, pp. 159–178.

Robinson, William S. (1950). “Ecological Correlation and the Behavior of Individuals”. In: American Sociological Review 15, pp. 351–357.

Rokeach, Milton (1960). The Open and the Closed Mind. Investigations into the Nature of Belief Systems and Personality Systems. New York: Basic Books.

Runciman, Walter G. (1966). Relative Deprivation and Social Justice. A Study of Attitudes to Social Inequality in Twentieth-Century England. Vol. 13. Reports of the Institute of Community Studies. London: Routledge & Kegan Paul.

Rydgren, Jens (2008). “Immigration Sceptics, Xenophobes or Racists? Radical Right-Wing Voting in Six West European Countries”. In: European Journal of Political Research 47.6, pp. 737–765. DOI: 10.1111/ j.1475-_6765.2008.00784.x.

— (2009). “Social Isolation? Social Capital and Radical Right-wing Voting in Western Europe”. In: Journal of Civil Society 5.2, pp. 129–150.

— (2011). “A legacy of “uncivicness”? Social capital and radical right-wing populist voting in Eastern Europe”. In: Acta Politica 46.2, pp. 132–157. DOI: 10.1057/ap.2011.4.

— ed. (2013). Class Politics and the Radical Right. London, New York: Routledge.

Rydgren, Jens and Patrick Ruth (2011). “Contextual Explanations of Radical Right-Wing Support in Sweden: Socioeconomic Marginalization, Group Threat, and the Halo Effect”. In: Ethnic and Racial Studies. DOI:10.1080/01419870.2011.623786.

Scheuch, Erwin K. and Hans-Dieter Klingemann (1967). “Theorie des Rechtsradikalismus in westlichen Industriegesellschaften”. In: Hamburger Jahrbuch für Wirtschafts- und Sozialpolitik 12, pp. 11–29.

Scheufele, Dietram A. and David Tewksbury (2007). “Framing, Agenda Setting, and Priming: The Evolution of Three Media Effects Models”. In: DOI: 10.1111/j.1460-_2466.2006.00326.x.

Sheets, Penelope, Linda Bos, and Hajo G. Boomgaarden (2015). “Media Cues and Citizen Support for Right-Wing Populist Parties”. In: International Journal of Public Opinion Research. DOI: 10.1093/ijpor/ edv014.

Smith, Jason Matthew (2010). “Does Crime Pay? Issue Ownership, Political Opportunity, and the Populist Right in Western Europe”. In: Comparative Political Studies 43.11, pp. 1471–1498. DOI: 10.1177/ 0010414010372593.

Stojarová, Věra (2012). “The extreme right in Croatia, Bosnia-Herzegovina and Serbia”. In: Mapping the Extreme Right in Contemporary Europe. From Local to Transnational. Ed. by Andrea Mammone, Emmanuel Godin, and Brian Jenkins. London et al.: Routledge, pp. 143–158.

Sumner, William Graham (1906). Folkways. A Study of the Sociological Importance of Usages, Manners, Customs, Mores, and Morals. Boston: Ginn.

Swank, Duane and Hans-Georg Betz (2003). “Globalization, the Welfare State and Right-Wing Populism in Western Europe”. In: Socio-Economic Review 1, pp. 215–245.

Swyngedouw, Marc (2001). “The Subjective Cognitive and Affective Map of Extreme Right Voters: Using Open-ended Questions in Exit Polls”. In: Electoral Studies 20, pp. 217–241.

Taggart, Paul (1995). “New Populist Parties in Europe”. In: West European Politics 18.1, pp. 34–51.

Tajfel, Henri and John C. Turner (1986). “The Social Identity Theory of Intergroup Behaviour”. In: Psychology of Intergroup Relations. Ed. by Stephen Worchel and William G. Austin. Chicago: Nelson-Hall Publishers, pp. 7–24.

Taylor, C. Marylee (2002). “Fraternal Deprivation, Collective Threat, and Racial Ressentment”. In: Relative Deprivation. Specification, Development, and Integration. Ed. by Iain Walker and Heather J. Smith. Cambridge: Cambridge University Press, pp. 13–43.

Vasilopoulou, Sofia (2011). “European Ingegration and the Radical Right. Three Patterns of Opposition”. In: Government and Opposition 46.2, pp. 223–244.

Walker, Iain (2001). “The Changing Nature of Racism: From Old to New?” In: Understanding Prejudice, Racism, and Social Conflict. Ed. by Martha Augoustinos and Katherine J. Reynolds. London, Thousand Oaks: Sage, pp. 24–42.

Werts, Han, Peer Scheepers, and Marcel Lubbers (2013). “Euro-scepticism and radical right-wing voting in Europe, 2002 2008: Social cleavages, socio-political attitudes and contextual characteristics determining voting for the radical right”. In:European Union Politics 14.2, pp. 183–205. DOI: 10.1177/1465116512469287.

Zaller, John R. (1992). The Nature and Origin of Mass Opinion. Cambridge, New York, Oakleigh: Cambridge University Press.

Zúquete, José Pedro (2008). “The European Extreme-Right and Islam. New Directions?” In: Journal of Political Ideologies 13.3, pp. 321–344. DOI: 10.1080/13569310802377019.

1At least at the attitudinal level, old and modern racism seem to be closely related (Walker, 2001).

2Although value orientations are sometimes grouped together with personality traits, they will be discussed in a separate section below.

3Similar arguments have been made about the rise of the right-wing extremist movements in the 1920s as well as about their resurgence in the postwar years (e.g. Scheuch and Klingemann, 1967).

4Marine Le Pen’s attempts to soften the image of the Front National (Mayer, 2013) and her public clashes with her father over his unreformed anti-semitism are a case in point.

Wahlforschung in der Vergleichenden Politikwissenschaft


1 Einleitung

Die Wahlforschung ist eines der wichtigsten Teilgebiete der Politischen Soziologie. Sie operiert damit an der Schnittstelle zwischen Politikwissenschaft und Soziologie. Zunächst war die Wahlforschung ganz auf die Erklärung nationaler Phänomene ausgerichtet. Verglichen wurden hier lediglich die Verhältnisse innerhalb eines politischen Systems, etwa in Frankreich (Siegfried, 1913) oder den USA (Key, 1959).

Seit etwa Ende der 1960er Jahre hat jedoch die international vergleichende Perspektive in der Wahlforschung stetig an Bedeutung gewonnen. Ausgangspunkt für diese Entwicklung war das Interesse der Vertreter des sozialpsychologischen Modells (Abschnitt 2.2), ihre Befunde in einem „most dissimilar“ Design zu validieren (Miller, 1994, S. 256). Umgekehrt zog die an der University of Michigan/Ann Arbor beheimatete Forschergruppe Kollegen aus der ganzen Welt, insbesondere aber aus Nord-West-Europa an. Auf diese Weise entstanden Kooperationsbeziehungen zwischen den Leitern verschiedener nationaler Wahlstudien, die teils über Jahrzehnte Bestand hatten (Miller, 1994, S. 256-259) und den Grundstein für die Institutionalisierung der vergleichenden Wahlforschung seit den 1970er Jahren legten (Mochmann, 2002).

Der vorliegende Beitrag gliedert sich in zwei Teile. Abschnitt 2 gibt zunächst einen knappen Überblick über die wichtigsten Ansätze der allgemeinen Wahlforschung. Abschnitt 3 stellt dann die wichtigsten Forschungsfelder, Datenquellen und Methoden der vergleichenden Wahlforschung vor.

  • 2 Theorien des Wählerverhaltens

Die Anfänge der Wahlforschung liegen in der offiziellen Statistik des 19. Jahrhunderts und im Werk André Siegfrieds, der zu Beginn des 20. Jahrhunderts damit begann, Wahlergebnisse mit kartographische Methoden darzustellen und Zusammenhänge etwa zwischen der Siedlungsstruktur und dem Abschneiden bestimmter Parteien zu untersuchen. Den Kern der modernen Wahlforschung bilden aber drei Theoriebündel, die Mitte des 20. Jahrhunderts in den USA begründet wurden und schlagwortartig als soziologischer, sozialpsychologischer und ökonomischer (oder rationalistischer) Ansatz bezeichnet werden.1 In der Forschungspraxis werden häufig in sehr pragmatischer Weise Elemente aus allen drei Ansätzen kombiniert. Wie eine explizite Verbindung der Theorien aussehen könnte, diskutieren Rudi und Schoen (2005).

  • 2.1 Soziologische Ansätze

Im Bereich der klassischen Wahlsoziologie lassen sich grob zwei Strömungen unterscheiden. Der sogenannte mikrosoziologische Ansatz geht auf Studien zurück, die seit den 1940er Jahren von Paul Lazarsfeld und seinen Kollegen an der Columbia University durchgeführt wurden (Berelson, Lazarsfeld und McPhee, 1954; Lazarsfeld, Berelson und Gaudet, 1944). Die Columbia-Gruppe nahm ursprünglich an, daß sich Wähler während des Wahlkampfes umfassend informieren, um dann wohlüberlegt eine Wahlentscheidung zu treffen. Um diese Hypothese zu prüfen, untersuchten Lazarsfeld et al. in sehr aufwendigen Studien die Inhalte regionaler Medien und versuchten diese mit individuellen Meinungsbildungsprozessen in Verbindung zu setzen, die sie mit Hilfe wiederholter Befragungen erfaßten.

Dabei zeigte sich jedoch rasch, daß die meisten Menschen Informationen über politische Inhalte nur indirekt über sogenannte Meinungsführer wahrnahmen. In vielen Fällen stand die Wahlentscheidung bereits zu Beginn des Wahlkampfes weitgehend fest und ließ sich durch Kenntnis einiger weniger sozio-demographischer Merkmale wie des Berufs, der ethnischen Gruppe oder der Religionszugehörigkeit recht gut voraussagen: „A person thinks, politically, as he is, socially“ (Lazarsfeld, Berelson und Gaudet, 1944, S. 27). Lazarsfeld et al. erklären diesen Befund mit der Dynamik kleiner Gruppen und dem Wunsch des Individuums, sich normkoform zu verhalten, gehen aber nicht auf die gesellschaftlichen Voraussetzungen für die Entstehung solcher Muster ein.

Diesen fehlenden Baustein liefert der makrosoziologische Ansatz, dessen Wurzel in den Arbeiten der Soziologen Stein Rokkan und Martin Semour Lipset zur Entstehung der westeuropäischen Parteiensysteme liegt (Lipset und Rokkan, 1967). Lipset und Rokkan führen diese auf eine Reihe sozio-politischer Großkonflikte (cleavages) zurück,2 in deren Verlauf es zu einer dauerhaften Verbindung zwischen bestimmten sozialen Gruppen und Parteien (z. B. Arbeiter ¡ê sozialistische/sozialdemokratische Parteien) gekommen sei. Die für ein Land charakteristische Konfiguration dieser Konflikte entscheidet aus Sicht von Lipset und Rokkan darüber, welche und wieviele Parteien existieren.

Mikro- und makrosoziologischer Ansatz zeichnen gemeinsam ein plausibles Bild davon, wie soziale und historische Faktoren das Wahlverhalten beeinflussen können. Damit sind sie insbesondere für die (international) vergleichende Wahlforschung bis heute von Bedeutung, weil die Wirkung dieser Faktoren naturgemäß nur in vergleichender Perspektive sichtbar werden kann. Außerhalb von Phasen revolutionärer Umbrüche tun sich die soziologischen Ansätze jedoch schwer damit, Veränderungen im Wählerverhalten zu erklären. Dies erklärt den Erfolg des sozialpsychologischen Ansatzes, der im nächsten Abschnitt vorgestellt werden.

  • 2.2 Der Sozialpsychologische Ansatz

In den späten 1940er Jahren begann sich am Survey Research Center der University of Michigan/Ann Arbor eine Arbeitsgruppe um den Sozialpsychologen Angus Campbell mit dem Wahlverhalten der Amerikaner zu beschäftigen. Dabei griffen sie auf die Konzepte und Methoden der repräsentativen Umfrageforschung zurück. Im Mittelpunkt des neuen Ansatzes standen drei Einstellungen (Dispositionen gegenüber politischen Objekten): Kandidaten- und Sachfragenorientierungen sowie die Parteiidentifikation, ein dauerhaftes, über die konkrete Wahlentscheidung hinausweisendes Gefühl der Verbundenheit mit einer der beiden großen amerikanischen Parteien.

Da diese Einstellungen der eigentlichen Wahlentscheidung unmittelbar vorgelagert sind, wurde eine erste Studie (Campbell, Gurin und Miller, 1954) als tautologisch kritisiert. Die Ann Arbor-Gruppe reagierte auf diese Kritik, indem sie in der Folgestudie „The American Voter“ (Campbell u. a., 1960) einen weitgespannten theoretischen Analyserahmen entwickelte, der historische, soziale, ökonomische und institutionelle Rahmenbedingungen als vorgelagerte Variablen mit einbezieht. Zugleich revidierten die Autoren ihre Sicht auf das Verhältnis der Einstellungen untereinander: Die Parteiidentifikation gilt nun als wichtigster Bestandteil der Variablentrias, die in der Lage ist, die Wahrnehmung von politischen Themen und Kandidaten zu beeinflussen.

Diese neue theoretische Konzeption dominierte die akademischen Debatte für mehr als eine Dekade. In der Folge wurde „The American Voter“ zu einer der bis heute am häufigsten zitierten Monographien in der Geschichte der Wahlforschung.

Aus dem „American Voter“ und weiteren Umfrageprojekten der Ann Arbor-Gruppe ging schließlich die US-amerikanische National Election Study hervor, die seit 1948 jede nationale Wahl in den USA untersucht und damit eines der größten sozialwissenschaftlichen Forschungsprojekte überhaupt darstellt. Im Laufe der Zeit wurde die National Election Study so zum Vorbild für nationale Wahlstudien auf der ganzen Welt. Auch der indirekte Einfluß der Ann Arbor-Gruppe auf die Wahlforschung ist somit enorm.

Im Laufe der Zeit wurde der ursprüngliche Ansatz immer wieder ergänzt, modifiziert und erweitert (Miller, 1994), teils sogar durch die ursprünglichen Autoren (Miller und Shanks, 1996). Eine der interessantesten Entwicklungen besteht dabei darin, daß in neuerer Zeit der sozialpsychologische Aspekt, der in der Praxis in den Hintergrund getreten war, betont und zugleich der Anschluß an die moderne Kognitionspsychologie gesucht wird (Weisberg und Greene, 2003).

Dennoch wurde Theorie und Forschungspraxis der sozialpsychologisch orientierten Wahlforschung immer wieder als dogmatisch und wenig innovativ kritisiert (siehe z. B. Achen, 1992). Aus dieser intellektuellen Unzufriedenheit heraus speist sich ein dritter Theoriestrang, der im nächsten Abschnitt vorgestellt werden soll.

  • 2.3 Der Ökonomische Ansatz

Ausgangspunkt des „ökonomischen“ oder „rationalistischen“ Zugangs zur Wahlforschung ist die „Economic Theory of Democracy“, mit der Anthony Downs (1957) zu einem Wegbereiter des Rational Choice Ansatzes in der Politikwissenschaft wurde. Obwohl Downs mit den Methoden und Ergebnissen der zeitgenössischen Wahlforschung vertraut war, ging es ihm nicht darum, selbst eine empirische Studie durchzuführen oder eine realistische Theorie des Wahlverhaltens zu entwickeln. Vielmehr konstruiert Downs eine Modellwelt, in der sich aus einigen wenigen axiomatischen Annahmen, die er aus der Mikroökonomie übernimmt, interessante Ergebnisse ableiten lassen.

Downs unterscheidet dabei zwischen zwei Klassen von Akteuren: Wählern, die ihr monetäres Einkommen aus der Regierungstätigkeit maximieren wollen, und Parteien, die möglichst viele politische Ämter besetzen möchten. Sowohl Wähler als auch Parteien sind dabei an die Regeln einer Verfassung gebunden, die freie, faire und regelmäßige Wahlen vorsieht.

In Anlehnung an das Vorgehen in der Ökonomie geht Downs zunächst davon aus, daß die Akteure über vollständige Präferenzen und Informationen verfügen. Diese zweite Annahme gibt Downs dann schrittweise auf um so zu zeigen, daß der Rückgriff auf Ideologien und ähnliche Konstrukte eine durchaus rationale Strategie sein kann, wenn die Kosten für die Beschaffung zusätzlicher politischer Informationen deren erwarteten Nutzen deutlich überschreiten.

Zu Downs’ bekanntesten Ergebnissen gehört neben dem auf Hotelling (1929) zurückgehenden Medianwähler-Theorem ¨C in einem Zweiparteiensystem mit einer einzelnen Policy-Dimension werden die Programme rationaler Parteien an der Position des Wählers konvergieren, der die ideologische Mitte des Elektorats repräsentiert ¨C das Wahlparadoxon, das sich aus der rein instrumentellen Motivation der Wähler ergibt.

Da Wähler sich nach den Modellannahmen ausschließlich für ihr Einkommen aus der Regierungstätigkeit interessieren, ergibt sich der Nutzen der Wahlteilnahme aus der Differenz zwischen dem Einkommen, das sie unter der von ihnen bevorzugten Partei erzielen, und dem Einkommen, das ihnen zufließt, wenn statt dessen die zweitplazierte Partei die Regierung übernimmt. Anders als bei einer Kaufentscheidung kann der einzelne Wähler aber nicht eigenständig darüber entscheiden, welche Partei die Wahl gewinnen soll. Vielmehr muß der potentielle Nutzen der Wahlteilnahme mit der Wahrscheinlichkeit gewichtet werden, daß der Wähler selbst die entscheidende Stimme abgibt, die der bevorzugten Partei zum Sieg verhilft. Diese Wahrscheinlichkeit ist unter den Bedingungen einer Massendemokratie verschwindend gering, so daß die Kosten der Wahlbeteiligung (vor allem die aufgewendete Zeit) deren erwarteten Nutzen stets übersteigen.3 Rationale Wähler sollten sich deshalb nicht an Wahlen beteiligen. Dennoch liegt die Wahlbeteiligung bei nationalen Wahlen in Demokratien meist deutlich höher als 50 Prozent.

Seit Erscheinen der „Economic Theory“ haben sich viele hervorragende Theoretiker darum bemüht, das Wahlparadoxon aufzulösen. Stärker empirisch orientierte Forscher hingegen sehen in den realen Wahlbeteiligungsraten „the paradox that ate Rational Choice Theory“ (Grofman, 1993). Dennoch konnte sich ca. seit den 1970er Jahren eine Strömung der empirischen Wahlforschung entwickeln, die sich explizit in die Tradition von Downs stellt. Dabei lassen sich vier Felder unterscheiden, auf denen besonders intensiv geforscht wird:


Die Re-Interpretation der Parteiidentifikation als Summe der (ökonomischen) Erfahrungen („running tally“), die ein Wähler im Laufe seines politischen Lebens mit den Parteien gemacht hat (Fiorina, 2002).


Die Modellierung von mehrdimensionalen issue- bzw. policy-Räumen, innerhalb derer Wähler Präferenzen entwickeln und Parteien programmatische Angebote machen (siehe als Überblick Pappi, 2000).


Die Bedeutung der Wirtschaftslage für die Erfolgsaussichten von Regierung und Oppostion (Lewis-Beck und Paldam, 2000).


Die Analyse von Anreizen zum taktischen Wählen,4 die der Kontext und insbesondere das Wahlsystem auf rationale Wähler ausüben (Cox, 1997).

Insbesondere die letzten drei Felder sind für die vergleichende Wahlforschung von großer Bedeutung.

  • 3 Vergleichende Wahlforschung

  • 3.1 Forschungsfelder

  • 3.1.1 Kontextvariablen

In der Forschungspraxis existiert faktisch keine Trennung zwischen (international) vergleichender und nationaler (oder subnationaler) Wahlforschung. Nur wenige Forscherinnen und Forscher, die in diesem Bereich aktiv sind, sehen sich ausschließlich als Länderspezialisten oder Komparativisten. Dennoch gibt es einige typische Forschungsfelder, die in der national orientierten Forschung keine oder eine geringere Rolle spielen. Dies erklärt sich daraus, daß einige für die Wahlforschung interessante Variablen innerhalb eines politischen Systems über längere Zeiträume völlig oder fast stabil sind.

An erster Stelle ist hier das Wahlsystem zu nennen, das in etablierten Demokratien nur höchst selten verändert wird, da eine Veränderung in der Regel nicht im Interesse der Parteien liegt, die darüber im Parlament zu entscheiden haben. Kommt es tatsächlich zu einem Wechsel des Wahlsystems wie etwa 1996 in Neuseeland, so stellt sich außerdem die Frage, inwieweit das Wählerverhalten vor und nach der Wahl überhaupt miteinander vergleichbar ist, und ob ein Wechsel des Wahlsystems möglicherweise eine Folge langfristiger Veränderungen im Wahlverhalten als deren Ursache ist. Deshalb bietet es sich an, die Wirkung von Wahlsystemen im internationalen Vergleich zu untersuchen. Im Mittelpunkt steht dabei häufig die Frage, ob, wie von Duverger (1951) in seinem berühmten „Gesetz“ behauptet, das Wahlsystem einen entscheidenden Einfluß auf das Format des Parteiensystems hat.

Dabei ist allerdings zu beachten, daß Wahlsysteme innerhalb einer Region oft kaum variieren. So zeigt ein Blick in die Datenbank der Internationalen Parlamentarischen Union, daß lediglich acht von 64 europäischen Staaten ein Mehrheitswahlsystem verwenden. Bei den karibischen Staaten hingegen sind es 14 von 22 Staaten, in denen nach diesem System gewählt wird (http://www.ipu.org/).

Neben dem Wahl- und Parteiensystem wurden und werden in der international vergleichenden Wahlforschung eine Vielzahl weiterer Kontextvariablen untersucht. Insbesondere im Bereich der Rechtsextremismusforschung wird etwa der Effekt von Sozialausgaben, Zuwanderungs- und Arbeitslosenquoten auf die Wahlabsicht zugunsten der Extremen Rechten untersucht (Lubbers, Gijsberts und Scheepers, 2002; Swank und Betz, 2003). In ähnlicher Weise kann auch das Verhalten anderer Parteien (Arzheimer, 2009) oder institutioneller Faktoren (Arzheimer und Carter, 2006) mit in die Modelle aufgenommen werden. Dabei zeigt sich in der Regel, daß diese (nationalen) Kontextvariablen durchaus erklärungskräftig sind, ohne daß individuelle Merkmale und solche Unterschiede zwischen den Ländern, die nicht durch Variablen abgebildet werden können, an Bedeutung verlieren würden.

  • 3.1.2 Class Voting und die Bedeutung der Religion

In Abschnitt 2.1 wurde die auf Lipset und Rokkan zurückgehende Theorie der sozialen Spaltungen (cleavages) vorgestellt. Zwei dieser Cleavages ¨C Arbeit vs. Kapital und Staat vs. (katholische) Kirche ¨C sind dafür verantwortlich, daß sozialdemokratische und christdemokratische Parteien bis in die 1980er Jahre hinein das politische Leben in vielen westeuropäischen Gesellschaften bestimmen konnten.

Zur sozialen Basis beider Parteifamilien liegt eine kaum überschaubare Literatur vor die zeigt, daß sich in der Tendenz der Effekt der Klassenzugehörigkeit deutlich abgeschwächt hat. Neuer Studien belegen aber, daß sich diese Entwicklung in verschiedenen Ländern durchaus unterschiedlich darstellt (Nieuwbeerta und Graaf, 2001). Auch von einem universellen Bedeutungsverlust der Religionszugehörigkeit für das Wahlverhalten kann keine Rede sein (Broughton und Napel, 2000). Im Ergebnis bedeutet dies: „Reports of the death of social cleavages are exaggerated“ (Elff, 2007). Gleichwohl ist festzuhalten, daß in den meisten demokratischen Gesellschaften sowohl die Zahl religiöser Menschen als auch die Zahl derjenigen, die Arbeiterberufe ausüben und/oder sich selbst als Arbeiter verstehen, langsam aber stetig sinkt.

  • 3.1.3 Nichtwahl

Abbildung 1: Wahlbeteiligung in den EU-15 Staaten (Quelle: IDEA)

In den letzten drei Dekaden ist die Wahlbeteiligung in Westeuropa (vgl. Abbildung 1), aber auch in anderen Weltregionen deutlich erkennbar gesunken. Da es sich hier klar um einen länderübergreifenden Trend handelt, ist das Phänomen der Nichtwahl zu einem wichtigen Gegenstand der vergleichenden Wahlforschung geworden. Im Zentrum des Interesses stehen dabei drei Variablenkomplexe:


Unterschiede in der Zusammensetzung der Elektorate und dabei besonders der Anteil der Jung- und Erstwähler


Politisch-kulturelle Unterschiede zwischen den Ländern


Institutionelle Unterschiede

Im Ergebnis zeigt sich, daß institutionelle Faktoren, die aus einer Rational Choice-Perspektive die Kosten der Wahlbeteiligung beeinflussen, einen erheblichen Teil der Varianz zwischen den Ländern erklären können. Besonders starke Effekte haben ¨C wenig überraschend ¨C das Bestehen einer Wahlpflicht sowie die automatische Registrierung von Wählern, Möglichkeiten zur Briefwahl sowie das Abhalten von Wahlen an arbeitsfreien Tagen (Franklin, Eijk und Oppenhuis, 1995). Ein hoher Anteil von Jung- und Erstwählern reduziert ceteris paribus die Wahlbeteiligung, da die Teilnahme an Wahlen für viele Bürger eine Gewohnheit darstellt, die sich im Lebensverlauf stabilisiert (Plutzer, 2002).

Beide Faktoren können aber das Absinken der Wahlbeteiligung nicht erklären, da in fast allen Gesellschaften das Durchschnittsalter der Wähler durch den demographischen Wandel steigt und die institutionellen Hürden für die Wahlteilnahme in vielen Ländern gesenkt wurden. Der Rückgang der Wahlbeteiligung muß deshalb primär auf politisch-kulturelle Wandlungsprozesse, d. h. auf das Verblassen von Wahlnormen und den Bedeutungszuwachs alternativer Beteiligungsformen (Norris, 1999) zurückzuführen sein, die in verschiedenen Ländern unterschiedlich weit fortgeschritten sind. Auch die Abschwächung und allmähliche Auflösung von Parteibindungen (dealignment, siehe Dalton und Wattenberg 2000), die mit dem oben beschriebenen Bedeutungsverlust der traditionellen cleavages einhergeht, gilt als wichtiger Faktor für das Sinken der Wahlbeteiligung.

  • 3.1.4 Economic Voting

Economic Voting“ ist ein breites und dynamisches Forschungsfeld, dessen Grundannahmen auf Downs’ ökonomische Theorie der Demokratie zurückgehen.5 Aus der „Economic Voting“ Perspektive machen die Bürger die jeweilige Regierung für die Wirtschaftslage eines Landes verantwortlich. Wenn sich wichtige makroökonomische Kenngrößen wie die Inflationsrate, die Arbeitslosenquote und das Bruttosozialprodukt verschlechtern bzw. nicht signifikant verbessern, bestrafen die Bürger die Regierungsparteien, indem sie ihnen in Abstimmungen und Umfragen ihre Unterstützung entziehen. Dieser Zusammenhang wird als „VP-Function“ bezeichnet (Nannestad und Paldam, 1994).

Aus einer Vielzahl von Studien, die seit den 1970er Jahren durchgeführt wurde, läßt sich ein Kern von weiteren Befunden extrahieren, die weithin akzeptiert sind (Lewis-Beck und Paldam, 2000, S. 114): Wähler haben einen kurzen Zeithorizont, sie orientieren sich stärker an der (unmittelbaren) Vergangenheit als an ihren Erwartungen für die Zukunft, die nationale Wirtschaftslage ist wichtiger als die persönlichen Finanzen und negative Entwicklungen werden von den Wählern stärker gewichtet als positive Veränderungen.

Zu den Besonderheiten der „Economic Voting“ Forschung gehört, daß die beschriebenen Effekte bei vielen Wahlen recht stark ausfallen, während sie sich in anderen Fällen nicht oder nur in geringem Umfang reproduzieren lassen (Lewis-Beck und Paldam, 2000, S. 113¨C114). Eine mögliche Erklärung dafür liegt in den Kontextvariablen und hier insbesondere in den institutionellen Unterschieden zwischen den Systemen. Während aus Sicht der Wähler in Mehrheitsdemokratien die Verantwortlichkeit der Regierung für die wirtschaftliche Entwicklung relativ definiert ist, kommt es in Konsensusdemokratien durch die Zwänge, die sich aus der Bildung von Koalitionsregierungen ergeben, aber auch durch die Intervention von unabhängigen Zentralbanken, zweiten Kammern oder starken Verfassungsgerichten zu einer Verantwortungsdiffusion. Dadurch schwächt sich der Zusammenhang zwischen Wirtschaftslage und Popularität der Regierung bzw. der größten Regierungspartei erkennbar und systematisch ab (Anderson, 2000; Nadeau, Niemi und Yoshinaka, 2002).

Besonders interessant sind vor diesem Hintergrund politische Systeme, in denen die „clarity of responsibility“ deutlich und in nachvollziehbarer Weise schwankt. Dies gilt neben der Bundesrepublik etwa für Frankreich, wo sich in Zeiten der cohabitation aus Sicht der Wähler die Verantwortung für die Wirtschaft vom Präsidenten zum Premierminister verschiebt (Lewis-Beck, 1997).

  • 3.2 Studien, Datenquellen

Die administrative und finanzielle Aufwand für die international vergleichbare Erhebung von Wählerdaten ist prohibitiv hoch. Wie oben bereits angedeutet, kam es jedoch schon früh zu einer Zusammenarbeit zwischen jenen Forschern, die für verschiedene nationale Wahlstudien verantwortlich waren. 1989 wurde diese Praxis durch die Gründung des „International Committee for Research into Elections and Representative Democracy“ (ICORE) formalisiert (Karvonen und Ryssevik, 2001, S. 44). Die führenden europäischen Datenarchive hatten sich bereits in den 1970er Jahren zum „Council of European Social Science Data Archives“ (CESSDA, http://www.cessda.org) zusammengeschlossen (Karvonen und Ryssevik, 2001, S. 45).

Neben einem Verzeichnis der von 1945-1995 durchgeführten nationalen Wahlstudien ist aus der Kooperation von ICORE, CESSDA und der University of Michigan/Ann Arbor die Comparative Study of Electoral Systems (CSES) hervorgegangen (Lagos, 2008, S. 589-590). Im Rahmen der CSES wird von den beteiligten Institutionen im Anschluß an die Interviews für die jeweilige nationale Studie eine weitere Batterie von einheitlichen Fragen gestellt. Die so erhobenen Individualdaten werden mit Meso- und Makro-Informationen zusammengespielt und stehen Wissenschaftlern auf der ganzen Welt online frei zur Verfügung (http://www.cses.org). Inzwischen sind für den Zeitraum von 1996 bis 2011 Daten aus mehr als 50 Ländern verfügbar. Damit ist die CSES für die vergleichende Wahlforschung eine Ressource von unschätzbarem Wert.

Ähnliche, aber spezifischere Ziele verfolgt die European Election Study (EES, http://eeshomepage.net), die seit 1979 die direkten Wahlen zum Europäischen Parlament begleitet. Aus vergleichender Perspektive sind diese Europawahlen von besonderem Interesse, weil hier in den Mitgliedsländern zum selben Zeitpunkt Kandidaten für dieselbe Institution gewählt werden. Neben den Interviews mit den Wählern, die für alle Wahlen vorliegen, wurden zu einzelnen Urnengängen zusätzliche Kandidatenbefragungen sowie Inhaltsanalysen der Parteiprogramme und der Medienberichterstattung durchgeführt. Hinzu kommen weitere Kontextdaten. Auch die Datensätze aus dem EES-Projekt stehen interessierten Wissenschaftlern über das Internet zur Verfügung.

Neben diesen reinen Wahlstudien existiert inzwischen eine große Zahl weiterer internationaler Surveys, die auch für die Zwecke der Wahlforschung genutzt werden können. Zu den wichtigsten dieser Studien zählen das Eurobarometer, der European Social Survey und der World Values Survey.

  • 3.3 Methoden

Dank der „Technological Revolution“ (Karvonen und Ryssevik, 2001) in der Erhebung und Verbreitung von Umfragedaten (die nicht zuletzt auch eine forschungspolitische und -kulturelle Revolution darstellt) verfügt die vergleichende Wahlforschung heute über Möglichkeiten, die in den 1990er Jahren noch als utopisch galten.

Eine ähnliche Revolution hat sich auch auf dem Gebiet der Analysetechniken und der Computerhardware vollzogen. Die Datensätze, die in der vergleichenden Wahlforschung verwendet werden, sind für sozialwissenschaftliche Verhältnisse sehr groß. So umfaßt beispielsweise die (partielle) Kumulation der Eurobarometer-Daten (Schmitt u. a., 2009) mehrere 100 000 Fälle, was einigen hundert Megabyte entspricht. Während frühere Versionen dieser Kumulation die zum Zeitpunkt ihrer Erstellung verfügbaren PCs an den Rande ihrer Leistungsfähigkeit brachten, lassen sich die heutigen, weitaus umfangreicheren Datensätze problemlos mit Geräten aus den Regalen der Discounter bearbeiten. Parallel dazu ist das technische Niveau der Analysen kontinuierlich gestiegen, weil Auswertungsverfahren, die früher eigene Programmierkenntnisse erforderten, in Summer Schools und Doktorandenprogrammen gelehrt werden und in Standardsoftware wie SPSS oder Stata implementiert sind.

Im Vergleich zur national orientierten Wahlforschung ergeben sich Besonderheiten zum einen aus der Natur der abhängigen Variable „Wahlverhalten“. Da die Wähler in den verschiedenen Ländern vor je unterschiedlichen Alternativen stehen ¨C selbst bei den Europawahlen kandidieren bisher nationale Listen ¨C müssen die nationalen Parteien bzw. Kandidaten einer Parteifamilie zugeordnet werden. Dies wirft einerseits die Frage auf, ob etwa eine Entscheidung zugunsten der deutschen SPD tatsächlich völlig äquivalent zu einer Stimme für die griechische PASOK ist. Andererseits ist unklar wie damit umzugehen ist, wenn sich die choice sets sehr stark unterscheiden, weil eine bestimmte Parteifamilie in einem Land nicht existiert oder faktisch keine politische Bedeutung hat. Forscher aus dem Umfeld der EES haben als Lösung dieses Problems lange Zeit eine spezielle Form der linearen Regression propagiert (Eijk und Kroh, 2002), die eine aufwendige Variante der Wahlabsichtsfrage erfordert, ohne daß klar ist, ob sich die Probleme damit wirklich lösen lassen. Die Mehrzahl der Arbeiten verwendet aber weiterhin eines der gängigen Verfahren zur Analyse polytomer Daten (in der Regel die multinomiale logistische Regression) oder dichotomisiert die abhängige Variable, indem das Stimmverhalten zugunsten einer bestimmten Parteienfamilie mit allen anderen Entscheidungen kontrastiert wird.

Eine zweite Besonderheit besteht darin, daß die moderne vergleichende Wahlforschung danach strebt, die Effekte von Variablen auf der Mikro- (Personen), Meso- (z. B. Kandidaten, Medieninhalte) und Makro-Ebene (Institutionen, nationale Wirtschaftslage etc.) gemeinsam zu modellieren. Als elegantes Verfahren dafür hat sich in den letzten Jahren die statistische Mehrebenenanalyse etabliert, die gegenüber älteren Methoden eine Vielzahl von Vorteilen bringt (Steenbergen und Jones, 2002). Allerdings wird in der Literatur kontrovers darüber diskutiert, wieviele Länder für eine Mehrebenenanalyse benötigt werden und ob sich das Verfahren überhaupt anwenden läßt, wenn diese Ländern nicht zufällig aus einer großen Population ausgewählt werden (Stegmueller, 2013).

Unabhängig von diesen Details der statistischen Modellierung bringt die Verwendung von Makro-Daten, die sich in einem Land nur sehr langsam oder gar nicht verändern, Probleme mit sich, die in dieser Form bei nationalen Wahlen selten auftreten: Wenn beispielsweise das Wahlsystem innerhalb eines Landes konstant ist, läßt sich sein Effekt auf die Wahl kleiner Parteien nicht schätzen, obwohl dieser möglicherweise sehr stark ausgeprägt ist. Eine Effektschätzung muß dann ausschließlich auf Unterschieden zwischen den Ländern basieren. Dieses Vorgehen ist aber mit neuen Problemen verbunden, weil die Zahl der Fälle in Relation zu den interessanten Variablen klein ist, diese Variablen auf der Makro-Ebene in der Regel sehr eng miteinander korreliert sind6 und überdies starke idiosynkratische Effekte einzelner Länder auftreten (unit effects). In diesem Sinne sind viele Makro-Datensätze „schwach“ (Western und Jackman, 1994) ein in der makro-quantitativen Forschung bekanntes grundsätzliches Problem, das auch durch die Verwendung moderner Analyseverfahren und die gleichzeitige Berücksichtigung von Mikro-Daten nicht zu lösen ist.

  • 4 Zusammenfassung

Bereits seit den 1970er Jahren haben sich durch den Vergleich von Ergebnissen aus nationalen Wahlstudien die Perspektiven der Wahlforschung erheblich erweitert. Die Anfänge der EES in den späten 1970er und dann der CSES in den 1990er Jahren markieren den Übergang zu einer Forschung, die von vornherein auf eine Äquivalenz der Konzepte und Instrumente ausgerichtet und damit genuin vergleichend angelegt ist. Die fast flächendeckende Freigabe von Datensätzen für Sekundärforscher über das Internet, die Fortschritte in der Computertechnik und die Verbreitung moderner statistischer Methoden haben in den letzten zwei Dekaden einen weiteren rasanten Fortschritt ermöglicht. Trotz der obengenannten Probleme zählt die international vergleichende Wahlforschung heute sowohl in der Wahlforschung als auch in der Vergleichenden Politikwissenschaft zu den am weitesten entwickelten Subdisziplinen.

  • Literatur

Achen, Christopher H. (1992). „Social Psychology, Demographic Variables, and Linear Regression: Breaking the Iron Triangle in Voting Research.“ In: Political Behavior 14.3, S. 195¨C211.

Anderson, Christopher J. (2000). „Economic Voting and Political Context: A Comparative Perspective“. In: Electoral Studies 19.2-3, S. 151¨C170.

Arzheimer, Kai (2009). „Contextual Factors and the Extreme Right Vote in Western Europe, 1980¨C2002“. In: American Journal of Political Science 53.2, S. 259¨C275.

Arzheimer, Kai und Elisabeth Carter (2006). „Political Opportunity Structures and Right-Wing Extremist Party Success“. In: European Journal of Political Research 45, S. 419¨C443. doi: 10.1111/j.1475-6765.2006. 00304.x.

Berelson, Bernard, Paul F. Lazarsfeld und William N. McPhee (1954). Voting. A Study of Opinion Formation in a Presidential Campaign. Chicago: Chicago University Press.

Broughton, David und Hans-Martien Napel, Hrsg. (2000). Religion and Mass Electoral Behaviour in Europe. London, New York: Routledge.

Bürklin, Wilhelm und Markus Klein (1998). Wahlen und Wählerverhalten. Eine Einführung. 2. Aufl. Bd. 3. Grundwissen Politik. Opladen: Leske und Budrich.

Campbell, Angus, Gerald Gurin und Warren E. Miller (1954). The Voter Decides. Evanston: Harper und Row.

Campbell, Angus u. a. (1960). The American Voter. New York: John Wiley.

Cox, Gary W. (1997). Making Votes Count. Strategic Coordination in the World’s Electoral Systems. Cambridge: Cambridge University Press.

Dalton, Russell J. und Martin P. Wattenberg, Hrsg. (2000). Parties without Partisans. Oxford: Oxford University Press.

Downs, Anthony (1957). An Economic Theory of Democracy. New York: Harper.

Duverger, Maurice (1951). Les Partis Politiques. Paris: Colin.

Eijk, Cees van der und Martin Kroh (2002). Alchemy or Science? Discrete Choice Models for Analyzing Voter Choice in Multi-Party Contests. Paper Delivered at the Annual Meeting of the American Political Science Association 2002. Boston: APSA. url: http://apsaproceedings.cup.org/Site/papers/008/008003vanderEijk.pdf(05.01.2005).

Elff, Martin (2007). „Social Structure and Electoral Behavior in Comparative Perspective: The Decline of Social Cleavages in Western Europe Revisited“. In: Perspectives on Politics 5 (2), S. 277¨C294. issn: 1541-0986. doi: 10. 1017/S1537592707070788.

Falter, Jürgen W. und Harald Schoen, Hrsg. (2005). Handbuch Wahlforschung. Wiesbaden: VS Verlag für Sozialwissenschaften.

Fiorina, Morris P. (2002). „Parties and Partisanship: A 40-Year Retrospective“. In:

Franklin, Mark N., Cees van der Eijk und Erik Oppenhuis (1995). „The Institutional Context: Turnout“. In: Choosing Europe? The European Electorate and National Politics in the Face of Union. Hrsg. von Cees van der Eijk und Mark N. Franklin. Ann Arbor: The University of Michigan Press, S. 306¨C331.

Grofman, Bernard (1993). „Is Turnout the Paradox That Ate Rational Choice Theory?“ In: Information, Participation, and Choice. An Economic Theory of Democracy in Perspective. Hrsg. von Bernard Grofman. Ann Arbor: Michigan University Press, S. 93¨C103.

Hotelling, Harold (1929). „Stability in Competition“. In: The Economic Journal 39.153, S. 41¨C57.

Karvonen, Lauri und Jostein Ryssevik (2001). „How Bright Was the future? The Study of Parties, Cleavages and Voters in the Age of the Technological Revolution“. In: Party Systems and Voter Alignments Revisited. Hrsg. von Lauri Karvonen und Stein Kuhnle. London: Routledge, S. 45¨C61.

Key, V.[ladimer] O.[rlando] (1959). „Secular Realignment and the Party System“. In: Journal of Politics 21.2, S. 198¨C210. doi: 10.2307/2127162.

Lagos, Marta (2008). „International Comparative Surveys: Their Purpose, Content and Methodological Challenges“. In: The Sage Handbook of Public Opinion Research. Hrsg. von Wolfgang Donsbach und Michael W. Traugoot. Sage, S. 580¨C593. doi: 10.4135/9781848607910.

Lazarsfeld, Paul F., Bernard Berelson und Hazel Gaudet (1944). The People’s Choice. How the Voter Makes up His Mind in a Presidential Campaign. Chicago: Columbia University Press.

Lewis-Beck, Michael S. (1997). „Who’s the Chef? Economic Voting Under a Dual Executive“. In: European Journal of Political Research 31.3, S. 315¨C325.

Lewis-Beck, Michael S. und Martin Paldam (2000). „Economic Voting: An Introduction“. In:

Lewis-Beck, Michael S. und Mary Stegmaier (2009). „American voter to economic voter: Evolution of an idea“. In: Electoral Studies 28.4. Special issue on The American Voter Revisited, S. 625¨C631. doi: DOI:10.1016/j. electstud.2009.05.023.

Lipset, Seymour Martin und Stein Rokkan (1967). „Cleavage Structures, Party Systems, and Voter Alignments: An Introduction“. In: Party Systems and Voter Alignments: Cross-National Perspectives. Hrsg. von Seymour Martin Lipset und Stein Rokkan. New York, London: Collier-Macmillan, S. 1¨C64.

Lubbers, Marcel, Mérove Gijsberts und Peer Scheepers (2002). „Extreme Right-Wing Voting in Western Europe“. In: European Journal of Political Research 41, S. 345¨C378.

Miller, Warren E. (1994). „An Organizational History of the Intellectual Origins of the American National Election Studies“. In: European Journal of Political Research 25, S. 247¨C265.

Miller, Warren E. und J. Merrill Shanks (1996). The New American Voter. Cambridge, London: Harvard University Press.

Mochmann, Ekkehard (2002). „Zur Institutionalisierung der international vergleichenden Wahlforschung“. In: Bürger und Demokratie in Ost und West. Studien zur politischen Kultur und zum politischen Prozeß. Hrsg. von Dieter Fuchs, Edeltraud Roller und Bernhard Weßels. Wiesbaden: Westdeutscher Verlag, S. 227¨C241.

Nadeau, Richard, Richard G. Niemi und Antoine Yoshinaka (2002). „A Cross-National Analysis of Economic Voting: Taking Account of the Political Context across Time and Nations“. In: Electoral Studies 21, S. 403¨C423.

Nannestad, Peter und Martin Paldam (1994). „The VP-Function – a Survey of the Literature on Vote and Popularity Functions after 25 Years“. In: Public Choice 79.3-4, S. 213¨C245.

Nieuwbeerta, Paul und Nan Dirk Graaf (2001). „Traditional Class Voting in Twenty Postwar Societies“. In: The End of Class Politics? Class Voting in Comparative Context. Hrsg. von Geoffrey Evans. Oxford: Oxford University Press, S. 23¨C56.

Norris, Pippa, Hrsg. (1999). Critical Citizens. Global Support for Democratic Government. Oxford u.a.: Oxford University Press.

Pappi, Franz Urban (2000). „Zur Theorie des Parteienwettbewerbs“. In: 50 Jahre Empirische Wahlforschung in Deutschland. Entwicklungen, Befunde, Perspektiven, Daten. Hrsg. von Markus Klein u. a. Wiesbaden: Westdeutscher Verlag, S. 85¨C105.

Pappi, Franz-Urban und Susumu Shikano (2007). Wahl- und Wählerforschung. Baden-Baden: Nomos.

Parsons, Talcott (1960). „Pattern Variables Revisited“. In: American Sociological Review 25, S. 467¨C483.

Plutzer, Eric (2002). „Becoming a Habitual Voter. Inertia, Resources, and Growth in Young Adulthood“. In: American Political Science Review 96, S. 41¨C56.

Roth, Dieter (2008). Empirische Wahlforschung. Ursprung, Theorien, Instrumente und Methoden. 2. Aufl. Wiesbaden: VS Verlag für Sozialwissenschaften.

Rudi, Tatjana und Harald Schoen (2005). „Ein Vergleich von Theorien zur Erklärung des Wählerverhaltens“. In: Handbuch Wahlforschung. Hrsg. von Jürgen W. Falter und Harald Schoen. Wiesbaden: VS Verlag für Sozialwissenschaften, S. 305¨C323.

Schmitt, Hermann u. a. (2009). The Mannheim Eurobarometer Trendfile 1970-2002. Data Set Edition 2.01. Codebook and Unweighted Frequency Distributions. Updated by Iris Leim and Meinhard Morschner, ZA Cologne. Gesis. url: http://info1.gesis.org/dbksearch/file.asp?file=ZA3521_cod_v2-0-1.pdf.

Siegfried, André (1913). Tableau politique de la France de l’Ouest sous la Troisieme Republique. Paris: A. Colin.

Steenbergen, Marco R. und Bradford S. Jones (2002). „Modelling Multilevel Data Structures“. In: American Journal of Political Science 46, S. 218¨C237. url: \url{u:\work\Texte\PDF-Dateien\Jones1997.pdf}.

Stegmueller, Daniel (2013). „How Many Countries for Multilevel Modeling? A Comparison of Frequentist and Bayesian Approaches“. In: American Journal of Political Science. doi: 10.1111/ajps.12001.

Swank, Duane und Hans-Georg Betz (2003). „Globalization, the Welfare State and Right-Wing Populism in Western Europe“. In: Socio-Economic Review 1, S. 215¨C245.

Weisberg, Herbert F. und Steven H. Greene (2003). „The Political Psychology of Party Identification“. In: Electoral Democracy. Hrsg. von Michael B. MacKuen und George Rabinowitz. University of Michigan Press, S. 83¨C124.

Western, Bruce und Simon Jackman (1994). „Bayesian Inference for Comparative Research“. In: American Political Science Review 88, S. 412¨C423.

1Naturgemäß können diese Ansätze hier nur in extrem verkürzter Form skizziert. Ausführlichere Darstellungen finden sich in den Lehrbüchern von Bürklin und Klein (1998), Pappi und Shikano (2007) und Roth (2008). Eine umfangreiche Würdigung dieser und anderer Ansätze bietet das von Falter und Schoen herausgegebene Handbuch Wahlforschung (Falter und Schoen, 2005).

2Die vier Grundtypen von Konflikten ¨C Staat vs. (katholische) Kirche, Zentrum vs. Peripherie, Arbeit vs. Kapital und Stadt vs. Land ¨C stehen im Zusammenhang mit revolutionären sozialen Umwälzungen in der Geschichte des jeweiligen Landes. Lipset und Rokkan begründen ihren ursprünglichen Ansatz unter Rückgriff auf das Werk Talcott Parsons’ (Parsons, 1960). Für die weitere Rezeptionsgeschichte spielte dieser systemtheoretische Unterbau aber keine Rolle.

3Dies gilt sofern die Wahlteilnahme keinen ergebnisunabhängigen (intrinsischen) Nutzen stiftet, was aber den Grundannahmen des Modells widersprechen würde.

4Taktisches Wählen liegt dann vor, wenn sich eine Wählerin bewußt nicht für die eigentlich bevorzugte Partei entscheidet, etwa weil sie glaubt, daß diese in einem Mehrheitswahlsystem ohnehin keine Chance hat, ins Parlament einzuziehen.

5Wie Lewis-Beck und Stegmaier (2009) zeigen, beschäftigte sich aber bereits die Ann Arbor-Gruppe mit diesem Thema. Der Ansatz ist damit auch mit einer sozialpsychologischen Perspektive kompatibel.

6So gibt es beispielsweise innerhalb der Europäischen Union keinen föderalen Staat mit einem Mehrheitswahlsystem.

Psephology and Technology, or: The Rise and Rise of the Script-Kiddie


1    Introduction

psephology and technology

Download this text as PDF

From its very beginnings, psephology has been at the forefront of methodology and has sometimes pushed its boundaries (see e.g. King, 1997 on ecological regression). Methods such as factor analysis or logistic regression, that were considered advanced in the 1990s, are now part of many MA and even BA programs. Driven by the proliferation of fresh data and the availability of ever faster computers and more advanced, yet more user-friendly software, the pace of technical progress has once more accelerated over the last decade or so. Hence, and somewhat paradoxically, this chapter cannot hope to give a definitive account of the state of the art: The moment this book will ship, the chapter would already be outdated. Instead, it tries to identify important trends that have emerged in the last 15 years as well as likely trajectories for future research.
More specifically, the next section (2) discusses the general impact of the “open” movements on electoral research. Section 3 is devoted to new(ish) statistical methods – readily implemented in open source software – that are necessitated by the availability of data – often available under comparably open models for data distribution – that are structured in new ways. Section 4 is a primer on software tools that were developed in the open source community and that are currently underused in electoral research, whereas the penultimate section discusses the double role of the internet as both an infrastructure for and an object of electoral research. Section 6 summarises the main points.

2    Open Source, Open Data, Open Science

Like many other subfields in the social sciences, psephology is heavily affected the rapid progress in computer and information technology. The two most significant developments in this respect are the twin open-source and open-data revolutions. Open source software has its roots in  the  free  software  movement  of  the  1980s  (Lakhani and Hippel, 2003), a rebellion against increasingly more restrictive software licences that, amongst other things, aimed at patenting algorithms and banned the “reverse engineering” of software installed on private computers. Proponents of the free software movement, on the other hand, made their software available for free (“free as in free beer”) and gave everyone and anyone the licence to modify their programs as they saw fit (“free as in free speech”), which required laying open the source code. The spread of the internet in the 1990s then facilitated large-scale collaboration on free software projects and gave rise to the current idea of open source software that is embodied in Raymond’s (1999) manifesto “The Cathedral and the Bazaar”, which emphasises the idea of distributed and only loosely co-ordinated teams as a strategy for quick and efficient development.
Whereas the free software movement had a certain anti-establishment bent, many of the largest and most successful open source projects, such as the Linux operating system, the Apache web server or the Firefox browser series, which collectively power much of the current internet, are happy to rely on the support of corporate backers who donate money, resources, and the time of some of their staff. In other instances, large companies have even created open “community editions” of their existing programs, or designed them as open source applications in the first place (Google’s Android operating system). Companies may do this to raise their profile, or in order to attract the best software engineers for their commercial projects, but two other motives are more interesting: They may want to use the open source software instead of a closed-source alternative to generate and deliver their own products (e.g. tech companies relying on Linux for running their server farms), or they may offer a service that is based on the open-source software (professional support or hosted versions). Either way, corporate support for open source makes commercial sense – neatly illustrating Olson’s (1965) argument about big players rationally investing in public goods – because open source is a, as Raymond suggests, a highly efficient model for organising large projects: It incorporates feedback from the user base almost instantaneously and turns the most capable and committed users into developers.
Open source is highly relevant for psepholog not only because it helped to build much of the internet infrastructure and some key tools – R (Ihaka and Gentleman, 1996; Crawley, 2013), Python (Lutz, 2013), and a plethora of others – but also because it has become the template for other “open” revolutions that impact on electoral research. In the broadest sense, open  data  refers  to  the  idea  that  research  data,  or data that could be used for research,  should be as accessible as possible.  As such, it is old news. In the quantitative social sciences, data archives such as the Roper Center (http://ropercenter.cornell.edu/) or Michigan’s Survey Research Center (https://www.src.isr.umich.edu/), which collect, archive, and  disseminate  existing data for secondary analyses, were established in the late 1940s. Patterns of co-operation and exchange between (European) archives were formalised with the formation of the Council of European Social Science Data Archives (CESSDA, http://cessda.net/) in the 1970s (Karvonen and Ryssevik, 2001, p. 45). In the public sector, one could argue that the practice of frequently publishing detailed information on key national statistics that was already well established in the late 19th century marks the beginning of open data. However, it was the key ingredients of the open source revolution – transparency, active involvement of the user base, and almost zero marginal transaction costs – that in the 2000s began to transform the production and use of data in unprecedented ways. Unless access is restricted for reasons of data protection, researchers no longer have to travel to a data archive to use a given data set, and for the distribution of data, physical media haven been all but abolished. Governments, large-scale research projects, and individual scholars are now opening up their raw data for download. Some agencies and some of the biggest internet companies (e.g. Google, Facebook, Twitter, and Yahoo) have even created application programming interfaces (APIs, see section 5.1) that give researchers the opportunity to access these data programmatically from a script.
The open data revolution has brought about some new problems of its own. While the body of data available for research is growing exponentially, researchers still have to know where and how to look, and the lack of a central repository and common interfaces seriously hampers progress. To be useful, data need to be stored, and, even more importantly, described and licenced in standardised ways that make them accessible and retrievable in the medium-to-long term. This, in turn, requires institutions that can be trusted, and that need funding. Moreover, the pressure on researchers to open up their own data is growing. Research councils now regularly make the deposition of research data and even the open access publication of the findings a precondition for funding. Similarly, more and more journals require that not only the data set itself but also the program code that generates the tables and graphs must be published along with thefinal article in some repository (see section 5.1).1 While such rules reinforce traditional scientific standards of honesty,  transparency,  and  reproducibility,  many  researchers are still anxious that they will be scooped if they are forced to reveal their data and methods at the beginning of a new project. Presumably due to the prevailing incentive structure, few social scientists currently adhere to the open source mantra of “release early, release often.” Others, however, embrace the ideal of a (more) open science by posting work-in-progress on their personal homepages, opening draft chapters on social network sites for scientists, or even by moving their data and manuscripts to open source development sites such as Github, which could in theory provide an ideal environment for scientific collaboration.

3    Data, Statistical Models, and Software

3.1    Complex Data Structures and Statistical Models

Pure formal theory and simulation exercises aside, all electoral research rests on data: a body of systematic and usually quantified observations that can be used to test assumptions about the ways in which citizens, politicians, organised interests, and the media interact and thereby affect political decisions. While early studies emphasised the importance of macro factors (Siegfried, 1913) and of clustered sampling and mixed methods (Lazarsfeld, Berelson, and Gaudet, 1944), the lasting influence of the American Voter (Campbell et al., 1960) has led many researchers to focus on micro-level data coming from nationally representative samples of mass publics for much of the 1960s, 1970s, and 1980s.
But theory suggests that credible (or at least plausible) accounts of human behaviour must encompass not just the individual (micro), but also the societal (macro) level, and ideally various “meso” layers and structures in between (see Coleman,  1994 for the general line of reasoning and Miller and Shanks, 1996 for an application to election studies). The thrust of this argument eventually led to a renewed interest in contextual variables and their effects (Jennings, 2007, pp. 35–38).  From the late 1990s and early 2000s on,  national election studies and comparative surveys alike began to include territorial identifier variables such as electoral district codes in their datasets. Using this information, it is possible to match data on individual respondents with government figures on the economy, on migration, and a whole host of other variables that can
1Pre-registration, a procedure that is becoming more prevalent in the Life Sciences and whose adoption in Political Science is now being discussed (Monogan, 2015), goes on step further by demanding that researchers submit sampling plans, outlines of the intended analyses, and mock reports to a journal that are peer-reviewed before they even begin to collect new data.

plausibly affect voting behaviour, while Multi-Level regression (see chapter 47) is a convenient tool for estimating the size of the alleged effects and their associated standard errors. Supplementing micro-level information with contextual variables leads to “nested” data, where each level-1 unit (respondent) belongs to one (and only one) level-2 unit (electoral) district. Each level-2 unit may in turn be part of one (and only one) level-3 unit (say, a province), resulting in a tree-like structure.
Multi-Level regression modelling with contextual covariates derived from official sources has become almost the de facto standard for analysing both large-scale comparative data sets (see chapter 48) and case studies of nations for which sub-national data are available. While the technique provides asymptotically correct standard errors and opens up a number of flexible modelling options (see section 3.2.1), it is no panacea. When nations are the relevant contexts, their number is often too low for Multi-Level Modelling (Stegmueller, 2013), and one may well ask if it makes sense at all to treat countries as if they were a random sample from a larger population (Western and Jackman, 1994). Comparing political behaviour within subnational units across nations is more informative and often more appropriate, but suffers from specific limitations, too: Even within the European Union’s complex and comprehensive system for the Nomenclature of Territorial Units for Statistics (NUTS, see Eurostat, 2015), subnational units that are supposed to be on the same level may differ vastly in terms of their size, population, and political, social and cultural relevance.2
Moreover, the integration of government statistics as regressors into a Multi-Level Model does not nearly exhaust the complexity of data that are now available for analysis. Building on earlier work by Lazarsfeld and Menzel (1961), Hox (2010) has developed a useful typology that clarifies the possibilities. On each level, there are global variables, which reflect inherent properties of the objects on the respective level. They are inherent in so far as they can neither be constructed by aggregating features of lower-level objects, nor by disaggregating features of higher-level contexts. Traditional (statistical) models of voting behaviour have focused on global variables at the indvidual level (level 1): an invidual vote for the Democrats is linked to the voter in question being female,
2NUTS-1 corresponds to the 16 powerful federal states in Germany, to clusters of provinces, states, or communities that have been grouped together for purely statistical purposes in Austria, Spain, and the Netherlands, and does not exist at all in many of the smaller countries (e.g. Croatia, Denmark, Luxembourg, or Slovenia). The lower-tier NUTS-2 level is equivalent to the federal states in Austria, the autonomous communities in Spain, the Regions in France, and the Provinces in the Netherlands, which all have their own elected assemblies. In other states such as Bulgaria, Finland, Germany, Romania, or Slovenia, NUTS-2 areas exist solely for the purpose of national planning and attracting EU funding, and citizens will be unaware of their existence. Similarly, NUTS-3 may be a district (Germany), a group of districts (Austria), a province (Denmark, Spain, Italy), a region (Finland), a statistical region (Slovenia), an island (Malta), or may not even exist (Cyprus, Luxembourg).

Level    1    2    3    . . .
Type of variable    global    →    analytical
relational    →    structural
contextual    ←    global    →    analytical
relational    →    structural
contextual    ←    global    →
relational    →
contextual    ←
→: Aggregation
←: Disaggregation
Source: Adapted from Hox (2010, p. 2)
Figure 1: A Typology of complex data structures

unemployed, and identifying as a Democrat. A prototypical Multi-Level Model would add the unemployment rate and the ethnic composition of the electoral district as level-2 regressors. These are analytical variables, which are created by aggregating global features of the lower-level units to form upper-level averages, ratios, or percentages. As a corollary, these variables can enter the model simultaneously on multiple levels (see section 3.2.1).
Other properties of the district may also be meaningful additions to the model, but they cannot be understood as an aggregation of individual-level qualities or disaggregation of higher-level features and are hence global variables at the district level. Gender and political experience of the main candidates are cases in point. Because there is no variable at the lowest level that would correspond to them, they are strictly contextual for individual voters and can enter the model only once, at the upper level.
Finally, relational data convey information regarding the ties (e.g. presence and intensity of face-to-face contacts) between objects on the same level. Such network data are crucial for any micro-sociological explantation of voting behaviour: Obviously, a person that is the hub of a Democratic clique of friends is more likely to turn out to vote, and to vote in accordance with her peers than someone who is socially isolated. Like global/analytical variables, network data can enter a Multi-Level Model simultaneously on multiple levels: Information on relations between individual voters within a district may be aggregated to form structural variables at the upper level, e.g. to compare districts with dense/sparse or homogeneous/fragmented communication networks.
Network data are extremely attractive in theory. But they introduce an additional level of complexity and require specialised statistical methods, because a tie by definition

involves two actors (see section 3.2.3). In addition, the collection of relational data necessitates specific (cluster) sampling plans, because a large number of the members of a given network needs to be surveyed to assess the properties of the network itself. This, in turn, raises issues of representativeness, data confidentiality, and cost-effectiveness and goes against the dogma of the nationally representative sample.
Election surveys sometimes contain items refering to so-called egocentric networks,
e.g. they might ask the respondent how many people she talks politics with, whether these are friends, family members, or just acquaintances, and how often she disagrees with them. But this information will be biased by the respondent’s perceptions and provides only a fractional glimpse into the full network, as normally not even the ties amongst the respondent’s immediate contacts can be reliably recovered.
As a readily available alternative, students of electoral behaviour are now turning to social media, where large and mostly complete political communication networks can be sampled and observed with ease. Just how well insights from these networks generalise to offline behaviour and the voting population as a whole is a different question. Either way, statistical procedures for analysing social networks are currently in the process of becoming part of the tool kit for electoral research.
Besides Multi-Level and network data, the use of spatial or geo-referenced data is another emerging trend in electoral studies. A geo-reference is simply a set of coordinates that locate an object in space. Coordinates can either define a point or an area (polygon). In the most simple sense, the territorial identifiers mentioned above record that a voter is living in a given (usually large) area and hence are geo-references, too. More precise coordinates for voters (e.g. census blocks, ZIP code segments, electoral wards, street addresses, or even GPS readings), however, allow researchers to locate voters within much smaller contexts, for which census and market research data – in other words, global and analytical variables that can be integrated into a Multi-Level Model of electoral choice – may be available. Whilst many researchers are familiar with the idea of coarse geo-references, the availability of very fine-grained data as well as the growing awareness of spatial dependencies necessitates specialised software and models for the proper analysis of geo-referenced data (see section 3.2.4)

3.2    Statistical techniques and software implementations

3.2.1    Multi-Level Models and Structural Equation Models

As outlined above, students of electoral behaviour routinely collect data that, reflecting the Multi-Level nature of the underlying theoretical explanations, exhibit complex structures. Statistical Multi-Level Models, which are also known as “mixed models” or “random coefficient models” are the most adequate means to deal with such data.

They account for the correlation of unmeasured disturbances within a given context and hence provide correct standard errors for the effects of macro-level variables. Moreover, they model context specific disturbances in the most efficient way possible by treating them as random. This is best illustrated with an example: In a study of N voters living in K electoral districts that aims at explaining individual turnout, one could try to capture the effects of unmeasured district-level variables (say local social capital) by introducing district specific intercepts (dummy variables). But this strategy has negative consequences for the identification of the model and becomes inefficient and impractical very quickly as the number of districts which are sampled grows (Steenbergen and Jones, 2002). A statistical Multi-Level Model will replace the K-1 estimates for the local intercepts with a single estimate of their variation over districts (a random intercept) and thus dramatically reduces the number of parameters.
Moreover, Multi-Level Models also provide for a number of additional advanced modelling options. If there are good reasons to believe that the impact of an explanatory variable (say ideology as measured by left-right self-placement) on turnout will vary considerably across the K districts, the analyst can specify a random effect for this variable, which supplements the estimate for the average effect of ideology (the traditional point estimate) with an estimate of its variation. As the name implies, random effects are adequate if the variation in the effect of an independent variable can plausibly be treated as random.
If, on the other hand, the impact of a variable varies in a systematic fashion, this can be modelled by specifying a cross-level interaction, e. g. between ideology (a micro-level variable) and the number of candidates standing in the district. Cross-level interactions need not be confined to variables that are as conceptually different as the two in this example. On the contrary, theory often suggests that a variable such as unemployment could in essence interact with itself, albeit on different levels, hence entering the model thrice: as an individual feature (a global variable on the micro-level), as an analytical variable (the unemployment rate on the district-level), and as an cross-level interaction between the two. A high unemployment rate may reduce the propensity to participate in an election for all citizens, and individual unemployment status will normally depress turnout in an even stronger fashion. But this micro-level effect may well be confined to low-unemployment level environments, whereas individual unemployment  may have no such negative impact or even increase the likelihood of voting in districts where the high unemployment rate attracts community organisers and other political entrepreneurs. Multi-Level Models are ideally suited for disentangling such complex causal relationships.
They can also deal with complexly structured political contexts  that  may  have many tiers (voters within households within wards within municipalities within districts

within provinces . . . ), and that may cross-cut and overlap instead of forming a neat, tree-like hierarchy: A voter is not just affected by the characteristics of the electoral district she is living in but has been politically socialised in completely different surroundings. Whilst Multi-Level Models can accommodate such complex structures, convergence will generally be slow, and estimates may be unstable. As with all other aspects of modelling,  analysts  should  therefore  strive  for  parsimony.  If  the  there are not variables at the higher levels and if the objective is simply to reflect the multi-stage nature of the underlying sampling process, traditional survey estimators or even Huber-White standard errors that account for clustering may provide a fast and robust alternative to a fully specified Multi-Level Model.
Having said that, Multi-Level Models are a very flexible tool, as contexts need not be defined in spatial terms. For the analysis of panel data, it often makes sense to think of individual respondents as “contexts” for the interviews conducted in successive panel waves. Particularly when panel data are imbalanced or collected at irregular intervals, Multi-Level models can alleviate some of the problems that plague the traditional approaches to panel data.
Another statistical technique that has become indispensable for students of electoral behaviour is Structural Equation Modelling (SEM). SEM is an extension of traditional factor analysis that lets researchers specify multi-indicator measurement models for otherwise unobservable (=latent) theoretical constructs such as political attitudes.  It is attractive, because it can simultaneously estimate coefficients for whole systems of equations, and because it can incorporate measurement models for attitudinal variables that account for relatively unreliable  indicators.  If  the  measurement  models  hold, SEM can also provide unbiased estimates of the equally unobservable, “structural” relationships amongst the latent variables. Given adequate data, it is possible to map a whole system of constructs and hypotheses about their relationships onto an equivalent system of equations.
In the past, its application in election studies was somewhat limited by the fact that they required measurements on a continuous scale that were distributed multivariate normal, whereas the key dependent variable in election studies as well as many relevant independent variables are categorical and usually distributed with considerable skew. In the 1990s, however, new estimators were developed that can accommodate non-normally distributed continuous data. In addition, generalisations of the original model allow for ordinal and nominal indicator variables and even for categorical latent variables (Jöreskog, 1990; Jöreskog, 1994; Muthén, 1979; Muthén, 2002). Moreover, Multi-Level Models and Structural Equation Models are closely related (Muthén, 2002; Skrondal and Rabe-Hesketh, 2004) and can be combined to form Multi-Level Structural Equation Models.

Until recently, the estimation of Multi-Level or Structural Equation Models required specialised (if relatively user-friendly) software: HLM or MLWin for Multi-Level Modelling, LISREL, EQS, or AMOS for SEM, and MPlus for either. This is no longer true: Recent versions of Stata – currently the most popular general purpose statistical package in political science – can estimate all but the most complex Multi-Level and Structural Equation Models and so greatly extend the potential user base of these techniques. SPSS, another popular package, has some Multi-Level capabilities and works closely with AMOS, an SEM software that SPSS Inc. acquired in 2003 before it was in turn bought by IBM in 2009.
Perhaps more importantly, there are packages available for the R programming language that provide similar features: Lme4 and Rstan for Multi-Level Modelling, and Lavaan and Sem for SEM. While they may be slightly less capable, slower, and generally more clunky than the commercial software, they are, like any other R-package and the core of the language itself, open source and freely available for almost any combination of hardware and operating system. Moreover, while they may lack professional documentation and customer service,  they are supported by a global community of enthusiasts, scriptable in a fully-fledged programming language with flexible data structures, and tie into the ever growing eco-system of more than 6000 user-written packages for R that aim at implementing the latest developments in statistics.

3.2.2    Bayesian methods

Most electoral researchers were trained within a “frequentist” framework of statistical reasoning that relies on the idea of a random sampling process that could be endlessly repeated under essentially identical conditions. So far, they have shown only modest interest in the (sometimes exaggerated) benefits of an alternative statistical framework: Bayesian statistics (Jackman, 2004). There are at least two causes for this inertia: The frequentist paradigm closely resembles the pattern of taking national large-scale random samples of the general population that has been the workhorse of election studies for most of the last seven decades,  and in such large samples Bayesian and frequentist estimates will normally closely resemble each other.
But whether applied researchers like it or not, the ever more popular Multi-Level Models are Bayesian models at their core (Gelman and Hill, 2007). While many political scientists still have some reservations regarding the underlying paradigm (or might be blissfully unaware of it), Bayesian statistics keeps making inroads into electoral research. There are a number of reasons for this. First, Bayesian models can sometimes be tailored to a problem for which no off-the-shelf frequentist solution has been implemented in any statistical package. Models that aim at predicting the distribution of seats in parliament from a rolling series of published opinion surveys are case in point.  Second, Bayesian

statistics may be able to provide an estimator that is better in terms of bias and efficiency than any frequentist alternative, as it is the case with Multi-Level Models and some SEMs. Third, Bayesian statistics, which for most of its existence was a rather arcane pasttime because of the computational demands implied, only gained practical relevance for applied researchers with the twin advent of simulation-based methods and affordable fast processors in the late-1990s-to-early 2000s. Even a decade ago, getting Bayesian estimates in MLWin for a reasonably complex Multi-Level Model could easily take an hour or more on a then-modern desktop computer, much as it was the case with SEM in the 1990s.
At the moment, most Bayesian estimation still requires access to specialised software (Winbugs, Openbugs, Jags, Stan . . . ), preferably via R. However, the implementation of Bayesian analysis in recent editions of Stata (from version 14 on) could be a game-changer in this respect.

3.2.3    Networks

So far, election studies have mostly missed out on the renaissance of Social Network Analysis (SNA) in political science (for some notable exceptions see e. g. Huckfeldt and Sprague, 1987; McClurg, 2006). Although interest in relational or network data has grown exponentially in Political Science, psephology has been somewhat late to the party because relevant data are generally not available. While large societies may display the properties of a “small-world” network in which everyone is related to everyone else through a relatively small number of contacts (say six), such network structures are very sparse and will rarely have an effect on political behaviour. Social embededness certainly plays a role for opinion formation and political behaviour, but mainstream election studies cannot hope to uncover the relevant networks. Traditional community studies as well as explorations of online communities, on the other hand, can do just that.
Although it is far from clear if and how findings gained here generalise to the electorate as a whole, statistical procedures for analysing social networks are currently in the process of becoming part of the tool kit for electoral research. Understanding these methods can present a formidable challenge.
By definition, network data break the mould of traditional data analysis, where cases correspond to the rows of the data matrix, and variables to its columns. In network applications, cases form both the rows and the columns of an (adjacency) data matrix, whose cells represent the existence, direction and possibly strength of ties between them. Recording traditional variables requires a second data matrix, specialised software, and, more importantly, and adjustment of the analyst’s mind set.
Once collected, data on ties between actors can be employed to calculate three broad

classes of statistical measures (Knoke and Yang, 2008): indices that reflect the position of an individual within the local or global network (e.g. one’s centrality), measures that refer to features of an actual or potential tie between two actors (e.g. the importance of this tie for the coherence of the network as a whole), and statistics that describe some features of the network as a whole (e.g. the degree to which it resembles the “small-world” scenario outlined above). Often, the aim of SNA is chiefly descriptive and the analysis would end with their calculation and interpretation, but in principle, all network measures can subsequently be employed as dependent or independent variables in a regression framework.
Relational data do not fit into the single-matrix paradigm of general statistical packages such as Stata or SPSS. Moreover, before the rise of social networking sites, there was little commercial interest in SNA. Therefore, most software that is potentially useful for students of electoral behaviour is developed by academics (often as an open source project) and available for free or at a very modest price. Historically, UCINET, which was created in the early 1980s and has been under constant development ever since has been a very popular choice. UCINET is grounded in the tradition of (mathematical) sociology and incorporates a multitude of procedures for manipulating and analysing relational data. However, according to its authors, many of these procedures become tediously slow in networks with more than 5000 nodes. Pajek and Pajek XXL, on the other hand, are slightly newer programs specifically geared towards large and very large networks of several million nodes. Their user interface is idiosyncratic, and the terminology used in the documentation as well as many of the procedures may be unfamiliar to social scientists, as the authors have their roots in mathematical graph theory and computer science. However, Pajek is unrivalled in terms of speed and sheer processing capacity.
UCINET, Pajek, and other SNA software make it possible to perform analyses that are unfeasible with standard statistical software. However, moving one’s data from a standard software suite to an external program for network analysis, then back to the general purpose package for further analysis is a disruptive, tedious, and error-prone process. The various SNA packages that exist for the R system are therefore an attractive alternative to the stand-alone SNA programs. The most generally useful are Statnet (a “meta” package that includes many procedures from more specialised packages), and Igraph, which seems to be slightly more accessible (and is also available as a package for the Python language). In all likelihood, either package will fulfil all but the most exotic needs of psephologists.

3.2.4    Geo-spatial analysis

Geo-spatial analysis is a broad term that encompasses at least two distinct (if related)
approaches: the use of geographical variables in “normal” regression models of electoral 12

behaviour on the one hand, and the estimation of specific statistical models that account for spatial dependencies on the other.
The  first  approach  may  simply  use  geo-references  to  merge  micro  data  with contextual information (see section 3.1). Under more advanced scenarios, psephologists will calculate geographical variables (most often distances) from sets of geo-references. This is best illustrated by an example: For various theoretical reasons, voters should ceteris paribus prefer local candidates, i. e. candidates that live closer to a given voter’s residence than other candidates. If candidates are obliged to have their home addresses on the ballot paper and the addresses of voters are known,3 the spatial distance between candidates and their prospective voters can be calculated (Arzheimer and Evans, 2012; Arzheimer and Evans, 2014).  This variable varies across voter-candidate pairs within districts (unless voters live at the same address) and is therefore a global variable at the
level of the individual voters.
Geo-spatial methods are required for (1) the translation of addresses into physical co-ordinates (a step known as geocoding) and (2) the calculation of various distance measures (e. g. travel time by car or public transport). Apart from the calculation of straight-line distance, which is a purely geometrical problem, the second step requires access to digital road maps, timetables, data on congestion, and routing algorithms. However, once the distance has been calculated, the analysis can proceed with the usual linear and non-linear regression models, which may account for nesting or clustering of the observations by imposing a structure on the variance-covariance matrix.
Various types of spatial regression models take this idea one step further. They correct for dependencies amongst the observations by taking the spatial co-ordinates of cases into account and adjusting the structure of the variance-covariance matrix accordingly. The relevance of spatial regression models for psephology is most obvious in the case of district-level aggregate analyses:  Whereas standard regression models assume that disturbances are identically and independently distributed, it stands to reason that neighbouring4  districts will be affected by similar disturbances and hence will display a  pattern  auto-correlation  that  renders  standard  errors  dubious  at  best.    In  spatial regression, the matrix of distances between the centroids of the electoral districts can be employed to estimate this auto-correlation, which in turn can be used to derive corrected standard errors within a spatial regression model (Ward and Skrede Gleditsch, 2008). Spatial regression could be applied to individual-level data too, but it is generally easier
3For reasons of data protection, usually only an approximate geo-reference of the respondent is recorded.
4Being neighbours is a somewhat fluid concept, as these shared influences will be stronger where districts are physically closer and less pronounced, yet still present, where a pair of districts is further apart. This scenario is very different from nesting, where there are clearly delineated, fixed groups of lower-level units.

and often also more appropriate in terms of the underlying theoretical assumptions about causal mechanisms to use a Multi-Level Model (possibly involving more than two levels) that accounts for nesting within political-administrative contexts.
Mapping and processing geo-referenced data traditionally required access to and training in the use of a Geographical Information System (GIS). A GIS is essentially a relational database with special capabilities for dealing with 2D- and 3D-coordinates. GIS software tends to be expensive, proprietary, and complex. In recent years, however, government agencies and other organisations aiming at opening up their data have set up websites that hide at least some of the complexity of the underlying system. In the most simple case, users may create chloropleth maps, or look up data for a single or a limited number of localities. More useful systems allow one to download pre-built or customised tables in some machine-readable format that can be merged with existing individual-level data. In very few ideal cases, there is an API that researchers can access programatically (see section 5.1).
Moreover, algorithms for collecting, storing, and processing geo-referenced data are now freely available and have been implemented in a host of stand-alone programs and/or packages for the R system. GRASS (Geographic Resources Analysis Support System) is a fully featured GIS that has a wide range of applications in engineering, the natural science, and the social sciences. GRASS runs on all major  operating systems. It can be used both interactively through its graphical user interface (GUI) and programmatically via scripts. Its real power, however, lies in its interfaces with two popular programming languages: Python and R. Through these interfaces (Pygrass for Python and Rgrass6/7), users can at the one hand program the GRASS system and extend its capabilities. On the other hand, researchers who regularly run their analyses in Python or R can selectively make use of data stored in GRASS and of the nearly 2,700 industry-grade functions available in the system. QGIS is a more light-weight alternative to GRASS. While it interfaces with R and Python, too, it is mostly geared towards interactive use.
In many cases, however, analysts working in R or Python will want to altogether avoid the overhead of plugging into a GIS. Much of the functionality of traditional GIS software is now available in the form of addons for these two languages. R in particular currently has more than a hundred packages for loading, manipulating, analysing, and mapping geo-referenced data (https://cran.r-project.org/web/views/Spatial. html).

4    Tools for successful, reproducible research

The previous two sections have dwelled on the rapid pace of technical progress in psephology. Somewhat paradoxically, this section suggests that in the face of ever more complex data and software, psephologists should turn to very basic tools, concepts and techniques that computer scientists developed decades ago: plain-text files and editors, directories (folders), and some utilities commonly used in medium-sized programming projects. As the old saying goes: In election studies, regression is progress.

4.1    Establishing a reproducible workflow

Data analysis involves a number of distinct phases (see Long 2009, chapter 1 for a similar outline):

1.    Data must be collected, either by the researchers themselves or by some third party, and stored electronically
2.    These machine readable data need to be transferred to the researchers, usually via the internet
3.    The data must be recoded or otherwise normalised, possibly after being converted to some other format first.
4.    A number of exploratory analyses and preliminary models are run on the data, perhaps using more than one computer program
5.    The researchers settle on a small set of final analyses and models whose results are stored
6.    For presentation and publication,  graphs and tables are produced from  these results, possibly using additional software

To be reproducible by the original researchers and their peers, every step as well as the rationale behind the decisions involved must be documented. Realistically, that means that as much as possible of the whole process should be automated by means of scripts: short sets of instructions for a computer program. Graphical user interfaces are useful for getting to know a program, and possibly for putting the finishing touches on graphs for publication, but scripts are infinitely more powerful, efficient, and reliable. When properly commented, scripts are also self-documenting, although researchers should strive to keep a separate research journal. For smaller projects, presentations, and teaching, researchers may even want to pursue a “literate programming” (Knuth,

1984) approach that combines code for several programs, text for publication, and documentation in a single document, from which intermediate tables and graphs as well as slides and PDF documents may be produced using either the Knitr package for R or the even more general Orgmode package for Emacs (see below). However, while literate programming is attractive in principle, it may not scale well to larger projects.
Most statistics packages have simple script-editing capacities built in, but in the long term, it is more efficient to use stand-alone text editors, which offer much more powerful editing features as well as syntax highlighting, proper indentation, and basic project managing capabilities. One of the most quirky and powerful of these editors is Emacs (https://www.gnu.org/software/emacs/), which was first released in the mid-1970s and has been under active development ever since. Despite its age, interest in Emacs has surged in recent years, and many quantitative social scientists swear by it. Emacs can be endlessly customised and extended, which can be baffling for new users. Cameron et al. (2005) provide a useful introduction, but documentation for the many more features and extensions is best searched on the internet. Psephologists may also want to install one of the configurations aimed specifically at social scientists that can be found online.
With the right set of extensions, Emacs supports almost every scripting language known to humankind, including the command languages of statistical packages such as Julia, OpenBUGS/JAGS, R, S-Plus, Stan, Stata, and SAS. At a minimum, “support” means syntax highlighting, indentation, and checking for balanced parentheses. Moreover, Emacs normally gives access to the respective help systems for these languages and can find documentation for related functions. It can insert boiler plate code (e.g. loops) and can execute snippets of code or whole scripts. Emacs was designed as an editor for computer programmers and so has the ability to keep track of variables and look up the definition of functions across an arbitrary number of files, making use of text tools such as Diff, Grep, or Find, version control systems such as Git (more on them below). The more complicated the toolchain becomes, the more it shines, as R, Stata, Python, and many other applications can be conveniently managed from a single keyboard- and script-centric  interface.

4.2    Buildtools, revision control, and other open source goodies

Ideally, there should be a separate script for each of the six steps outlined in section 4.1. Shorter scripts are easier to maintain, and it would be inefficient to re-run the whole process only to add a horizontal line to a table. It is also vitally important that data are only ever edited in a non-destructive way: Each script must save its results as a new file, keeping the data collected and transferred in steps one and two in pristine condition. It is also good research practice to keep all files belonging to a given project in a directory of

their own, and to create separate sub-directories for scripts, graphs, tables, and datasets (Long, 2009).
Once a project grows beyond a handful of individual scripts, further automation of the process or meta-scripting becomes a necessity, because the individual jobs need to be executed in a certain order. In principle, a degree of automation can be achieved within the statistical package of choice itself: Both Stata and R are capable of processing scripts that it turn “include” or “source” other scripts. Moreover, both programs have rudimentary infrastructure for starting external programs and can so, at least in theory, manage a tool chain. In practice, however, it is easier and less error-prone to rely on an external scripting language, e.g. Python or the scripting language of the operating system’s native command line interpreter (shell), to manage complex workflows.
If some of the tasks involved are time-consuming or otherwise  expensive  (i.e. model estimation by numerical means or metered data acquisition from the internet), psephologists should rely on “build tools”: software that is normally used by computer programmers to compile (“build”) complex software from hundreds of text files via a potentially large number of intermediary files. If a single text file is edited, it is normally sufficient to recompile a small fraction of the whole project that is directly affected by this change. Build tools can identify, manage, visualise and most importantly exploit such dependencies, thereby making substantial efficiency gains possible.
On average, workflows for software project are more complex than workflows for the analysis of electoral data by several orders of magnitude, but psephologists can still benefit from learning to use build tools. This is best illustrated by an example. Consider the following simple workflow:

1.    Download (with R, or with a specialised program such as wget) a data set (say the European Social Survey) from the internet if the file on the web has changed
2.    Save a relevant subset of the data after recoding some variables
3.    Load the subset, estimate some complex model, and save the parameters to a file
4.    Illustrate the findings by
•    Producing a number of graphs from the parameters and save them as separate files
•    Producing a number of tables from the parameters and save them as separate files

5.    Generate a PDF report with the LATEX document preparation system by processing a text file that includes the graphs and tables

For an efficient and managable workflow, each task should be performed by a single program acting on a single set of instructions (either a script or simply a number of options and arguments submitted when the program starts). Moreover, each task takes one ore more inputs, and leaves behind one or more outputs.5 The way in which these individual tasks are listed makes it very easy to recognise the dependencies amongst them: If a new version of the European Social Survey is published, all steps must be repeated in that exact order. If, on the other hand, the researcher decides to change the coding of the variables (step 2), the estimator employed by the model (step 3), or the look of the graphs (step 4), only the subsequent steps must be repeated. Incidentally, the latter modification would not require rebuilding the tables: If the dependencies were visualised as a tree, both tasks would appear on the same level, as they are completely independent of each other. In a computing environment with sufficient ressources, they could be executed in parallel, thereby further speeding up the process.
Build tools  such  as  the  venerable Make  program  (Mecklenburg, 2005, generally available on Unix-like systems) and its many modern successors require that the dependencies are specified in yet another textfile. While this may sound like a chore, it is usually just a matter of writing down which script generates what files (“targets”) from which inputs. Moreover, this helps clarifying and streamlining the workflow. Once this set of rules is in place, the build tool will analyse the dependencies and execute the tasks in the required order. After this initial run, targets will only be regenerated if the scripts or inputs from which they originate change.
A  final  tool  that  psephologists  should  borrow  from  the  world  of  software development are revision control systems.  Most researchers will be (painfully) aware of the value of automated backup systems, which retain a number of old copies to avoid the situation where a good backup is replaced by a corrupted copy.  Modern systems usuallly provide a number of hourly or daily snapshots alongside increasingly older (weekly, monthly, yearly) copies.  Revision control systems take this idea of snapshots one step further by retaining a complete history of changes to each (text) file in a project directory.6  Modern revision control systems such as the somewhat unfortunately named Git (Loeliger and McCullough, 2012) can track the whole state of a directory and quickly reset all files in a directory to the state in which they were yesterday evening, or show which changes were made to a specific file since Monday night. They provide tools for finding the exact point at which some changes in the way the variables were recoded
5Ideally, the number of inputs and outputs should be as low as possible (i.e. by writing one individual script for every graph that goes into the final document), but that can become very tedious and is not always feasible.
6Unless they are kept in sync with a remote repository, revision control systems operate on local files only and hence do not protect against data loss through hardware failure. Researchers still need to make sure that they backup their working files as well as their revision control repository regularly.

stopped the model from converging or brought about a dramatic change in the estimates further down the line.
But most importantly, using a revision control system introduces another layer of reliability and reproducibility. Modern revision control sytems cannot just easily revert unwanted changes to one’s project files, they can effortlessly maintain an arbitrary large number of timelines (“branches”) for a project directory. This is great tool for testing code and ideas: One can easily try out a variety of operationalisations, model specifications, or graphical styles in various branches, once more recording all the changes made to the files, then switch back to a more stable line of development that represents the current state of the analysis and selectively copy over anything that worked. Revision control sytems are based on the assumption that each of these changes should be documented in a comment and so strongly encourage the analysts to keep a log of the rationale behind the myriad of tiny decisions they take while analysing their data and presenting their findings.
Like many other tools discussed in this chapter, revision control systems have been used by computer programmers for decades. Their modern incarnations are designed to deal with millions of lines of code spread across hundreds of files, on which large teams of developers may work concurrently. Psephologists may well think that a system like Git (which is relatively difficult to learn but can be tamed through a number of GUIs) is ridiculously overpowered for their needs. However, experimenting with one’s code and data in a safe environment where each change, each experiment is documented and can be reverted, modified and even re-applied at any later point is ultimately much more rational, rewarding, and productive than the common practice of endlessly commenting in and out lines of code, or creating lots of increasingly cryptically named scripts whose exact purpose we cannot remember after a couple of weeks.

5    The  Internet  as  an  Infrastructure  for  and  as  an Object of Electoral Studies

5.1    Infrastructure

Psephology has been transformed by the availability of large-scale comparative opinion surveys such as the ISSP, the EES, or the Eurobarometer series (see chapter 48). The websites of CESSDA members and other large archives are now the default option for the distribution of these datasets, allowing speedy and cost-efficient proliferation of data, whereas physical media (e.g. DVDs or CD ROMs) have been phased out, unless partiuclarly restrictive usage rules apply.

While the archives are unrivalled when it comes to providing documentation and long-term safe storage for large numbers of datasets, preparing one’s data for a release through the archive system is seen as a chore by many researchers. Thus, there has always been a tradition of more informal data-sharing in psephology with friends and close colleagues. With the advent of the web, individual researchers and small teams began to put their datasets on personal or departmental websites. However, data on such sites is often difficult to find, because there is no central catalogue, and may disappear at any moment, because it is not backed up by a professional infrastructure.
Moreover, data may be stored in any number of formats and without even minimal documentation. The open source Dataverse  project  (http://dataverse.org/)  and some related initiatives aim at solving these problems by providing means for the (semi-)automatic conversion, documentation, versioning, and retrieval of data. They also provide globally unique identifiers and checksums to solve the problem of data integrity. Journals, research groups, and individual scholars can easily create their own repositories to facilitate re-analysis and replication. Dataverse and similar software go a long way towards making data from smaller, often self-funded projects that would otherwise be lost to the scientific community available for secondary analyses. But they still rely on a professional and sustainable IT infrastructure. At the moment, this infrastructure is provided for free by Harvard University and some other global players. Wheter they will continue to provide this service to the community if usage of the system picks up remains to be seen.
Besides traditional data archives and more individual repositories, countless government agencies and other public institutions around the globe have set up websites where they share parts of their records and hence have become data providers. Especially at the subnational level, the main problem with these sites is fragmentation. Even if they would adhere to common standards for their website design and the presentation of their data, finding and navigating hundreds or thousands of individual sites to collect, say, data on candidates in local elections, is obviously inefficient and often infeasible. Thankfully, governments around the world have woken up to the potential value of free access to their data and are implementing open data legislation. As a result, government-sponsored regional, national, or even supra-national “portal” or “data store” sites, which gather and distribute fine-grained data from lower levels, are becoming more prevalent. While these initiatives are often primarily aimed at policy makers and business communities, the social sciences also benefit from the emerging consensus that government data should be open in principle. For psephologists, the growing availability of geo-referenced electoral results and other statistical data (e.g. census or landuse data) is of particular importance.
In an ideal world, websites would offer the exact data set required by a researcher in

a format that can be read directly into one’s favourite statistical package. In reality, datasets are often offered in the guise of Excel sheets or textfiles that need to be imported. While this is not too problematic, such files are often created “on the fly” from an underlying data base according to some specifications that need to be entered manually. If the same dataset (or different variants and iterations of the same dataset) needs to be downloaded more than a couple of times, it may be worthwhile to do this programmatically by means of a script. Moreover, there still exist (government) websites that present the required data not as a file for download, but rather as a series of formatted tables on the screen, possibly in a paginated format. In these cases, researchers should consider writing a “scraper”, i.e. a small program that simulates the activities of a website user and stores the results as a dataset. While Python has a whole suite of libraries that make it an ideal tool for scraping tasks, some modern packages for the R system offer very similar capabilities from within the statistical package. Munzert et al. (2015) provide an excellent introduction to “scraping” and “mining” the internet. While they focus on R, the techniques and standards they discuss translate easily to workflows that are based on other tools.
Finally, many service providers – amongst them a handful of government agencies
– provide “application programming interfaces” (APIs) to their data. APIs bypass the traditional website altogether. They represent complex and very specific mechanisms for interacting with the underlying database of a service as a series of simple commands for high-level programming languages such as R or Python. Using these commands, scripts can directly access these services without even simulating the activities of a human user of a website. From the point of view of the person writing a script, accessing a service on the internet is not different from calling a function that is hardwired into the respective programming language.
For instance, psephologists may have a variable that contains the addresses of candidates as stated on the ballot papers (a messy combination of names and numbers, possibly with typos). To convert these to proper geographical co-ordinates, they would want to make use of a “geocoding” service. There are APIs for various such services (e.g. Google Maps, Bing Maps, and the OpenStreetMap project), which wrap the necessary low-level instructions into a simple function call. Usage limits and possibly payment options aside, switching from one service to another is usually just a matter of applying slightly different functions to a variable. Using yet another API, the resulting coordinates could then be mapped to census tracts, for which a host of socio-economic and demographic data are available that could provide a rough-and-ready approximation of the respective environment the candidates live in.

5.2    The internet as an object

Since its inception as a research infrastructure, the internet has been thoroughly transformed. While the usual caveats about selective access and use apply, the internet’s role as a political medium is becoming more and more relevant for psephologists. Current research is very much focussed on political communication as it happens on social networking platforms, with Facebook, Twitter, and Instagram being the most prominent ones. Scraping these sites with simple scripts would not just violate their terms of use, but is virtually impossible due to their heavy use of interactive web technology, the networked nature of communication on these sites, and the sheer volume of posts. However, once more there are APIs available through which these services can be mined programmatically. While limits apply, analysts will often find the free tiers perfectly suitable for their needs. Moreover, both Twitter and Facebook have strong research departments that are open to forming partnerships with social scientists.
Research into social-networked communication on the internet is currently dominated by computer scientists and linguists, who often operate without any underlying theory of social behaviour. Psephologists interested in this field will have to learn a whole host of techniques and concepts, and will have to link these with their own substantive interests. Grimmer and Stewart (2013) provide a useful introduction to automatic content analysis, whereas Ward, Stovel, and Sacks (2011) give a tour d’horizon of concepts in social network theory that matter for political scientists.
Using the internet for analysing conventional media sources is less problematic in many ways. Although many publishers aim at implementing paywalls to secure their revenue streams, many mainstream outlets still put all or at least most of their content online. Moreover, Google, Microsoft and  other  companies have  created aggregator sites that can be accessed programmatically. Using these sources, psephologists can retrospectively track the development of a given issue during a campaign, or assess the tonality of media reports on a set of candidates.  In short, using the internet and a scripting language, researchers can achieve most of what would have required a host of research assistants and an extensive newspaper archive (or an expensive database subscription) only a few years ago.
The Google-supported Global Data on Events, Location and Tone (GDELT, http:
//www.gdeltproject.org/) database takes this idea one step further. GDELT, which is based on older event databases (Gerner et al., 1994), aims at automatically extracting information on actors and events from newswire reports and making them available on a global scale. The GDELT project is somewhat controversial, because its original founders fell out, and because of worries over the quality of the inferences that are drawn from the raw inputs. However, the project, which has generated enormous interest in the IR community, has great potential for psephology, too.

6    Conclusion

From its very beginnings, electoral research has been a beneficiary, and often a driver of technological and methodological progress in the larger field of Political Science. In recent years, this progress has accelerated: User-friendly software, ever faster computers, and last not least the proliferation of data mean that yesterday’s advanced methods quickly turn into today’s new normal. By and large, this chapter has argued that psephologists should continue to embrace technology in general and the open source and open data revolutions in particular. As the examples in the natural sciences (e.g. biology) show, psephologists can do more, and more reliably, if they think a little bit more like software developers and make use of freely available tool chains that haven been tried and tested for decades in much harsher environments.
There is, however, a flip side. Technology is a valuable tool, but it can be a distraction, too, and psephologists should never lose sight of their core competency: the ability to put singular findings into a larger context, making use of nearly a century of theory-building. The world is full of “data scientists”, who will happily and rapidly analyse electoral data just as they would analyse any other kind of data. Trying to compete with them on a purely technical level would be a hopeless endeavour. As a profession, we can only stand our ground if we can base our own analyses on profound theoretical insights.


Arzheimer, Kai and Jocelyn Evans (2012). “Geolocation and Voting: Candidate-Voter Distance  Effects  on  Party  Choice  in  the  2010  General  Election  in  England”.  In:
Political Geography 31.5, pp. 301–310. doi: 10.1016/j.polgeo.2012.04.006.
—   (2014). “Candidate Geolocation and Voter Choice in the 2013 English County Council Elections”.  In:  Research  &  Politics.  doi:  10.1177/2053168014538769.
Cameron, Debra et al. (2005). Learning GNU Emacs. A Guide to the World’s Most Extensible Customizable Editor. 3rd ed. Sebastopol: O.
Campbell, Angus et al. (1960). The American Voter. New York: John Wiley.
Coleman, James S. (1994). Foundations of Social Theory. Cambridge, London: The Belknap Press of Harvard University Press.
Crawley, Michael J. (2013). The R Book. Chichester: Wiley.
Eurostat (2015). Nomenclature of Territorial Units for Statistics NUTS 2013/EU-28. Regions in the European Union. Luxembourg: Publications Office of the European Union, url:
http://ec.europa.eu/eurostat/documents/3859598/6948381/KS- GQ- 14- 006-EN-N.pdf/b9ba3339-b121-4775-9991-d88e807628e3.

Gelman, Andrew and Jennifer Hill (2007). Data Analysis Using Regression and Multilevel/Hierarchical Models. Cambridge: Cambridge University Press.
Gerner, Deborah J et al. (1994). “Machine Coding of Event Data Using Regional and International Sources”. In: International Studies Quarterly 38.1, pp. 91–119.
Grimmer, Justin and Brandon M. Stewart (2013). “Text as Data. The Promise and Pitfalls of Automatic Content Analysis Methods for Political Texts”. In: Political Analysis
21.3, pp. 267–297. doi: 10.1093/pan/mps028.
Hox, Joop J. (2010). Multilevel Analysis. Techniques and Applications. 2nd ed. New York: Routledge.
Huckfeldt, Robert and John Sprague (1987). “Networks in Context: The Social Flow of Political Information”. In: The American Political Science Review 81.4, pp. 1197–1216.
url: http://www.jstor.org/stable/1962585.
Ihaka, Ross and Robert Gentleman (1996). “R: A Language for Data Analysis and Graphics”. In: Journal of Computational and Graphical Statistics 5.3, pp. 299–314. doi:
Jackman, Simon (2004). “Bayesian Analysis for Political Research”. In: Annual Review of Political Science 7, pp. 483–505.
Jennings, M. Kent (2007). “Political Socialization”. In: The Oxford Handbook of Political
Behaviour. Ed. by Russell J. Dalton and Hans-Dieter Klingemann. Oxford: Oxford University  Press,  pp.   29–45.  doi:   10.1093/oxfordhb/9780199270125.003.0002.
Jöreskog, Karl G. (1990). “New Developments in LISREL: Analysis of Ordinal Variables Using Polychoric Correlations and Weighted Least Squares”. In: Quality and Quantity
24.4, pp. 387–404. doi: 10.1007/BF00152012.
— (1994). “On the Estimation of Polychoric Correlations and Their Asymptotic Covariance Matrix”. In: Psychometrika 59.3, pp. 381–389.
Karvonen, Lauri and Jostein Ryssevik (2001). “How Bright was the future? The Study of Parties, Cleavages and Voters in the Age of the Technological Revolution”. In: Party Systems and Voter Alignments Revisited. Ed. by Lauri Karvonen and Stein Kuhnle. London: Routledge, pp. 45–61.
King, Gary (1997). A Solution to the Ecological Inference Problem. Reconstructing Individual Behavior from Aggregate Data. Princeton: Princeton University Press.
Knoke, David and Song Yang (2008). Social Network Analysis. 2nd ed. Thousand Oaks: Sage.
Knuth, Donald Ervin (1984). “Literate Programming”. In: The Computer Journal 27.2, pp. 97–111.
Lakhani, Karim R. and Eric von Hippel (2003). “How Open Source Software Works: “‘”Free”’  User-to-User  Assistance”.  In:  Research  Policy  32.6,  pp.  923–943.  doi:  10 .

Lazarsfeld, Paul F., Bernard Berelson, and Hazel Gaudet (1944). The People’s Choice. How the Voter Makes up His Mind in a Presidential Campaign. Chicago: Columbia University Press.
Lazarsfeld, Paul F. and Herbert Menzel (1961). “On the Relation Between Individual and Collective Properties”. In: Complex Organizations. A Sociological Reader. Ed. by Amitai Etzioni. New York: Holt, Rinehart & Winston, pp. 422–440.
Loeliger,  Jon  and  Matthew  McCullough  (2012).  Version  Control  with  Git.  2nd  ed.
Sebastopol: O’Reilly.
Long, J. Scott (2009). The Workflow of Data Analysis. Principles and Practice. College Station: Stata Press.
Lutz, Mark (2013). Learning Python. 5th ed. Sebastopol: O’Reilly.
McClurg,  Scott  D.  (2006).  “The  Electoral  Relevance  of  Political  Talk:  Examining
Disagreement and Expertise Effects in Social Networks on Political Participation”. In: American Journal of Political Science 50.3, pp. 737–754. url: http://www.jstor.
Mecklenburg, Robert (2005). Managing Projects with GNU Make. Sebastopol: O’Reilly. Miller, Warren E. and J. Merrill Shanks (1996). The New American Voter. Cambridge,
London: Harvard University Press.
Monogan, James E. III (2015). “Research Preregistration in Political Science. The Case, Counterarguments, and a Response to Critiques”. In: PS: Political Science & Politics
48 (3), pp. 425–429. doi: 10.1017/S1049096515000189.
Munzert, Simon et al. (2015). Automated Data Collection with R. A Practical Guide to Web Scraping and Text Mining. Chichester: Wiley.
Muthén, Bengt O. (1979). “A Structural Probit Model with Latent Variables”. In: Journal of the American Statistical Association 74, pp. 807–811.
— (2002). “Beyond SEM. General Latent Variable Modeling”. In: Behaviormetrika 29, pp.  81–117.  url:  /home/kai/Work/Texte/Muthen2002.pdf.
Olson, Mancur (1965). The Logic of Collective Action. Public Goods and the Theory of Groups. Cambridge: Harvard University Press.
Raymond, Eric S. (1999). The Cathedral and the Bazaar. Musings on Linux and Open Source by an Accidental Revolutionary. Beijing et al.: O’Reilly.
Siegfried, André (1913). Tableau politique de la France de l’Ouest sous la Troisieme Republique. Paris: A. Colin.
Skrondal, Anders and Sophia Rabe-Hesketh (2004). Generalized Latent Variable Modeling.
Boca Raton u.a.: Chapman & Hall.
Steenbergen, Marco R. and Bradford S. Jones (2002). “Modelling Multilevel Data Structures”. In: American Journal of Political Science 46, pp. 218–237.

Stegmueller,   Daniel   (2013).   “How   Many   Countries   for   Multilevel   Modeling?   A
Comparison of Frequentist and Bayesian Approaches”. In: American Journal of Political Science. doi: 10.1111/ajps.12001.
Ward,  Michael  D.  and  Kristian  Skrede  Gleditsch  (2008).  Spatial  Regression  Models.
Quantitative Applications in the Social Sciences 155. Thousand Oaks: Sage.
Ward, Michael D., Katherine Stovel, and Audrey Sacks (2011). “Network Analysis and Political Science”. In: Annual Review of Political Science 14.1, pp. 245–264. doi: 10.
Western,  Bruce  and  Simon  Jackman  (1994).  “Bayesian  Inference  for  Comparative Research”. In: American Political Science Review 88, pp. 412–423.

Another Dog that didn’t Bark? Less Dealignment and More Partisanship in the 2013 Bundestag Election


1 Introduction

For the last twenty-five years or so, party identification has been said to be in decline in Germany. And yet, those two parties which are most closely associated with traditional concepts of partisanship, i.e. the Christian Democrats (CDU/CSU) on the right and the Social Democrats (SPD) on the left – are once more jointly governing Germany, with the CDU/CSU coming tantalisingly close to an outright majority in parliament. This paper tries to shed some light by re-visiting the major stations of the debate before considering new longitudinal data and finally turning to the 2013 Bundestag election.

2 The controversy over partisan dealignment in Germany

The question whether Michigan-style identifications do exist in West Europe, where politics was shaped along the lines of ideologies and cleavages, was hotly debated in the 1970s (see Dalton, Flanagan, and Beck, 1984 for a useful summary). However, towards the end of the decade a consensus emerged that the concept could indeed be transplanted to the polities on the old continent including Germany, conditional on an operationalisation that caters for multi-party systems (Falter, 1977). Such an operationalisation has been employed since the first Politbarometer surveys (dating back to the late 1970s) and has been replicated in Germany’s general social survey (ALLBUS), in the national election studies, and in countless other opinion surveys.

Yet, the late 1970s may very well have marked the height of partisanship in Germany. Mutually re-enforcing processes of socio-economic modernisation, secularisation, and value-change began to undermine the cleavage base of the German party system, which in turn facilitated the rise of the Green party in the 1980s. Moreover, according to one very influential account (Dalton, 1984), the expansion of higher education and the increase in the availability of political information reduced the heuristic value of party identification as a device that reduces cognitive costs.

The political crises of the 1980s and early 1990s, on the other hand, had very little effect on levels of party identification in Germany: The decline in partisanship was never sudden but rather glacial and concentrated in those social groups whose loyalties have shaped the modern German party system: working class voters, catholics, and churchgoers more generally (Arzheimer, 2006).

More recently, Dassonneville, Hooghe, and Vanhoutte (2012) have argued that the decline in partisanship has accelerated and is now most prevalent amongst voters with low levels of formal education, which could in the long run lead to an underrepresentation of vulnerable socio-economic groups in the German party system. Moreover, a positive correlation between formal education on the one hand and party identification on the other goes against the grain of Dalton’s original argument about cognitive mobilisation and dealignment (see also Albright, 2009 and Dalton, 2014).

Nevertheless, unlike many other studies on dealignment in Germany (but see Schmitt-Beck and Weick, 2001 and Arzheimer and Schoen, 2005) Dassonneville, Hooghe, and Vanhoutte’s work is based on the Socio-Economic Panel (SOEP), an annual survey of more than 12,000 households that has been running since 1984. While the SOEP provides unrivalled insights into the individual dynamics of partisanship, it also suffers from a number of drawbacks. First and foremost, after three decades in the field, panel mortality is a serious issue. While the SOEP team claims that they can compensate for attrition by recruiting new households, the structure of the data set and the attached weights have become unwieldy to say the least. Second, the research agenda of the SOEP is primarily driven by economists. Its questionnaire contains very items with genuinely political content and therefore lacks the priming context that is provided by ordinary opinion surveys. Finally, field work for the SOEP is usually drawn out over a lengthy period of time, whereas polling for other surveys that are used to study partisanship is either continuous or focused on campaigns, i.e. periods of intense political mobilisation.

While none of these issues rule out the SOEP as a valuable data source for analysing dealignment in general and issues of attitude stability at the micro level in particular, the SOEP is less than ideally suited for plotting the long-term levels of partisanship in Germany, or its importance in any given election. Therefore, the next section will rely on the monthly Politbarometer survey series to chart the decline of partisanship, while the penultimate section will make use of the German Longitudinal Election Study (GLES) to assess the relevance of party identification for voters in the 2013 Bundestag election.

3 Is partisanship in (Western)Germany in decline?

Forschungsgruppe Wahlen have been tracking German political attitudes with their monthly Politbarometer surveys since the golden age of party identification in the late 1970s. The Politbarometer follows a classic repeated cross-sectional survey design, where each group of interviewees is sampled independently and thought to be representative for the German population in the respective year and month.

Although Forschungsgruppe is a commercial operation, their raw data are made available for secondary analysis after an embargo of two to three years. Previous analyses of these data for the 1977-2002 period have shown that in line with theories of secular dealignment, party identification in Western Germany declines fairly slowly and steadily at a rate of less than one percentage point per year (Arzheimer, 2006).

Since then, Forschungsgruppe has released ten years’ worth of new data, which cover the upheaval caused by the ‘Agenda 2010’ following the 2002 election and the onset of the second Grand Coalition (2005) as well as the merger between the Eastern PDS and the Western WASG (2007) and the short but meteoric rise of the FDP (2009).


Figure 1: Partisanship in West Germany, 1977-2012

Source: own calculation based on Politbarometer series, ZA2391

The series is rather noisy with a standard deviation of 5.4 percentage points. This is to be expected, as sampling error alone should result in a standard deviation of roughly 1.5 percentage points, disregarding any additional error due to multistage sampling. Even after applying a moving average smoother using a five-month (2 1 2) window, the series is rather jittery (see Figure 1), with some of the noise probably being the result of campaign effects (the diamond-shaped symbols mark the dates of federal elections). However, it also seems clear that the downward trend of the 1980s and 1990s has slowed down considerably in the new millenium, with the average yearly attrition rate falling well below 0.5 percentage points.

As the micro data are readily available, it is possible to model the decline in partisanship directly without resorting to the aggregated time series (see Arzheimer, 2006). A simple descriptive model would start with a logistic regression of holding a party id (a dichotomous variable) on calendar time, controlling for campaign effects. For simplicity’s sake, only federal elections and Land elections in Bavaria, Baden-Wurttemberg, and North Rhine-Westphalia – the three most populous states which are collectively home to more than half of the West German population — were considered, and campaigns were assumed to uniformly run for three months, including the month in which the election was held. Logistic regression enforces an S-shaped link between partisanship and its predictors, which given the empirical distribution of party identifications in the sample (between 59 and 84 per cent) will result in a nearly linear relationship. To accommodate the apparent non-linear decline of partisanship, following Royston and Sauerbrei (2008) a number of fractional polynomial transformations of calendar time were included in a bivariate model (not shown), with an additional square root transform providing the best fit.

Since the purpose of the model is descriptive, only two variables were included to account for changes in the composition of the population that occurred over the 35-year period: Formal education (people who were educated beyond Mittlere Reife vs. everyone else), and age. As outlined in section 2, formal education is interesting in itself, but it also serves as a useful proxy for not belonging to the working class and not attending church frequently, rendering a durable affiliation with either the SPD or the CDU/CSU much less likely.

Age, or rather the time at which person was born will affect partisanship in two ways. On the one hand, partisanship is partly a habit, which is reinforced over the course of one’s life (Converse, 1969). Therefore, older voters should be more likely to identify with a party. On the other hand, dealignment theory suggests that independent of individual age and across the span of their lives, members of younger cohorts are less likely to identify with a party compared to those who were socialised into the largely stable German party system of the 1960s and 1970s.

Life cycle and cohort effects are notoriously difficult to separate (Oppenheim Mason et al., 1973). Because age is only recorded in a categorised fashion in the Politbarometer surveys anyway, no such attempt was made. Instead, respondents were split into three broad categories (under 35, 35 to 60, and over 60) to control for the slow but momentous demographic changes Germany is undergoing. Finally, the effects of age and education were allowed to vary over time to account for generational replacement and the new relationship between education and partisanship postulated by Dassonneville, Hooghe, and Vanhoutte (2012).

Although the additional complexity introduced by the interaction terms is a setback, model comparisons (not shown) based on the Bayesian Information Criterion (BIC) demonstrate that such a fully interactive model fits the data much better than either a non-interactive variant or a model that regresses partisanship on calendar time and campaign effects alone.


Figure 2: Estimated overall levels of partisanship in West Germany, 1977-2002 (adjusted predictions at representative values (APR))

Source: own calculation based on Politbarometer series, ZA2391. Predictions derived from parameter estimates shown in Table 1.

Party ID
Campaign (all)0.0400∗
Age: 35-59-2.923∗∗∗
Age: 60--3.117∗∗∗
Educ: high0.0941
Age: 35-59 × Sqrt(Time)0.317∗∗∗
Age: 60- × Sqrt(Time)0.299∗∗∗
Age: 35-59 × Time-0.00747∗∗∗
Age: 60- × Time-0.00579∗∗∗
Educ: high × Sqrt(Time)-0.0210
Educ: high × Time0.00134


Table 1: Micro Model of Partisanship in West Germany, 1977-2012

Source: own calculation based on Politbarometer series, ZA2391.

Table 1 shows the results. However, since the substantive meaning of logit coefficients is hard to grasp, particularly in the face of additional non-linearities and interactions, the interpretation will focus on a graphical representation. Figure 2 shows that the decline of partisanship has slowed down considerably indeed. In theory, anything could have happened in the nine months between the current end of the time series and the election, but the graph makes it abundantly clear that dealignment has effectively halted during the last decade under study. The estimated attrition rate for the five-year period from December 2007 to December 2012 is a mere 0.8 percentage points, just over the estimated yearly average for the 1980s.


Figure 3: Estimated levels of partisanship in West Germany by formal education, 1977-2002 (adjusted predictions at representative values (APR))

Source: own calculation based on Politbarometer series, ZA2391. Predictions derived from parameter estimates shown in Table 1.

Including education, age, and their interaction with time in the model makes it possible to look into group-specific trends in dealignment. Figure 3 shows that partisanship has fallen much more rapidly amongst those with higher formal qualifications, leading to a gap that has become increasingly wider in recent years, as dealignment has essentially petered out amongst those with higher levels of educational attainment. Yet, dealignment has slowed down for the lower attainment group, too: The change from e.g. 2000 to 2010 is much less dramatic than the development for the 1990 to 2000 period, hinting once more at stabilisation on a lower level.


Figure 4: Estimated levels of partisanship in West Germany by age group, 1977-2002 (adjusted predictions at representative values (APR))

Source: own calculation based on Politbarometer series, ZA2391. Predictions derived from parameter estimates shown in Table 1.

One intriguing aspect of this pattern is that levels of formal education are negatively correlated with age as a result of the ongoing expansion of education. Figure 4 offers a more direct look into the age-specific trajectories of dealignment. One first insight is that – at least according to the underlying model – age did not matter much in the late 1970s and early 1980s but quickly became a factor over the course of this decade as younger respondents were increasingly less likely than their older compatriots to report an identification with a party. Relevant segments of the new cohorts entering the political system either never acquired such an identification or did not retain it at the same rate as their predecessors. Given how steep the estimated decline of their partisanship is compared to the other groups, it seems safe to assume that the dealignment of the 1980s and mid-1990s that reduced the number of partisans by nearly a quarter must have been driven largely by this group.

However, once more the estimated attrition rate in this group began fall appreciably around the turn of the century. Moreover, nearly everyone who belonged to this group in the 1980s had now moved on to the next age band, which exhibits a nearly linear pattern of decline that is currently steeper than that of the youngest group, although levels of partisanship are still noticeably higher.

Finally, the over sixties, who began at roughly the same level as the middle age group, did outstrip them in terms of partisans by the mid-1990s. Levels of partisanship have been essentially stable in this group for more than a decade now. Once more one must keep in mind that by the early 2000s, everyone who was in the middle group in the 1980s had moved on to this upper age band.

Demographic changes that the mean age of people belonging to an age group will somewhat fluctuate over time: From the 1940s until the mid-1960s, almost every birth cohort was bigger than the one before, but since then, this pattern has been reversed. Yet, even accounting for this effect and for the rising life expectancy, the changes in the impact of age on party identification are too big to be the result of stable life cycle effects. They point either at massive shift in what it means for partisanship to be young, middle-aged, or old, or, equivalently, at substantial cohort effects.

One final aspect that must be considered is the relative size of the three age groups. During the first five years of polling, 29 per cent of all respondents were under 35, while 26 per cent of those interviewed were older than 60. For the 2008-2012 period, this balance has been reversed. The share of older citizens has risen to just under 30 per cent, and only 18 per cent of all respondents are younger than 35. Voters aged 35 to 59 currently make up 52 per cent of the sample, but their share is now peaking, while the oldest group is rapidly growing and already stands at 33 per cent in the 2012 data. In essence, this means that dealignment in Germany is slowed down by demographic change, because the combined shares of middle aged and older voters, who are more likely to be partisans, is growing.

Either way, party identification has neither collapsed nor withered away in West Germany. Assessing the state and trajectory of party identification in the former East Germany is less straightforward. First, theories of dealignment do not apply because there should not have been any alignment in the first place. After all, Easterners had not been exposed to the West German party system before 1990 and, more generally, had had no experience with free elections since the (partially free) Land elections of 1946. While it has been argued that many Easterners had access to West German TV and hence could form “quasi-attachments” to West German parties (Bluck and Kreikenbom, 1991), these attachments can hardly have been comparable to Michigan-type identifications. After all, the latter are the result of socialisation effects in the family and intermediary associations, exposure to fellow partisans, party members and party communication, first-hand experience of policies and policy outcomes, and last not least the habit-forming experience of repeatedly voting for one’s party. Accordingly, the number of self-reported partisans in the East was lower than in the West all through the 1990s, while attachments were weaker and less stable.


Figure 5: Partisanship in East Germany, 1991-2012

Source: own calculation based on various Politbarometer samples

Second, the East German subsamples of the Politbarometer poll are often relatively small. Until 1995, East Germans were massively overrepresented in the polls: Essentially, Easterners were sampled separately and in numbers approaching those for West Germany (roughly 1000 per month and region) to account for the idiosyncratic and very fluent nature of public opinion in the post-unification East. From 1996 to 1998, Forschungsgruppe used a single sampling frame, interviewing about 1000 respondents per month in total. In 1999, Forschungsgruppe reinstated separate regional subsamples of roughly equal size, but from the early 2000s on, they considerably reduced Eastern sample sizes for most months, boosting it occasionally to cover election campaigns. As a result, the Eastern time series is very noisy even after applying the moving average smoother (Figure 5).

Despite these fluctuations, it is clear that the massive decline of self-reported identifications in the early 1990s was a temporary phenomenon. From the mid-1990s on, the number of identifiers moved up, although in fits and starts. This pattern is at least compatible with a process of social-political learning, during which East Germans became familiar with the party system and wider liberal-democratic political system. Then, for the last decade or so, levels of partisanship in East Germany have been by and large stable in the 55-to-65 per cent range, roughly five percentage points below West German levels.

Given the relatively small East German sample sizes (particularly for younger and highly educated voters), the comparatively short time series, and the absence of any clear trends, I refrain from modelling developments in subgroups. At this stage, the more important point to note is that partisanship was clearly still an important at the time of the 2013 election. While the group of non-partisans is large, in both regions, more than half of the voters report a party identification, and there is no sign of a sudden and imminent decline.

4 The role of party identification in the 2013 election

4.1 Party identification and party choice

Just because respondents report identifications, they need not necessarily be politically meaningful. In this section, a simple model of voting in the 2013 election is presented in order to assess the political relevance of party identification.

Modelling electoral choice in multi-party systems is not entirely straightforward. Perhaps the most commonly employed statistical model is the multinomial logit (MNL). One problem of the MNL, however, is the large number of parameters which must be estimated, because each possible outcome (minus a reference category) is given its own set of coefficients: For k parties and l variables, the total number of parameters is (k − 1) × (l + 1). Even if CSU voters are lumped together with voters of the CDU, and non-voters and voters of “other” parties are disregarded, there were are at least five relevant choices (Christian Democrats, SPD, FDP, Greens, and the Left) that need to be considered, so that even simple models become unwieldy very quickly.

Fortunately, there is another option. The Conditional Logit Model (CLM, Alvarez and Nagler, 1998) has only a single parameter for the effects of each variable that varies across alternatives within voters. This includes many variables which are deemed to affect electoral behaviour: evaluations of candidates, policies, and parties. The CLM resembles the MNL in that it can be extended to also incorporate variables that are constant across alternatives (Long and Freese, 2006, p. 307), like more general attitudes, or socio-demographic variables, but for these, the number of parameters is once more proportional to k − 1.


Evaluation: Candidate0.555∗∗∗0.625∗∗∗
Ideolocal Distance-0.374∗∗∗-0.423∗∗∗
Union × Tax vs Welfare-0.00368-0.121
FDP × Tax vs Welfare0.259∗0.220
B90Gruene × Tax vs Welfare-0.01180.224
Left × Tax vs Welfare-0.0122-0.0614
Union × Immigration-0.0750-0.0729
FDP × Immigration-0.0658-0.176
B90Gruene × Immigration-0.124-0.260
Left × Immigration-0.151-0.379∗∗


Table 2: Micro Model of Electoral Choice in the 2013 Bundestag Election (East vs. West)

Source: own calculation based on GLES 2013 pre-election cross-section, ZA5700. “Observations” are observed choices. The number of cases is 888 for the West and 206 for the East. Standard errors take into account the nesting of choices within electors and the complex survey design, including the weights supplied by the GLES team.

Table 2 shows the estimates for the parameters of a very simple conditional logistic model of electoral choice in the 2013 election. Data come from the pre-election cross-sectional survey component of the German Longitudinal Election Study (GLES). The model itself is built around the Michigan triad of party identification, candidate evaluations, and issue considerations. The latter are operationalised in multiple ways. For the “ideological distance” measure, respondents were asked to place themselves and the main parties on a standard left-right scale to gauge the general agreement between voters’ preferences and the parties’ policy proposals. To get a more rounded impression of the impact of policy considerations, preferences on two more specific positional issues that were deemed to be important in the 2013 election were included as well: lower taxes vs. more welfare spending, and immigration.1

While respondents were asked for their perceptions of party positions on these issues so that alternative-specific measures of distance could be calculated, the number of missing values for these items is quite high. Hence, only voters personal preferences regarding immigration and tax/welfare enter the model. Including such case-specific variables in a CLM of electoral choice requires one to include a series of party specific constants and interaction terms (Long and Freese, 2006, p. 305), which pick up the effect of a change in the case-specific variables on the chance of choosing the respective party vs. some arbitrary baseline alternative (in this case, the SPD).

To account for any differences between East and West Germany, parameters were estimated separately for both regions.2 While the interpretation is slightly complicated by the presence of multiple interaction terms, it is clear from Table 2 that such differences played a role in the 2013 election. To see why this is the case, consider a voter who is both in favour of raising welfare spending (0) and facilitating immigration (0). For these persons, all interaction terms drop out of the equation so that the constant reflects the odds of voting for the respective party vs. voting for the SPD. In the West, the odds seem to favour the Left (e0.528≈ 1.7), but the coefficient is not statistically different from zero. In the East, however, the Left’s advantage is significant, and massive (e2.077≈ 8). Even for Eastern voters who hold a more centrist position (5) on the immigration scale, the Left will be slightly more attractive, ceteris paribus, whereas in the West, the balance is tilting towards the SPD.

While these differences are certainly interesting, the main concern of this section is the role of party identification. From the first line of Table 2, it can be gleaned that in both regions, identifying with a party has a very strong effect on the odds of actually voting for this party even after controlling for specific issue positions, general ideological distance, and candidate evaluations.

The latter two do certainly matter, too. Because of the range of the underlying scales (0-10 and 1-11, respectively), their potential effect is even bigger than that of party identification. But in practice, the perceived ideological distances between voters and parties are relatively small, with a median of 2 points and a mean of 2.3. Candidate evaluations display more variation with a mean of 6.2 and a median of 6, implying that a plausible candidate could possibly compensate for a lack of attachment to the party.

Yet, one should bear in mind that for candidate evaluations (and ideological distances), only the differential is relevant, because all candidates will appeal to some degree. If a voter likes or dislikes all candidates in equal measure, their joint effect on her voting behaviour is nil. For the average voter, the standard deviation of candidate evaluations is just 1.9 points, suggesting that in many cases, the differential and hence the candidate effect will be considerably smaller than the potential effect. Having a party identification, on the other hand, will be definition benefit only a single party, to whom the maximal potential effect will apply.

One intuitive (though potentially problematic, see Long and Freese, 2006, p. 111) approach towards assessing the relevance of party identifications is to compare actual electoral choices to those expected given the data and the parameter estimates. In both areas, about 85 per cent of voters are classified correctly.3 However, simply assuming that those who hold an identification will vote in accordance with it works just as well, with a 85 per cent of the subgroup correctly classified in the West and 92 per cent in East. Accordingly, the match between party identification and model-derived predictions is almost perfect (98 per cent) for identifiers.

This shows that at least in this election, candidate evaluations and policy concerns were rarely able to offset the effect of longstanding loyalties amongst those who have an identification and turned out to vote. Nonetheless, they will shape voting decisions amongst the slowly growing group of those who do not identify with a party.

4.2 The importance of being left: Ideology, party identification and choice amongst left parties

In German Politics, one of the most interesting developments in recent years has been the breakaway of the WASG from the SPD following the enactment of the “Agenda 2010” reforms, and the ensuing PDS/WASG merger (Hough, Koß, and Olsen, 2007). As a result, the left camp is now more fragmented than the right, at least for the time being. Moreover, the (ongoing) conflict over the “Agenda” and its legacy has re-asserted the importance of distributional issues (which were over-shadowed by moral questions, at least in many academic analyses) for party competition.

The question of whether this new divide within the left camp has already become entrenched in the guise of (new) party identifications has rarely been addressed. After all, it is not implausible that the vote for the Left (particularly in the West) could be driven by policy concerns alone or even by more generalised “protest”.

Yet, the short answer to the question is that this does not seem to be the case. Admittedly, voters of the Left party position themselves significantly closer to the left end of the political spectrum than voters of the SPD or the Greens. This even holds when the analysis is restricted to the subsample of voters who self-identify as leftists by reporting position on the continuum that is clearly left of the centre (4 or less). Moreover, voters of the Greens are slightly more in favour of immigration than voters of the other two parties. Again, this holds for both regions, and for the general population and the leftist subsample (not shown as a table).


SPD × East1.606∗
B90Gruene × East1.987∗
Left × East1.296


Table 3: Leftist Voters’ Positions on Taxes/Welfare Spending as a Function of Party Choice and Region

Source: own calculation based on GLES 2013 pre-election cross-section, ZA5700. The size of the subpopulation is 333. Standard errors take into account the complex survey design, including the weights supplied by the GLES team.


no/other × West7.003(0.400)
no/other × East4.564(0.436)
SPD × West4.474(0.332)
SPD × East3.641(0.280)
B90Gruene × West4.137(0.315)
B90Gruene × East3.685(0.525)
Left × West4.588(0.540)
Left × East3.444(0.349)


Table 4: Leftist Voters’ Positions on Taxes/Welfare Spending (Adjusted Predictions at Representative Values)

Source: own calculation based on GLES 2013 pre-election cross-section, ZA5700. Adjusted predictions derived from model presented in Table 3. The size of the subpopulation is 333. Standard errors take into account the complex survey design, including the weights supplied by the GLES team.

But on the crucial tax/spending issue, there are hardly any differences between the supporters of the three parties. Here, the real difference is that between Easterners and Westerners, and this gap is particularly pronounced amongst those who consider themselves to be left-wing. Table 4 lists the adjusted predictions derived from a simple linear model (Table 3) that regresses tax/spending preferences amongst leftist (self-placement on scale points 1-4) voters on region and electoral choice. Lines 1-4 shows national estimates by party choice. Clearly, the differences between the respective supporters of the SPD, the Greens, and the Left are small and statistically insignificant, whereas any other voters position themselves more than two points closer to the “lower taxes” pole of the scale on average.

Perhaps even more striking are the estimates for the overall difference between East Germans and West Germans given in the next two lines. Although all respondents in this subsample consider themselves to be on the left, Western respondents lean slightly towards the “lower taxes/fewer benefits” pole of the continuum. Eastern respondents, on the other hand, position themselves 1,6 points closer to the “higher taxes/more benefits” pole.

The rest of the table breaks down the preferences of leftist along party lines and region. Because of the small sample sizes, the regional differences within electorates are not statistically significant, but the clearly show that within each region, the voters of the three parties hold broadly similar views on taxation and welfare.




N 1282

Table 5: Party identification of leftist voters in West Germany by vote choice

Source: own calculation based on GLES 2013 pre-election cross-section, ZA5700. The size of the subpopulation is 254. Standard errors take into account the complex survey design, including the weights supplied by the GLES team.

While policies seem hence to matter less than one would have expected, party identification once more plays a prominent role. Table 5 shows the party affiliation of Western leftist voters by electoral choice. From the main diagonal, it can be seen that between 72 and 85 per cent report a party identification that is congruent with their electoral choice. Crucially, this also holds for the Left party, which is still relatively new by West German standards. Here, 73 per cent of the voters claim to be longstanding supporters. Although the sampling error is relatively large for this small group, one can be confident that more than half of the Left’s Western voters are identifiers.

In the East, the results are virtually identical (not shown as a table). Flipping the perspective demonstrates that similarly high numbers of identifiers vote for the “correct” party, and again, this holds for both regions (not shown as a table). Taken together, these findings suggest that the fragmentation of the left electorate has indeed become entrenched. Obviously, this does not bode well for any attempts of the SPD to win (back) voters from the Left.

5 Conclusion: Party identification in Germany: not Dead yet

The notion of party decline in Western countries is as old as the post-war political order (Reiter, 1989). But at least for the old Federal Republic, and then for the Western states during the first decade after unification, there is no evidence of any sudden collapse of the party loyalties. Instead, the available data from the Politbarometer series point to an almost glacial process of dealignment that is driven by social and generational change (Arzheimer, 2006).

This article expands on earlier contributions by first extending the study of the Politbarometer series by a full decade to the whole 1977-2012 period. The most important finding from this analysis is that dealignment in Western Germany has slowed down even further, coming to a virtual halt in recent years.

One reason for this is the emerging positive relationship between formal education and partisanship, coupled with the ongoing expansion of the German education system. This positive effect of education (which confirms some of Dassonneville, Hooghe, and Vanhoutte (2012)’s finding using a less idiosyncratic data base) is both unexpected and remarkable, because it contradicts classic cleavage theory as well as the original argument about cognitive mobilisation. Whether it hails a new age of “cognitive partisans” (Dalton, 2014, p. 140) remains to be seen, although the results are certainly suggestive.

Demographic changes play an important part, too. While it is not quite clear whether this is primarily a result of life-cycle or of cohort effects, late-middle-aged voters and younger pensioners are more likely to be partisans than younger voters, whose share of the electorate is rapidly shrinking.

Turning from the longitudinal to a cross-sectional perspective, it could further be demonstrated that in both East and West Germany, party identifications are a very strong predictor of voting intentions, even if the other elements of the Ann-Arbor-Model – candidate evaluations and issues orientations – are controlled for in various ways. Those voters who identify with a party rarely report diverging voting intentions so that issues and candidates matter almost exclusively for the apartisans.

Although the analysis was restricted to the pre-election survey to avoid any post-hoc rationalisations on behalf of the respondents, the spectre of endogeneity obviously looms large in any such model. After all, it is reasonable to assume that at least some respondents cannot distinguish between their current voting intentions and any long-term loyalties they may or may not harbour. However, measures of candidate evaluations and issue orientations are equally or even more so prone to contamination by voting intentions. Therefore, the estimate for the relative importance of party identification should be unaffected even if the absolute size of its effect may be overstated.

Finally, a detailed analysis of leftist voters interviewed for the GLES showed that even in the (small) subgroup of Western voters of the Left party, most respondents claimed to be identifiers. Again, this is a significant and largely unexpected finding. The formation of the WASG and ultimately the WASG/PDS merger were triggered by the SPD’s shift to the right on social and economic policy, yet the leftists amongst the voters of the SPD and of the Left take broadly similar positions on these issues while claiming to identify with their respective parties. This suggests that the fragmentation of the left camp has become entrenched and cannot be easily overcome by another programmatic shift of the SPD.


Albright, Jeremy J. (2009). “Does Political Knowledge Erode Party Attachments? A Review of the Cognitive Mobilization Thesis”. In: Electoral Studies 28.2, pp. 248–260. DOI: 10.1016/j.electstud.2009. 01.001.

Alvarez, R. Michael and Jonathan Nagler (1998). “When Politics and Models Collide. Estimating Models of Multiparty Elections”. In: American Journal of Political Science 42, pp. 55–96.

Arzheimer, Kai (2006). “’Dead Men Walking?’ Party Identification in Germany, 1977-2002”. In: Electoral Studies 25, pp. 791–807. DOI: 10. 1016/j.electstud.2006.01.004.

Arzheimer, Kai and Harald Schoen (2005). “Erste Schritte auf kaum erschlossenem Terrain. Zur Stabilität der Parteiidentifikation in Deutschland”. In: Politische Vierteljahresschrift 46, pp. 629–654.

Bluck, Carsten and Henry Kreikenbom (1991). “Die Wähler in der DDR: Nur issue-orientiert oder auch parteigebunden?” In: Zeitschrift für Parlamentsfragen 22, pp. 495–502.

Converse, Philip E. (1969). “Of Time and Partisan Stability”. In: Comparative Political Studies 2, pp. 139–171.

Dalton, Russell J. (1984). “Cognitive Mobilization and Partisan Dealignment in Advanced Industrial Democracies”. In: Journal of Politics 46, pp. 264–284.

— (2014). “Interpreting Partisan Dealignment in Germany”. In: German Politics 23.1-2, pp. 134–144. DOI: 10.1080/09644008.2013.853040.

Dalton, Russell J., Scott C. Flanagan, and Paul Allen Beck, eds. (1984). Electoral Change in Advanced Industrial Democracies: Realignment or Dealignment. Princeton: Princeton University Press.

Dassonneville, Ruth, Marc Hooghe, and Bram Vanhoutte (2012). “Age, Period and Cohort Effects in the Decline of Party Identification in Germany: An Analysis of a Two Decade Panel Study in Germany (1992-2009)”. In: German Politics 2, pp. 209–227.

Falter, Jürgen W. (1977). “Zur Validierung theoretischer Konstrukte – Wissenschaftstheoretische Aspekte des Validierungskonzepts”. In: Zeitschrift für Soziologie 6, pp. 349–369.

Hough, Dan, Michael Koß, and Jonathan Olsen (2007). The Left Party in Contemporary German Politics. Houndmills: Palgrave Macmillan.

Long, J. Scott and Jeremy Freese (2006). Regression Models for Categorical Dependent Variables Using Stata. 2nd ed. College Station: Stata Press.

Oppenheim Mason, Karen et al. (1973). “Some Methodological Issues in Cohort Analysis of Archival Data”. In: American Sociological Review 38, pp. 242–258.

Reiter, Howard L. (1989). “Party Decline in the West. A Skeptic’s View”. In: Journal of Theoretical Politics 1, pp. 325–348.

Royston, Patrick and Willi Sauerbrei (2008). Multivariable Model-building: A Pragmatic Approach to Regression Analysis Based on Fractional Polynomials for Modelling Continuous Variables. Chichester, UK: Wiley.

Schmitt-Beck, Rüdiger and Stefan Weick (2001). “Die dauerhafte Parteiidentifikation – nur noch ein Mythos? Eine Längsschnittanalyse zur Identifikation mit den politischen Parteien in West- und Ostdeutschland”. In: Informationsdienst soziale Indikatoren 26, pp. 1–5.

1Taxes/spending: “And what is your own opinion regarding taxes and social welfare services? 0 – lower taxes, even if this means a reduction in the benefits offered by the social state; 10 – lower taxes, even if this means a reduction in the benefits offered by the social state”. Immigration: “And what is your opinion regarding immigration? 0 – Immigration should be facilitated; 10 – immigration should be restricted”.

2Obviously, it would have been possible to estimate a single model for all of Germany by including appropriate interaction terms, but this would have introduced an additional layer of complexity.

3The correction presented by Long and Freese (2006) yields a slightly lower rate of 72 per cent.

Versöhnen statt spalten? Das Ergebnis der Bundestagswahl 2009 und die Rolle der PDS/Linkspartei in Ost-West-Perspektive


1 Einleitung, Fragestellung, Daten

Der Urnengang vom 27. September 2009 war bereits die sechste gesamtdeutsche Bundestagswahl. Ähnlich wie bei den vorangegangenen Wahlen unterschieden sich auch dieses Mal die Wahlergebnisse in Ost und West sehr deutlich. Diese Unterschiede, vor allem aber die Rolle, die die PDS/Linkspartei bei ihrem Zustandekommen spielte, sind Gegenstand des vorliegenden Beitrages. Dieser gliedert sich in zwei große Teile: In Abschnitt 2 beschreiben und analysieren wir zunächst die Unterschiede im Wahlergebnis auf der Makro-Ebene. Die Abschnitte 3 und 4 beschäftigen sich im Anschluß daran mit den Ursachen für diese Unterschiede. Dabei konzentrieren wir uns vor allem auf die Wahrnehmung und die Wahl der Linkspartei/PDS, der es 19 Jahre nach der Vereinigungswahl von 1990 gelungen zu sein scheint, sich erfolgreich nach Westen auszudehnen.

Im ersten Teil unserer Beitrages stützen wir uns auf die amtlichen Wahlergebnisse, die von der Homepage des Bundeswahlleiters und aus der Regionaldatenbank Genesis bezogen werden können. Im zweiten Teil verwenden wir die kombinierte Vorwahl-/Nachwahl-Komponente der German Longitudinal Election Study (GLES) 2009 in der Version Pre1.3. Die Daten wurden von Juli bis zur Bundestagswahl (Vorwahl-Komponente) bzw. von der Bundestagswahl an bis in den November (Nachwahl-Komponente) erhoben und können unter der ZA-Nummer 5302 vom Datenarchiv der GESIS bezogen werden. Der Einfachheit halber wird diese Komponente der GLES im Text zumeist kurz als „Wahlstudie“ bezeichnet. Die für die Replikation der Ergebnisse benötigten Stata-Files werden über das Dataverse von Kai Arzheimer zur Verfügung gestellt (http://dvn.iq.harvard.edu/dvn/dv/arzheimer).

Autorenversion (Preprint) Finale Version: Kai Arzheimer und Jürgen W. Falter (2013). „Versöhnen statt spalten? Das Ergebnis der Bundestagswahl 2009 und die Rolle der PDS/Linkspartei in Ost-West-Perspektive“. In: Wahlen und Wähler. Analysen aus Anlass der Bundestagswahl 2009. Hrsg. von Bernhard Weßels, Oscar W. Gabriel und Harald Schoen. Wiesbaden: VS Verlag für Sozialwissenschaften, S. 118–150

Wie bei Wahlstudien üblich weicht die Verteilung der berichteten Wahlabsicht bzw. Entscheidung in den GLES- Daten selbst bei Verwendung aller Gewichtungsfaktoren in einigen Aspekten vom amtlichen Endergebnis der Wahl ab. Dies betrifft vor allem den Anteil der Nichtwähler und der sonstigen Parteien, aber auch die Entscheidung für Union und SPD, deren jeweilige Anteilswerte in der Wahlstudie erkennbar über- bzw. unterschätzt werden. Diese Abweichungen erklären sich zum einen aus den bekannten Effekten der sozialen Erwünschtheit und der selektiven Ausfälle von Respondenten, andererseits aber auch aus der Tatsache, daß die Daten über einen langen Zeitraum hinweg erhoben wurden, innerhalb dessen sich unter dem Eindruck des Wahlkampfs und anderer politischer Ereignisse Verhaltensabsichten bzw. sogar die Erinnerung an tatsächliches Verhalten verändern kann (zu den Effekten des Wahlkampfes 2009 siehe Krewel, Schmitt-Beck und Wolsing, 2011).

Für unsere Fragestellung sollte dies jedoch vergleichsweise unproblematisch sein: Zum einen zeigen sich trotz der Abweichungen bezüglich des absoluten Niveaus auch in den GLES-Daten die bekannten Ost-West-Differenzen in der relativen Mobilisierungsleistung der Parteien. Zum anderen geht es uns weniger um eine exakte Prognose bzw. Retrodiktion als vielmehr um die Analyse von Zusammenhängen. Selbst dann, wenn es zu subgruppenspezifischen Ausfällen kommt (beispielsweise weil sich zu wenige politisch desinteressierte Wähler an Umfragen beteiligen), sollten die Schätzungen für Zusammenhänge stabil sein, sofern die Variablen, durch die die Subgruppen definiert werden, im Modell enthalten sind (Allison, 2002).

2 Die Bundestagswahl 2009 in (Ost-West)-Perspektive

2.1 Globale Ost-West-Differenzen

Bereits ein erster Blick auf die Wahlkarten zeigt, daß sich das Ergebnis der Bundestagswahl 2009 im alten Bundesgebiet deutlich vom Wahlausgang in den neuen Ländern unterscheidet. Wie aber läßt sich das Ausmaß dieser Unterschiede mit einer einzigen Maßzahl quantifizieren? In unseren Analysen zu früheren Bundes- und Landtagswahlen (Arzheimer und Falter, 1998, 2002, 2005) haben wir vorgeschlagen, die Ost-West-Unterschiede mit einer Variante des bekannten Pedersen-Index (Pedersen, 1983) zusammenzufassen.

Dazu betrachten wir getrennt für die Unionsparteien, die SPD, die FDP, die Grünen, die PDS/Linkspartei sowie die (heterogene) Gruppe aller „sonstigen“ Parteien1 die absoluten Prozentpunktdifferenzen zwischen den Wahlergebnissen in den alten Ländern (einschließlich des früheren Westteils von Berlin) und den neuen Ländern (einschließlich des früheren Ostteils von Berlin).2 Als Prozentuierungsbasis dient dabei jeweils die Anzahl der Wahlberechtigten, da nur so die tatsächlichen Mobilisierungsleistungen der Parteien sichtbar werden. Zur Summe dieser absoluten Differenzen wird dann noch die absolute Differenz der Nichtwähler addiert und das Ergebnis durch zwei geteilt. Im Ergebnis erhält man so eine Maßzahl, deren theoretischer Wertebereich zwischen 0 (keine Ost-West-Unterschiede) und 100 (es gibt ausschließlich reine „Ost-“ bzw. „Westparteien“) liegt.

Bei den vergangenen Bundestagswahlen hat dieser Index empirisch Werte zwischen 14.2 (1990) und 21.6 (1998) erreicht. Während die Werte 2002 und 2005 im Bereich von 20 Punkten lagen, wurde 2009 wiederum ein Wert von 21.6 erzielt. Von einer Annäherung im Wahlverhalten kann mithin – zumindest was die Verteilung im Aggregat betrifft – keine Rede sein. Ursachen für diesen hohen Indexwert sind neben den bekannten ostdeuschen Besonderheiten – starke Stellung der Linkspartei/PDS und vergleichsweise niedrigen Werte für Grüne und FDP – die sehr niedrige Wahlbeteiligung und sowie das sehr schwache Abschneiden der SPD.

Allerdings sind die Regionen3 wie schon bei früheren Wahlen (Arzheimer und Falter, 2005) in sich durchaus heterogen. Dies gilt vor allem für Westdeutschland. Hier weichen trotz des vergleichsweise schwachen Abschneidens der CSU viele Kreise und Städte in Bayern stärker vom westdeutschen Ergebnis ab als das Ostdeutschland vom gesamtdeutschen Ergebnis tut (vgl. Karte 1). Die für die Bundestagswahl 2002 beschriebene elektorale Dreiteilung Deutschlands (Pappi und Shikano, 2003, S. 4-6) war also keineswegs nur der Kandidatur des damaligen CSU-Vorsitzenden Stoiber geschuldet.

Interessanter als das bloße Faktum der Ost-West-Unterschiede ist aber selbstverständlich, wie, wo und wann die Ost-West-Differenzen in der Stimmenverteilung auftreten. Betrachtet man innerhalb von alter Bundesrepublik und neuen Ländern die Aggregatveränderungen von Bundestagswahl zu Bundestagswahl (dies entspricht der üblichen Berechnungsweise des Pedersen-Index), so zeigt sich, daß die Aggregatverschiebungen in Westdeutschland mit Indexwerten im Bereich von 4 bis 8 Punkten jeweils recht überschaubar waren. Im Osten hingegen wurden vor allem in den 1990er Jahren Werte in einer Größenordnung verzeichnet, die man sonst nur aus der Phase der Neuformierung des westdeutschen Parteiensystems während der 1950er Jahre kannte. Dieser scheinbare Widerspruch zwischen konstanten Ost-West-Unterschieden und hoher ostdeutscher Aggregatvolatilität erklärt sich aus der relativ stabilen Unterstützung für die Linkspartei/PDS in Kombination mit erheblichen Fluktuationen zwischen den anderen Parteien. Das Amalgam von Kontinuität und Wandel galt in der Vergangenheit als das eigentliche Spezifikum des ostdeutschen Wahlverhaltens (Arzheimer und Falter, 2005).

Bei der Bundestagswahl 2009 hat sich das Verhältnis beider Landesteile jedoch umgekehrt: Mit 14.3 Punkten liegt der Index für Westdeutschland nicht nur deutlich über dem entsprechenden Wert für Ostdeutschland (10.9) sondern übertrifft auch alle historischen westdeutschen Werte seit 1953. In diesem Indexwert spiegelt sich eine ganze Reihe von westdeutschen Entwicklungen wider: der Anstieg des Nichwähleranteils auf fast 28 Prozent, die dramatischen Verlust der SPD, das Erstarken der FDP und nicht zuletzt die Zugewinne der Linkspartei/PDS, die (ausgehend von einem 2005 immer noch recht niedrigen Niveau) ihren auf die Wahlberechtigten bezogenen Stimmenanteil um mehr als 50 Prozent steigern konnte.

Auch wenn ihr Stimmenanteil im Osten weiterhin rund dreimal höher liegt als im alten Bundesgebiet, kann die Linkspartei/PDS damit erstmals seit der Wiedervereinigung als gesamtdeutsche Partei betrachtet werden: Mehr als die Hälfte, nämlich 42 ihrer 76 Abgeordneten sind über Listen in den 10 alten Bundesländern (ohne Berlin) ins Parlament eingezogen.4 Dies ist ohne Zweifel eines der interessantesten Ergebnisse der jüngsten Bundestagswahl.

2.2 Der Durchbruch der Linkspartei/PDS im Westen, 2002-2009

Bekanntlich entstand die PDS durch die zweifache Umbenennung der früheren Staatspartei SED (Bortfeldt, 1992). Dementsprechend handelte es sich zunächst um eine rein ostdeutsche Partei. Zur Beginn des neuen Jahrhunderts mußte die 1990 begonnene Strategie der Westausdehnung der PDS als gescheitert gelten. Im Jahr 2002 verfügte die PDS in den alten Ländern (ohne Berlin) über lediglich rund 3 000 Mitglieder. Selbst in großen Flächenländern wie Bayern und Baden-Württemberg hatten die jeweiligen Landesverbände nur rund 500, in Nordrhein-Westfalen gerade einmal 1 300 Mitglieder (Niedermayer, 2009a, S. 11).

Hierbei handelte es sich zu einem großen Teil um frühere Mitglieder des Bundes Westdeutscher Kommunisten (BWK), ehemalige DKP-Mitglieder sowie parteipolitisch ungebundene junge Linke (Hough, Koß und Olsen, 2007, S. 135), die mit den über 60 000 ostdeutschen PDS-Mitgliedern oft kaum etwas gemein hatten und auf die westdeutschen Wähler wenig attraktiv wirkten. Bei der für die PDS ohnehin verheerenden vierten gesamtdeutschen Bundestagswahl von 2002 konnte die Partei in den alten Ländern nur in zwei der hier betrachteten Gebiete – den Wahlkreisen Hamburg-Mitte und Hamburg-Altona – mehr als zwei Prozent der Wahlberechtigten für sich mobilisieren. Über Hamburg hinaus fand die PDS noch in einigen norddeutschen Großstädten (Bremen, Kiel), in Teilen Südhessens (Darmstadt, Frankfurt), im Westteil Berlins sowie einigen früheren industriellen Zentren (Duisburg, Wuppertal, Kassel) Zuspruch.

In drei Viertel der Gebiete stimmte jedoch weniger als ein Prozent der Wahlberechtigten für die PDS. Ironischerweise schnitt die Partei bei den als Nebenwahlen geltenden Landtagswahlen in Westdeutschland häufig noch schlechter ab als bei den Bundestagswahlen (Arzheimer und Falter, 2005), was sich vermutlich daraus erklärt, daß letztere vom (vergleichsweise) positiven Image der Bundespartei und deren professionellen Wahlkämpfen dominiert werden (Hough, Koß und Olsen, 2007, S. 135).

Mehr als zehn Jahr nach der Wiedervereinigung war die PDS somit immer noch eine reine Ostpartei, deren Erfolge sich vor allem auf ostdeutsche Identitäten und ein Bedürfnis nach einer speziellen Interessenvertretung gründeten (Neller, 2006; Neller und Thaidigsmann, 2002). Die Wahrscheinlichkeit, daß es in absehbarer Zeit gelingen könnte, im Westen schlagkräftige Parteigliederungen aufzubauen und damit das elektorale Überleben auf Bundesebene zu sichern, schien denkbar gering. Dementsprechend galten die westdeutschen Landesverbände innerhalb der PDS als Sorgenkinder. Noch im Frühjahr 2005 sprach Gregor Gysi in einem Interview, das bei den westdeutschen Parteimitgliedern für großen Unmut sorgte, davon, daß die PDS im Westen fremd bleibe und „eher wie eine ausländische Partei“ wirke.5

In dieser Situation boten das rechtlich wie politisch mit erheblichen Risiken behaftete Wahlbündnis mit der WASG für die überraschend angesetzte Bundestagswahl 2005 sowie die Perspektive einer möglichen späteren Verschmelzung beider Gruppierungen der Führung der PDS die völlig unerwartete Chance, die Partei kurz- und mittelfristig zu stabilisieren. Bekanntermaßen entschied sich die Parteispitze dafür, diese Chance zu nutzen, indem sie – teils gegen erheblichen Widerstand der lokalen und regionalen Gliederungen – die Landeslisten der PDS für WASG-Kandidaten öffnete. Im Ergebnis gelangte die PDS – 2005 nun unter dem neuen Namen „Die Linkspartei.PDS“ – im Westen erstmals in die Nähe der Fünfprozenthürde und erzielte dank des sehr guten Abschneidens im Osten insgesamt sogar mehr Mandate als die Grünen.

Vor dem Hintergrund dieser politischen Entwicklungen ist es nicht überraschend, daß in den alten Ländern ohne Berlin auf der Ebene der Kreise, kreisfreien Städte und Wahlkreise mit r = 0.49 kein allzu enger Zusammenhang zwischen den PDS-Erfolgen von 2002 und 2005 besteht. Regrediert man den PDS-Anteil von 2005 auf das entsprechende Ergebnis der Vorgängerwahl, so zeigt sich ein ausgeprägtes räumliches Muster der (positiven) Residuen: Im Saarland sowie in den angrenzenden Gebieten in Rheinland-Pfalz erreichte die Partei sehr viel höhere Zustimmungsraten, als dies nach den Ergebnissen von 2002 zu erwarten gewesen wäre, die hier die bisherige räumliche Verteilung der PDS-Anhänger sowie indirekt auch die organisatorische Aufbauleistung der westdeutschen Landesverbände repräsentieren.

Dieses besondere Muster erklärt sich vermutlich aus der starken Verwurzelung Oskar Lafontaines in der Region. Lafontaine war zwar erst im Frühsommer 2005 in die Partei eingetreten, wurde aber als deren Spitzenkandidat wahrgenommen.6 In Karte 2 sind die vor dem Hintergrund der Vorgängerwahl unerwartet großen Erfolge für die Linkspartei in dieser Region deutlich zu erkennen. Ein interessanter Aspekt ist dabei die Ausstrahlung nach Rheinland-Pfalz, d. h. über das Gebiet des saarländischen Landesverbandes hinaus. Dieses kann zum einen als Beleg für die persönliche Wirkung Lafontaines, zum anderen als Hinweis auf die noch nicht sehr stark verfestigte organisatorische Struktur der Partei gedeutet werden, für die die Grenzen zwischen den Landesverbänden hier offensichtlich keine große Rolle spielen.

Auch absolut betrachtet erzielte die Linkspartei im Südwesten mit Zuwächsen von sieben bis 14 Prozentpunkten7 und Stimmenanteilen von bis zu 15 Prozent der Wahlberechtigten die mit weitem Abstand besten Resultate in den alten Bundesländern. Weitere Hochburgen der Partei waren Teile des Ruhrgebietes, Frankfurt/Main, Hamburg, Bremen und Bremerhaven sowie das bayrische Schweinfurt, die Basis des WASG-Mitbegründers und heutigen Parteivorsitzenden Klaus Ernst. Trotz der bemerkenswerten Zugewinne war die Unterstützung für die Linkspartei in Westdeutschland deshalb sehr stark regionalisiert.

Diese ausgeprägte räumliche Konzentration der Unterstützung für die Linkspartei zeigt sich nicht nur im Kartenbild, sondern läßt sich auch quantifizieren: Moran’s I als Maß der globalen räumlichen Autokorrelation (O’Loughlin, 2002) erreicht sowohl für den Stimmenanteil der PDS/Linkspartei bei den Wahlen von 2002 und 2005 (I = 0.39 bzw. I = 0.45) als auch für die Residuen aus der einfachen Regression (I = 0.41) recht hohe Werte.8 Im Falle der Residuen von 2005 geht dieser Wert zu einem großen Teil auf die südwestdeutschen Gebiete zurück.9

Bei der Bundestagswahl 2009 hat sich das für 2005 beschriebene Muster der Linkspartei-Erfolge im wesentlichen fortgesetzt. Bezogen auf die Wahlberechtigten hat die mittlerweile mit der früheren WASG verschmolzene Linkspartei in den alten Ländern nochmals rund 2.2 Prozentpunkte hinzugewonnen. Ihre maximalen Zugewinne im Bereich von 4 bis 4.7 Prozentpunkten erreichte sie dabei in norddeutschen Gebieten, wo sie bereits 2005 durchschnittlich oder leicht überdurchschnittlich abgeschnitten hatte (Bremerhaven, Salzgitter, Wilhelmshaven, Aurich, Emden). Die geringsten Zuwächse von 0.3 bis zu einem Punkt waren einerseits in Bayern, wo die Partei vielerorts auf niedrigem Niveau stagniert, andererseits im Saarland zu verzeichnen, wo die Partei ihr Potential offenbar weitgehend ausgeschöpft hat. Dennoch bilden das Saarland und die angrenzenden rheinland-pfälzischen Gebiete auch 2009 zusammen mit Hamburg, Bremen, Bremerhaven, Teilen des Ruhrgebietes und einigen norddeutschen Gebieten den elektoralen Schwerpunkt der Partei.

Trotz ihrer bedeutenden Zugewinne, die man bei der Bundestagswahl 2002 und auch noch 2005 kaum für möglich gehalten hätte, bleibt die Linkspartei damit auch 2009 im Westen eine Gruppierung, die sich sehr stark auf einige regionale Hochburgen stützt. Dies zeigt sich zum einen an dem sehr hohen Wert von 0.5 für Moran’s I, zum anderen daran, daß sich mehr als 80 Prozent der räumlichen Varianz im Wahlergebnis der Linkspartei auf die Ergebnisse bei den beiden vorangegangenen Bundestagswahlen zurückführen lassen. Vor dem Hintergrund dieser Befunde stellt sich die Frage, ob die Linkspartei in beiden Gebieten unterschiedlich wahrgenommen wird und ob jeweils unterschiedliche Motive hinter ihrer Wahl stehen.


PICKarte 1: Lokale Abweichungen vom regionalen Ergebnis 2009


PICKarte 2: Residuen PDS-Wahl 2005 in den alten Ländern außer Berlin

3 Wahl und Wahrnehmung der Linkspartei/PDS in Ost und West

3.1 Soziodemographie und Einstellungen der Linkspartei/PDS-Wähler in Ost und West


Tabelle 1: Die Wähler der Linken im Ost-West-Vergleich

Mit Blick auf die Parteigeschichte steht zu erwarten, daß sich die Linkspartei in beiden Regionen Deutschlands auf durchaus unterschiedliche Elektorate stützt. Zugleich hat die Linkspartei nicht nur im Westen, sondern auch in den neuen Bundesländern erheblich an Zuspruch gewonnen. Deshalb vermuten wir, daß es gegenüber früheren Wahlen auch im Osten zu Verschiebungen innerhalb der Wählerschaft gekommen sein dürfte. Tabelle 1, in der getrennt nach Regionen die Wähler der Linkspartei allen übrigen Befragten gegenübergestellt werden, bestätigt beide Vermutungen.10

Mit Blick auf die Soziodemographie ist zunächst festzuhalten, daß in Westdeutschland Männer unter den Wählern der Linkspartei klar überrepräsentiert sind. Dies ist ein für die Elektorate nicht-etablierter Parteien typisches Muster. In Ostdeutschland hingegen ist (in Einklang mit den bisherigen Befunden zu den Wählern der PDS) dieser Effekt sehr viel schwächer ausgeprägt. Ebenfalls altbekannt ist die Tatsache, daß die ostdeutschen Wähler der Linkspartei überdurchschnittlich alt sind und der Anteil der Rentner und Pensionäre überdurchschnittlich hoch ist. Im Westen hingegen sind die Wähler der Linken im Mittel jünger als die übrigen Befragten. Dementsprechend ist auch der Anteil der Rentner deutlich geringer als unter den übrigen Befragten. Auch daß sich in beiden Landesteilen ein überproportionaler Anteil der Linksparteiwähler als „Arbeiter“ einstuft (auch wenn dies nicht unbedingt in Einklang mit dem ausgeübten oder früheren Beruf steht) ist im Lichte der bisherigen Befunde wenig überraschend.

Bemerkenswert ist jedoch, daß das Bildungsniveau der Linksparteiwähler in beiden Regionen deutlich unter dem der anderen Befragten liegt. In den bisherigen Studien zur ostdeutschen PDS-Wählerschaft war deren überdurchschnittlich hohe formale Bildung stets eins der hervorstechenden Kennzeichen gewesen. Zugleich ist der Anteil der Arbeitslosen unter den Wählern der Linkspartei im Westen rund dreimal so hoch, im Osten immerhin rund 50 Prozent höher als unter den übrigen Befragten. Diese Befunde deuten darauf hin, daß es der Linkspartei bei der Bundestagswahl 2009 in den neuen Ländern im größeren Umfang gelungen sein dürfte, über ihre bisherige Kernklientel hinaus in die Arbeiterschicht vorzudringen.

Dafür spricht auch der im Vergleich zur übrigen Bevölkerung sehr hohe Anteil von Gewerkschaftsmitgliedern, in dem sich zugleich die Verwurzelung der WASG im linken Gewerkschaftslager widerspiegeln dürfte. Offensichtlich hat es die Linkspartei 2009 geschafft, traditionelle oder zumindest potentielle SPD-Wähler zu mobilisieren.

Die Ursache dafür dürfte in der Unzufriedenheit mit den „Agenda“-Reformen und der von der SPD mitgetragenen Politik der großen Koalition liegen. Diese Unzufriedenheit zeigt sich in der Verteilung der Einstellungsvariablen. In beiden Regionen sind die Wähler der Linken überdurchschnittlich unzufrieden mit dem Funktionieren der Demokratie in der Bundesrepublik. Sie nehmen die aktuelle Wirtschaftslage negativer wahr und blicken pessimistischer in die ökonomische Zukunft als die Wähler anderer Parteien. Vor allem aber ist bei ihnen das Gefühl sehr stark ausgeprägt, daß die bundesdeutsche Gesellschaftsordnung ungerecht ist.

Ebenfalls sehr aufschlußreich ist die Bewertung des Sozialismus als abstrakter Staatsidee. Wie in der Vergangenheit wird diese politische Ordnung von den ostdeutschen Wählern der Linkspartei extrem positiv beurteilt. Die westdeutschen Wähler der Linken hingegen beurteilen die Idee des Sozialismus zwar im Mittel deutlich positiver als die übrigen westdeutschen Befragten, sind in ihrem Urteil aber zugleich weniger enthusiastisch als jene ostdeutschen Befragten, die nicht die Linkspartei gewählt haben bzw. wählen wollen.

In diesem Antwortmuster spiegeln sich zum einen – fast zwanzig Jahre nach dem Fall der Mauer – die nach wie vor bestehenden Einstellungsunterschiede zwischen Ost- und Westdeutschen wieder. Zum anderen ist dies einer der wenigen Punkte, an dem sich eine mögliche Spaltung der Linken-Wählerschaft entlang der regionalen Konfliktlinie abzeichnet.

Ein weiterer möglicher Konflikt betrifft das (damalige) Führungspersonal der Partei. Von den westdeutschen Wählern der Linken werden sowohl Lafontaine als auch Gysi fast identisch, nämlich klar positiv bewertet. Unter den ostdeutschen Wählern hingegen ist die Zustimmung zu Lafontaine erkennbar schwächer ausgeprägt, die Unterstützung für Gysi hingegen fast euphorisch.

Ein letzter in der jüngeren Geschichte der Partei begründeter Unterschied zeigt sich bei der Zahl und Zusammensetzung der Parteiidentifizierer. In den alten Bundesländern liegt der Anteil derjenigen Linken-Wähler, die sich längerfristig an die Partei gebunden fühlen, bei 49 Prozent. Dieser Wert ist für sich betrachtet erstaunlich hoch, liegt aber deutlich unter der Rate von 62 Prozent Identifizierern im Osten. Noch deutlicher sind die Unterschiede bezüglich der Wähler, die sich mit einer anderen (linken) Partei identifizieren. Im Westen sind dies rund 21, im Osten hingegen nur 6 Prozent. Offensichtlich ist die Wählerschaft der Linken im Osten derzeit noch deutlich stärker konsolidiert als im Westen. Dies zeigt sich auch darin, daß 13 Prozent der westdeutschen Wähler der Linkspartei die Grünen als eine mögliche Alternative betrachten. In den neuen Ländern liegt der entsprechende Anteil bei lediglich einem Prozent.

Zusammenfassend läßt sich festhalten, daß sich die Wähler der Linkpartei in beiden Regionen recht deutlich von den übrigen Befragten unterscheiden. Zugleich sind sie sich trotz einiger zu erwartender Unterschiede über die ehemalige innerdeutsche Grenze hinweg erstaunlich ähnlich.

3.2 Position der Linkspartei/PDS im Parteienspektrum

Die Wahlstudie 2009 enthält eine ganze Batterie von Items, mit deren Hilfe die Befragten sich selbst und die relevanten Parteien im politischen Raum verorten können. Neben der globalen Links-Rechts-Selbsteinstufung betrachten wir in diesem Abschnitt auch die wahrgenommene Position in Bezug auf die beiden Hauptkonfliktlinien des Parteienwettbewerbs in Deutschland (Pappi, 1984; Shikano, 2008): die ökonomische und die libertär-autoritäre Dimension. Für diese beiden Dimensionen stehen in der Wahlstudie zwei Indikatoren zur Verfügung, die sich auf den Konflikt zwischen einem Ausbau sozialstaatlicher Leistungen einerseits und einer Senkung der Steuern andererseits (ökonomische Dimension) sowie die Position in der Zuwanderungspolitik (libertär-autoritäre Dimension) beziehen.

Ein erstes, schon mit Blick auf den neuen Namen der Partei wenig überraschendes Ergebnis betrifft die Einstufung der Partei auf der globalen ideologischen Dimension mit den Endpunkten „links“ (1) und „rechts“ (11). Jeweils rund 90 Prozent der Befragten in beiden Landesteilen ordnen die Partei hier auf einer Position am linken Rand des Spektrums (Werte 1-3) ein. Dementsprechend sind die Differenzen zwischen den Mittelwerten (2.1 im Westen und 1.9 im Osten) zwar statistisch signifikant, inhaltlich aber wenig bedeutsam und vermutlich vor allem auf einen immer noch etwas geringeren Bekanntheitsgrad der Partei in den alten Ländern zurückzuführen.

Etwas deutlicher fallen die Unterschiede in Bezug auf die oben angesprochene ökonomische Subdimension aus. In den alten Ländern liegt die mittlere Einstufung hier bei 4.6 Skalenpunkten, während die Partei in den neuen Ländern im Mittel bei einem Wert von 5.0 schon relativ nahe am Skalenmittelpunkt von 6 eingestuft wird. Hierbei handelt es sich möglicherweise um einen Ankerpunkteffekt: Da sich die Ostdeutschen auf dieser Dimension im Mittel etwas weiter links einstufen als die Westdeutschen (5.9 vs. 6.5 Punkte) wird die Linkspartei/PDS selbst bei einer identischen Position womöglich als weniger extrem wahrgenommen.

Dramatische (und ebenfalls statistisch signifikante) Unterschiede zeigen sich schließlich bei der Einordnung auf der Zuwanderungsdimension. Mit einem Skalenwert von 4.6 Punkten wird die Partei im Osten als moderater Migrationsbefürworter wahrgenommen. In den alten Ländern liegt die mittlere Einstufung hingegen bei 5.3 Punkten, d. h. sie wird hier als eher neutral wahrgenommen.


Abbildung 1: Wahrnehmung der Linkspartei auf zwei Policy-Dimensionen

Es liegt nahe, diese Unterschiede mit Oskar Lafontaines umstrittener „Fremdarbeiter“-Rede vom Sommer 2005 und ähnlichen Äußerungen in Zusammenhang zu bringen. Tatsächlich dürften die höheren, d. h. rechteren Einstufungen der Partei in den alten Ländern vor allem darauf zurückgehen, daß viele Wahlberechtigte in Westdeutschland mit den entsprechenden Positionen der Partei kaum vertraut sind und deshalb mehr oder minder zufällig antworten. Während in den neuen Ländern 71 Prozent der Befragten die Linkspartei/PDS bezüglich dieser Frage links der Mitte einordnen, tun dies im Westen nur 58 Prozent der Bürger. Zudem gibt es im Westen eine ausgeprägte Häufung der Antworten auf der Mittelkategorie.

Der Anteil derjenigen, die nach eigener Einschätzung auf wenigstens einer der beiden Dimensionen überhaupt nicht in der Lage (oder nicht willens) sind, die Partei einzuordnen, ist mit 24 Prozent unter den westdeutschen Wählern anderer Parteien bzw. Nichtwählern am höchsten. In der ostdeutschen Vergleichsgruppe liegt der Wert mit 21 Prozent aber kaum niedriger. Selbst unter den ostdeutschen Wählern der Partei wollen sich rund 16 Prozent der Befragten nicht auf eine Einstufung der Partei einlassen.11 Dagegen ist der Anteil der Antwortverweigerer unter den westdeutschen Wählern der Linkspartei/PDS mit 8 Prozent vergleichsweise gering.

Abbildung 1 zeigt die kombinierte Wahrnehmung der Partei auf den beiden genannten Dimensionen noch einmal im Überblick für vier verschiedene Personengruppen: ost- und westdeutsche Wähler der Linkspartei (linke Spalte) sowie ost- und westdeutsche Nichtwähler bzw. Wähler anderer Partein (rechte Spalte). Die Linien verbinden dabei – analog zu den Höhenlinien in einer topographischen Karte – Punkte mit gleicher Wahrscheinlichkeitsdichte.12 Deutlich ist hier zu erkennen, daß viele Wähler trotz des scheinbar klaren Profils Schwierigkeiten damit haben, die Partei (richtig) einzuordnen.

Geht man davon aus, daß die Linkspartei/PDS tatsächlich bezüglich beider Dimensionen im linken Bereich des politischen Spektrums verortet ist,13 dann sind in allen vier Gruppen maximal die Hälfte derjenigen Befragten, die überhaupt ein solches Urteil abgeben, in der Lage, die Partei korrekt zu positionieren. Am niedrigsten ist dieser Anteil paradoxerweise bei den ostdeutschen PDS-Wählern, obwohl diese Gruppe am besten mit der Programmatik der Partei vertraut sein sollte. Lediglich 35 Prozent dieser Personen ordnen die PDS auf beiden Dimensionen links der Mitte ein.

In Abbildung 1 ist dies recht gut zu erkennen. Viele ostdeutsche PDS-Wähler ordnen die Partei in der Mitte oder sogar etwas rechts von der Mitte des ideologischen Raumes ein. Hinzu kommen zwei schwer zu erklärende lokale Maxima: Knapp zehn Prozent der ostdeutschen PDS-Wähler ordnen die Partei im rechten oberen Quadranten (wirtschaftspolitisch rechts und gegen Zuwanderung) ein. Weitere neun Prozent glauben, daß die Partei für eine Erweiterung der Zuzugsmöglichkeiten und den Abbau von Sozialleistungen stehe.

Nur marginal korrekter fällt die Einschätzung der Partei durch die westdeutschen Befragten aus: Hier plazieren 42 Prozent der Linkspartei-Wähler bzw. 44 Prozent der anderen Befragten die Partei auf beiden Dimensionen im linken Spektrum. Lediglich unter den ostdeutschen Nichtwählern und Wählern anderer Parteien gelangen zumindest 50 Prozent der Befragten zu einer korrekten Einschätzung der PDS.

Für sich genommen scheinen diese Befunde darauf hinzudeuten, daß ideologische Überlegungen bei der Wahl der Linkspartei/PDS keine große Rolle spielen dürften. Denkbar ist aber auch, daß die beiden Dimensionen durch die Indikatoren nur unzureichend erfaßt werden.

Für diese letzte Interpretation spricht, daß die Wahrnehmung der Linkspartei/PDS auf der allgemeinen Links-Rechts-Skala und die Einstufung auf der Sozialleistungen/Steuersenkungs-Skala praktisch unabhängig voneinander sind, obwohl normalerweise angenommen wird, daß die allgemeine Links-Rechts-Dimension wesentlich von ökonomischen Verteilungskonflikten geprägt wird (Fuchs und Klingemann, 1989). Die bivariate Korrelation beider Maße liegt in den vier hier betrachteten Gruppen zwischen −0,10 (Linksparteiwähler West) und 0,19 (andere Befragte Ost).14 Während sich im Falle der Linkspartei argumentieren ließe, daß diese vielen Wählern immer noch nicht vertraut ist, zeigt sich bei der Einstufung der SPD ein sehr ähnliches Muster. Aus unserer Sicht spricht dies dafür, daß zumindest das ökonomische Item keine valide Messung der latenten Dimension ermöglicht.15

Im Ergebnis bleibt festzuhalten, daß die große Mehrheit der Wähler in Ost und West die Linkspartei am linken Rand des Parteienspektrums einordnet. Eine differenziertere Einschätzung entlang der beiden Hauptdimensionen des deutschen Parteinwettbewerbs scheitert am diffusen Erscheinungsbild der Partei, den Unzulänglichkeiten der Operationalisierung oder an einer Kombination beider Faktoren.

4 Die Wahlentscheidung bei der Bundestagswahl 2009 im Ost-West-Vergleich

4.1 Die Wahrnehmung der Parteien in Ost und West

Bevor wir uns der eigentlichen Wahlentscheidung zuwenden, stellt sich die Frage, ob die zur Wahl stehenden Alternativen in beiden Landesteilen überhaupt in gleicher oder zumindest ähnlicher Form wahrgenommen werden. Für die Linkspartei/PDS haben wir diesen Punkt in Abschnitt 3.2 mit Bezug auf zwei Policy-Dimensionen bzw. die allgemeine Links-Rechts-Dimension bereits relativ ausführlich erörtert. In diesem Abschnitt wollen wir der Frage nachgehen, wie Ost- und Westdeutsche die Gesamtheit der zur Wahl stehenden (relevanten) Parteien, d. h. das Parteiensystem wahrnehmen.

In den Jahren seit der Wiedervereinigung wurde die Entwicklung des deutschen Parteiensystems vor allem unter dem Gesichtspunkt einer Regionalisierung diskutiert (zusammenfassend Niedermayer, 2009b, S. 406-408): Während sich im Westen das 2+2-Parteiensystem der 1980er Jahre erhalten hatte, fiel es der FDP und vor allem den Grünen schwer, in den neuen Ländern Fuß zu fassen. Statt der aus der alten Bundesrepublik bekannten Konstellation hatte sich dort ein regionales Dreiparteiensystem aus CDU, PDS und SPD etabliert.

Dabei avancierte die PDS auf kommunaler und regionaler Ebene häufig zur zweitstärksten oder sogar zur stärksten Kraft und beteiligte sich in Mecklenburg-Vorpommern und in Berlin gemeinsam mit der SPD an der Bildung von Landesregierungen. Auf Bundesebene und in Westdeutschland gilt eine solche Zusammenarbeit hingegen immer noch als ausgeschlossen bzw. hoch problematisch (zu den veränderten Mustern der Koalitionsbildung nach der Wiedervereinigung vgl. ausführlich Kropp, 2010). Schon aus diesem Grund stünde zu erwarten, daß sich die Wahrnehmung des Parteiensystems in beiden Regionen unterscheidet.

Andererseits gibt es aber auch die These, daß sich das deutsche Parteiensystem mit der Bundestagswahl 2005 strukturell, nämlich zu einem „fluiden Fünfparteiensystem“ (Niedermayer, 2001, 2008) gewandelt habe. In einem solchen System treten zwar weiterhin regionale Unterschiede auf, diese sind aber nicht mehr notwendigerweise von Dauer. Ein wichtiges Indiz für die Gültigkeit dieser Hypothese ist das häufig sehr gute Abschneiden von FDP und Grünen in den neuen Ländern während der letzten Jahre sowie selbstverständlich das Erstarken der Linkspartei/PDS im Westen.

Empirisch läßt sich die Wahrnehmung des Parteiensystems durch die Bürger in unterschiedlicher Weise erfassen. Einen einfach zu operationalisierenden und für die Befragten wenig belastenden Zugang haben Arzheimer und Klein (1997) vorgeschlagen: Wie viele andere Wahlstudien enthält auch die GLES eine Batterie von elfstufigen Ratingskalen, mit deren Hilfe die Befragten ihre Sympathie oder Antipathie gegenüber den fünf16 relevanten Parteien ausdrücken können. Aus der Korrelationsmatrix dieser Skalometerwerte lassen sich Informationen über die wahrgenommene Ähnlichkeit der Parteien ableiten, ohne daß (1) den Befragten eine Vielzahl von paarweisen Vergleichen der Parteien abverlangt wird und (2) ohne daß den Befragten Vorgaben bezüglich der Dimensionen gemacht werden, die sie ihren Ähnlichkeitsurteilen zugrunde legen.

Die zehn impliziten Ähnlichkeitsurteile (Pearsonsche Korrelationen) skalieren wir so um, daß sie als Distanzen interpretiert werden können, und unterziehen – getrennt nach alten und neuen Bundesländern – diese Distanzen einer klassischen Multidimensionalen Skalierung, um die wahrgenommenen Distanzen zwischen den Parteien in einem zweidimensionalen Raum abzubilden.



Abbildung 2: Wahrnehmung des Parteiensystems in West- und Ostdeutschland

Abbildung 2 zeigt das Ergebnis der Skalierung. In beiden Regionen können die Parteien grundsätzlich sehr gut in den zweidimensionalen Raum eingepaßt werden.17 Anders als man vermuten könnte, ergeben sich dabei für Ost- und Westdeutschland praktisch identische Konfigurationen, die partiell die Einordnung der Parteien auf einer Links-Rechts-Dimension widerspiegeln. Union/FDP und SPD/Grüne bilden in den Augen der Befragten jeweils eine Art Protokoalition. Die Linkspartei/PDS wird in maximaler Entfernung von den bürgerlichen Parteien und in der Nähe der beiden anderen linken Parteien eingeordnet. Auffällig ist dabei aber die relativ große Entfernung von der SPD, die in etwa der Distanz zwischen Union und SPD entspricht. Die zentrale Aussage von Abbildung 2 ist jedoch, daß das Verhältnis der Parteien untereinander und insbesondere die Position der Linkspartei/PDS gegenüber den anderen Parteien im Umfeld der Bundestagswahl 2009 in beiden Landesteilen sehr ähnlich wahrgenommen wurde. Dies ist eine klare Veränderung gegenüber früheren Befunden, etwa von Arzheimer und Klein (1997)

4.2 Wahlteilnahme und die Rolle von Parteiidentifikationen

In der Tradition des Ann-Arbor-Modells (Campbell, Gurin und Miller, 1954; Campbell u. a., 1960) ist die Parteiidentifikation der zentrale Prädiktor für das Wahlverhalten. Auch die „revisionistische“ Neuinterpretation des Konzeptes durch Vertreter des Rational-Choice-Ansatzes (Fiorina, 1981, 2002; Popkin, 1994) sowie neuere Ansätze innerhalb des sozialpsychologischen Paradigmas haben an dieser grundsätzlichen Bewertung wenig geändert. Unabhängig von den Debatten über den exakten Status der Parteiidentifikation und deren optimaler Operationalisierung hat sich die konzeptuelle Unterscheidung zwischen kurzfristigen Einflüssen auf das Wahlverhalten und einer längerfristigen Loyalität gegenüber einer bestimmten Partei, die wie eine Art Voreinstellung wirkt, über den engeren Kreis der Vertreter des Ann-Arbor-Modells hinaus etabliert (Rudi und Schoen, 2005; Schmitt-Beck, 2011).

In unseren Beiträgen zu den bisherigen gesamtdeutschen Bundestagswahlen (Arzheimer und Falter, 1998, 2002, 2005; Kaspar und Falter, 2009) haben wir wiederholt darauf hingewiesen, daß Parteibindungen einen wesentlichen Beitrag zum Verständnis der Ost-West-Unterschiede im Wahlverhalten leisten können. Aufgrund der jüngsten Geschichte und der nach wie vor bestehenden sozialstrukturellen Unterschiede sind Parteibindungen in den neuen Bundesländern deutlich seltener und – dort wo sie vorhanden sind – auch schwächer ausgeprägt als im Westen. Diese strukturellen Unterschiede sind aus unserer Sicht mit dafür verantwortlich, daß der Anteil der Nicht- und Wechselwähler im Gebiet der früheren DDR deutlich höher ist als in der alten Bundesrepublik.

Auswertungen der monatlichen Politbarometerstudien (nicht tabellarisch ausgewiesen) deuten darauf hin, daß der in Westdeutschland seit den 1980er Jahren zu beobachtende Trend eines langsamen, aber kontinuierlichen Abschmelzens der Parteibindungen (Arzheimer, 2006) in den letzten Jahren zu einem Stillstand gekommen ist. Auch im Osten scheint der Anteil der Parteiidentifizierer weiterhin weitgehend stabil zu bleiben.

Diese Abschwächung des Abwärtstrends spiegelt sich auch in der Wahlstudie wider. Dort geben 70 Prozent der westdeutschen und 59 Prozent der ostdeutschen Befragten an, über eine langfristige Parteibindung zu verfügen.18 Diese relativ hohen Werte dürften partiell allerdings auch auf eine Aktivierung von Parteiidentifikationen durch den Wahlkampf zurückgehen. Zudem besteht nach wie vor ein deutlicher Unterschied zwischen beiden Regionen, da der Anteil der nach eigenen Angaben keiner Partei besonders verbundenen Befragten im Osten nach wie vor fast anderthalbmal so hoch ist wie im Westen.


Identifikation mit …WestOstWest:Ost
Union28221 , 3

Tabelle 2: Verteilung der Parteiidentifikationen in West- und Ostdeutschland

Auch bei der Verteilung der Identifikationen selbst zeigen sich deutliche Unterschiede, die den Erwartungen entsprechen (vgl. Tabelle 2). Langfristige Bindungen an die früheren Bonner Parteien sind in den neuen Ländern immer noch deutlich seltener als im alten Bundesgebiet. Vergleichsweise gut schneidet hier noch die Union ab, die im Westen rund 1,3-mal soviele langfristige Anhänger hat wie im Osten. Für FDP, Grüne und SPD liegt das Verhältnis West:Ost im Bereich von 1,5 bis 1,7. Die Linkspartei hingegen verfügt bezogen auf die Zahl der Befragten trotz ihrer Zuwächse in der alten Bundesrepublik im Osten über fast dreimal soviele Anhänger wie im Westen. Diese Unterschiede in der Verteilung der Parteiidentifikationen erklären einen erheblichen Teil der Ost-West-Unterschiede im Wahlverhalten.


sehr schwach13
ziemlich schwach68
ziemlich stark4539
sehr stark1510

Tabelle 3: Stärke der Parteiidentifikationen in West- und Ostdeutschland

Was schließlich die Qualität der Parteiidentifikationen betrifft, so sind diese in Ostdeutschland tatsächlich etwas schwächer ausgeprägt als im Westen. Während dort die Mehrheit (45 Prozent) der Wähler angibt, über eine „ziemlich starke“ Identifikation zu verfügen, ordnen sich im Osten nur 39 Prozent der Respondenten in dieser Gruppe ein (vgl. Tabelle 3). Insgesamt sind die Unterschiede zwischen beiden Regionen in dieser Hinsicht aber relativ klein.19


NichtwahlWahl entgegen PI
Stärke PI−0,689∗∗∗−0,248∗∗
Ost×Stärke PI−0,253−0,281
N 2414

Tabelle 4: Effekt der Parteiidentifikation auf die Wahlbteiligung/-entscheidung in Ost und West

Wenn man die Vorstellung einer Parteiloyalität ernst nimmt, dann sollten bei der Modellierung der Entscheidung von Parteianhängern über die Wahlteilnahme mindestens drei Ausprägungen des Wahlverhaltens unterschieden werden: die Wahlentscheidung im Sinne der PI, die Wahlentscheidung gegen die PI und die Nichtwahl, die einem überzeugten Parteigänger, der mit der Programmatik oder den Kandidaten der eigentlich bevorzugten Partei unzufrieden ist, womöglich leichter fällt als die Wahl einer „falschen“ Partei. Betrachtet man das Wahlverhalten bei der Bundestagswahl 2009 nach diesen drei Kategorien, so zeigen sich deutliche Unterschiede sowohl zwischen den Parteien als auch zwischen den beiden Regionen, die wiederum für einen Teil der Unterschiede in den regionalen Wahlergebnissen verantwortlich sind.

Tabelle 4 enthält die Schätzungen für eine multinomiale logistischen Regression, die diese Unterschiede modelliert.20 Neben der Region sowie der Stärke und Richtung der Parteiidentifikation enthält das Modell zwei Variablen, die vor allem als Prädikatoren der Wahlbeteiligung eine wichtige Rolle spielen: das politische Interesse sowie einen Dummy für das Vorliegen eines (Fach-)Abiturs als Indikator für den Grad der formalen Bildung. Etwaige Ost-West-Unterschiede werden durch Interaktionen mit dem Regional-Indikator abgebildet. Befragte ohne Parteiidentifikation bleiben aufgrund der gewählten Perspektive außer Betracht. Die Referenzkategorie für das politische Verhalten ist die Wahlentscheidung für jene Partei, mit denen sich die Befragten identifizieren. Die Referenzgruppe sind westdeutsche Unionsanhänger mit sehr schwachem politischem Interesse und sehr schwacher Parteibindung (jeweils =0 ) ohne Abitur.

Aus der Konstante läßt sich ablesen, daß die Handlungsvariante „Nichtwahl“ für die Referenzgruppe rund 1,7-mal (= exp(0,503)) wahrscheinlicher ist als die Wahlentscheidung im Sinne der Parteiidentifikation. Eine gegen die Parteibindung gerichtete Wahlentscheidung ist hingegen sehr unwahrscheinlich: die entsprechende Wahrscheinlichkeit beträgt nur rund ein Drittel derjenigen für die Wahl im Sinne der Parteiidentifikation.21

Die Koeffizienten in den Zeilen vier bis sieben repräsentieren die Kontraste zwischen den Anhängern der Union und den Unterstützern der übrigen Parteien. Auffällig ist hier zunächst, daß sich (wiederum bezogen auf westdeutsche Befragte ohne Abitur, mit geringem politischem Interesse und schwach ausgeprägter Parteibindung) die Anhänger der Grünen und vor allem der FDP signifikant von den Anhängern der Union (und implizit auch von denen der SPD und der Linken) unterscheiden: Für beide Gruppen ist dem Modell zufolge nicht nur eine Wahlenthaltung, sondern auch eine Entscheidung gegen die eigentliche Identifikation wahrscheinlicher als eine Wahlentscheidung für die eigentlich präferierte Partei. Dieses Ergebnis spricht dafür, daß zumindest die schwachen Bindung an diese beiden kleinen Parteien kaum im Sinne einer echten Loyalität interpretiert werden sollten.

Die Zeilen drei und acht zeigen die Koeffizienten für die Effekte der Stärke der Parteiidentifikation und des politischen Interesses. Erstere reduziert erwartungsgemäß sehr stark die Wahrscheinlichkeit von Nichtwahl bzw. abweichendem Wahlverhalten. Bei der Bewertung der Effekte ist zu bedenken, daß diese Variable eine Spannweite von vier Skalenpunkten und damit einen sehr großen maximalen Effekt hat. Für Befragte mit sehr engen Bindungen an die bevorzugte Partei ist es deshalb fast ausgeschlossen, daß diese nicht gewählt wird.

Auch das politische Interesse, das ebenfalls auf einer Skala mit einer Spannweite von vier Punkten gemessen wurde, hat einen beträchtlichen Effekt auf das Wahlverhalten. Dieser konzentriert sich jedoch auf die Wahlbeteiligung. Für das Verhältnis der Wahrscheinlichkeiten von identifikationsgeleiteter und abweichender Parteienwahl ist das politische Interesse hingegen unerheblich. Die Zeilen neun bis fünfzehn schließlich enthalten die Interaktionseffekte, aus denen abzulesen ist, wie sich die Wirkung von Bildung, Stärke und Richtung der Parteiidentifikation und politischem Interesse in beiden Regionen unterscheiden. Bemerkenswert ist hier vor allem, daß formale Bildung und Stärke der Parteibindung die Wahrscheinlichkeit von Nichtwahl bzw. Wahl entgegen der Identifikation stärker reduzieren als im Westen.


Abbildung 3: Wahl gemäß Parteiidentifikation nach Region, Richtung und Stärke der Parteibindung

Wegen der großen Zahl von Koeffizienten und der Ambiguitäten, die sich aus den verschiedenen Kodierungsmöglichkeiten ergeben, gestaltet sich die weitergehende Interpretation der Tabelle schwierig. Im folgenden konzentrieren wir uns deshalb auf die graphische Analyse des mit der Parteibindung konformen Wahlverhaltens in Abhängigkeit von der Region sowie der Richtung und Stärke der Parteibindung. Abbildung 3 zeigt die entsprechenden Schätzungen.22

Im Ergebnis ist zunächst klar zu erkennen, daß die Intensität der Identifikation für alle Parteien von entscheidender Bedeutung ist: Mit zunehmender Stärke der Bindung steigt die Wahrscheinlichkeit einer im Sinne der Identifikation korrekten Entscheidung ganz klar an.

Ebenfalls deutlich zu erkennen ist nun, wie stark die Bindekraft der Identifikation über die Parteien hinweg variiert: Bei der Linken, der SPD und vor allem bei der Union führen selbst relativ schwache Identifikationen mit relativ großer Wahrscheinlichkeit zur Wahl der Partei. Bei FDP und Grünen hingegen haben selbst sehr intensive Identifikationen nur einen eingeschränkten Effekt auf das Wahlverhalten.

Dritten schließlich zeigen sich fast keine signifikanten Ost-West-Unterschiede. Die einzige Ausnahme davon sind die besonders engagierten Anhänger der Union in Ostdeutschland, die sich als geringfügig loyaler erweisen als die entsprechende Gruppe in den alten Ländern. Alle anderen Differenzen sind nicht signifikant und häufig auch sehr klein.

Inhaltlich bedeutet dies, daß sich rund zwei Jahrzehnte nach der Vereinigung keine Ost-West-Differenzen mehr nachweisen lassen, die den Charakter der Parteibindungen an sich betreffen. Die vorhandenen Unterschiede im Wahlverhalten unter den Parteianhängern gehen vielmehr auf die Verteilung und Intensität der Identifikationen sowie auf die Unterschiede in der Verteilung anderer Variablen zurück.

4.3 Ein multivariates Modell der Wahlentscheidung bei der Bundestagswahl 2009

Im letzten Teil unserer Analyse erweitern wir die Perspektive, in dem wir zum einen auch jene Befragten in die Analyse aufnehmen, die keine Parteibindung aufweisen, zum anderen einige zusätzliche Variablen berücksichtigen, die einen Einfluß auf die Wahlentscheidung haben sollten. Dabei handelt es sich einerseits um die Bewertung der Spitzenkandidaten der Parteien,23 andererseits um die generalisierte Links-Rechts-Selbsteinstufung der Befragten, die wir hier als summarischen Indikator für allgemeine Policy-Präferenzen betrachten.24 Um auch jene Befragten berücksichtigen zu können, die nach eigenem Bekunden keine langfristige Loyalität gegenüber einer Partei empfinden, haben wir die Informationen zur Richtung und gegebenenfalls Stärke der Parteibindung in einem Set von fünf Variablen zusammengefaßt. Diese haben jeweils den Wert 0, wenn ein Befragter nicht an diese Partei gebunden ist. Wenn hingegen eine Bindung an die betreffende Partei vorliegt, nimmt die entsprechende Variable je nach deren Intensität Werte zwischen 1 und 5 an. Als Basiskategorie betrachten wir die Nichtwahl bzw. Wahl einer „sonstigen“ Partei.25


Tabelle 5: Ein umfassendes Modell der Wahlentscheidung für die Bundestagswahl 2009

Angesichts der Vielzahl von Variablen, die eng mit Wahlbeteiligung und -entscheidung verbunden sind, überrascht es nicht, daß das Modell eine hervorragende Anpassung an die Daten erreicht.26 Die resultierende Tabelle enthält 140 nicht-redundante Parameterschätzungen und entzieht sich damit einer einfachen Interpretation. Klar erkennbar ist in erster Linie, daß auch in dieser Modell mit steigendem politischen Interesse die Wahrscheinlichkeit der Wahl einer (beliebigen) Partei gegenüber der Wahrscheinlichkeit der Nichtwahl zunimmt. Dies gilt für beide Regionen.

Alle weitergehenden Interpretationen erfordern aber wiederum eine graphische Darstellung. Dabei liegt unser Hauptaugenmerk zunächst auf der Einflußwirkung der Ideologie (Links-Rechts-Selbsteinstufung) auf die parteipolitisch ungebundenen Wähler, da wir hier die interessantesten Effekte erwarten.

Allerdings stellt sich hier das Problem, daß die (auf die Ebene der erwarteten Wahrscheinlichkeiten bezogene) Wirkung einer Variablen in einem non-linearen Modell stets vom Niveau aller anderen Variablen abhängt. In der Literatur wird deshalb häufig empfohlen, die erwarteten Wahrscheinlichkeiten zu berechnen, indem ein oder zwei fokale unabhängige Variablen über ihren Wertebereich variiert und alle anderen unabhängigen Variablen auf ihren Mittelwert oder Modus gesetzt werden (King, Tomz und Wittenberg, 2000; Long und Freese, 2006).

Die auf diese Weise berechneten Effekte („marginal effects at the mean“) können aber in endlichen Stichproben in die Irre führen (Greene, 2003, S. 669), vor allem wenn zwischen den unabhängigen Variablen enge Beziehungen bestehen. In vielen Fällen ist es deshalb sinnvoller, die „average marginal effects“ zu bestimmen (Bartus, 2005). Diese errechnen sich, indem für jeden einzelnen Befragten die erwarteten Wahrscheinlichkeiten berechnet werden. Dabei wird die fokale Variable (in unserem Fall die Zugehörigkeit zu einer Region) variiert, während alle anderen unabhängigen Variablen mit ihren realen Werten in die Schätzung eingehen. Anschließend werden die Mittelwerte über diese Schätzungen errechnet. Die Differenzen zwischen den geschätzten Mittelwerten entsprechen den geschätzten Effekten der fokalen unabhängigen Variablen auf die Wahrscheinlichkeit der Wahl.


Abbildung 4: Links-Rechts-Selbsteinstufung und Sympathien für die Spitzenkandidaten

In unserem Fall empfiehlt sich diese etwas komplexere Methode, weil zwischen den unabhängigen Variablen erfahrungsgemäß recht enge Zusammenhänge bestehen. Dies betrifft einerseits die Beziehung zwischen der generalisierten Ideologie und der Bewertung der Spitzenkandidaten: In beiden Landesteilen finden eher rechte Wähler Merkel bzw. zu Guttenberg und Westerwelle im Mittel deutlich sympathischer als Steinmeier und umgekehrt (vgl. Abbildung 4).27


Abbildung 5: Links-Rechts-Selbsteinstufung und Bindungen an die Parteien

Ebenfalls recht eng sind die Zusammenhänge zwischen der Links-Rechts-Selbsteinstufung und den Bindungen an die Parteien. Hier zeigen sich einige inhaltlich sehr interessante Muster (vgl. Abbildung 4). So ist zunächst noch einmal zu erkennen, daß FDP und Grüne über relativ wenige feste Anhänger verfügen. Zudem sind deren Bindungen an die jeweilige Partei häufig nur sehr schwach ausgeprägt.

Bemerkenswert ist darüber hinaus, daß in beiden Regionen Deutschlands im rechten Teil des politischen Spektrums Bindungen an die Unionsparteien immer noch weit verbreitet sind. Verglichen damit ist das linke Lager gespalten. Die SPD hat vor allem im Osten deutlich weniger langfristige Anhänger als die Union und scheint auch ein deutlich schmaleres Spektrum im Mitte-Links-Bereich abzudecken. Die relativ wenigen festen Anhänger der Grünen positionieren sich vor allem in den alten Bundesländern deutlich links von der Mitte.


Abbildung 6: Links-Rechts-Selbsteinstufung in den beiden Regionen

Von besonderem Interesse ist in Ost-West-Perspektive selbstverständlich die Situation der Linkspartei. Hier zeigt sich nochmals, daß diese selbst in den alten Ländern am linken Rand des Spektrums über eine erstaunlich große Zahl selbstdeklarierter fester Anhänger verfügt. Ebenfalls klar zu erkennen ist darüber hinaus, daß die Linkspartei/PDS in den neuen Ländern im gesamten linken Spektrum über eine große Zahl fester Anhänger verfügt und auf diese Weise den beiden anderen linken Parteien nur wenig Raum läßt.

Hinzu kommt ein weiterer Faktor, den wir bisher noch nicht angesprochen haben: In den neuen Ländern ordnen sich rund drei Viertel der Befragten links der Mitte ein, während sich im Westen linke und rechte Überzeugungen in etwa die Waage halten (Abbildung 6).


Abbildung 7: Erwartete Wahlentscheidung ungebundener Wähler nach Region und ideologischer Selbsteinstufung

Vor dem Hintergrund dieses doch gravierenden Unterschiedes in der ideologischen Zusammensetzung der beiden Teilelektorate, der wie oben gezeigt eng mit unterschiedlichen Bewertungen der Kandidaten und unterschiedlichen Parteibindungen verknüpft ist, stellt sich die Frage, ob die Ost-West-Unterschiede im Wahlverhalten primär auf diese ganz generellen Einstellungsunterschiede zurückgehen.

Abbildung 7 zeigt die auf der Grundlage des vollständigen Modells geschätzten Wahlwahrscheinlichkeiten zugunsten der fünf Parteien in Abhängigkeit von ideologischer Selbsteinstufung und Region für die besonders interessante Gruppe der ungebundenen Wähler. Für die einzelnen Parteien ergeben sich dabei durchaus relevante Unterschiede, wobei allerdings die relativ großen Konfidenzintervalle zu beachten sind. So wird die Wahrscheinlichkeit einer Entscheidung zugunsten der FDP innerhalb dieser Gruppe in den alten wie in den neuen Ländern kaum von der ideologischen Selbsteinstufung beeinflußt. Bei den Grünen in Ostdeutschland (und nur in Ostdeutschland) hingegen gibt es Hinweise auf ideologische Effekte, die aber nicht die Schwelle der statistischen Signifikanz erreichen.

Etwas klarer sind die Ergebnisse bei der Wahlentscheidung zugunsten der SPD. In beiden Regionen scheint die Wahlentscheidung von ideologischen Überlegungen beeinflußt zu werden, d. h. die Wahlwahrscheinlichkeit sinkt im rechten Spektrum ab, wobei diese Differenzen wiederum nicht signifikant sind. Zugleich ist die Partei im Osten ceteris paribus weniger erfolgreich als im Westen, wobei diese Differenzen nur im mittleren und im Mitte-Rechts-Bereich signifikant sind. Fast spiegelbildlich stellt sich die Situation der Union dar: Diese ist – wenig überraschend – bei rechteren Wählern beliebter als bei Mitte-Links-Wählern. Dieser Effekt kommt in Ostdeutschland stärker zum Tragen, wobei auch hier in der Mehrzahl der Konstellationen die Konfidenzintervalle überlappen.

Von besonderem Interesse für unsere Fragestellung ist schließlich der Effekt der Links-Rechts-Selbsteinstufung auf die Wahl der Linkspartei. Hier zeichnen sich – wiederum erwartungsgemäß – klare ideologische Effekte ab, d. h. Personen im linken Spektrum haben mit großer Sicherheit eine sehr viel stärkere Tendenz, diese Partei zu wählen, als Bürger aus dem Mitte-Rechts-Bereich. Anders, als man es aufgrund der Geschichte der Partei vielleicht erwarten könnte, unterscheidet sich die Stärke dieses Effektes in den beiden Regionen jedoch nicht.

Aus der letzten Teilgrafik schließlich geht hervor, daß die Links-Rechts-Selbsteinstufung bei parteipolitisch ungebundenen Bürgern auch einen Effekt auf die Wahrscheinlichkeit der Nichtwahl zu haben scheint. Ähnlich wie bei der Wahl der Grünen sind aber weder die Unterschiede zwischen den verschiedenen ideologischen Gruppen noch die Unterschiede zwischen den Regionen statistisch signifikant.

In der Summe läßt sich festhalten, daß die Wirkung der ideologischen Selbsteinstufung in beiden Regionen im wesentlichen identisch ist. Dies gilt auch und gerade für die Wahl der Linkspartei.

Nimmt man statt der besonders volatilen Gruppe der parteipolitisch ungebunden Bürger das gesamte Elektorat in den Blick (nicht ausgewiesen), so lassen sich überhaupt keine statistisch signifikanten Ost-West-Unterschiede im Effekt der Links-Rechts-Selbsteinstufung nachweisen. Aus der Perspektive unseres Modells erklären sich die nach wie vor erheblichen Ost-West-Differenzen im Wahlverhalten deshalb primär durch die unterschiedliche Verteilung der Einstellungsvariablen, nicht aber durch übergeordnete Kontexteffekte, etwa in Bezug auf die Wahrnehmung des Parteiensystems.

5 Fazit

Auch bei der sechsten gesamtdeutschen Bundestagswahl haben sich wieder tiefgreifende Unterschiede zwischen der alten Bundesrepublik und den neuen Bundesländern gezeigt. Diese betreffen sowohl die Wahlbeteiligungsraten als auch die Stimmenanteile der Parteien. Zur erklären sind diese Differenzen vor allem über die weiterhin bestehenden Unterschiede in der Verteilung von Einstellungen und sozio-demographischen Merkmalen.

Zugleich gibt es aber Hinweise auf eine gewisse Angleichung im Wahlverhalten der beiden Regionen: Erstmals seit den 1950er Jahren hat es in den alten Ländern Aggregatverschiebungen gegeben, die in ihrer Größenordnung mit der aus den neuen Ländern bekannten Volatilität vergleichbar sind. Damit hat sich unsere in früheren Beiträgen geäußerte These, daß das Wahlverhalten in Ostdeutschland einen Eindruck von den zukünftigen Entwicklungen in Westdeutschland vermittelt, bestätigt.

Zurückzuführen sind die Aggregatveränderungen auf den Anstieg des Nichtwähleranteils, die Stimmenverluste der früheren Volksparteien sowie den Aufstieg der Linkspartei, die mit dem Ergebnis der Bundestagswahl ihren Anspruch, als gesamtdeutsche Partei wahrgenommen zu werden, unterstreichen konnte.

Innerhalb der Wählerschaft der Linkspartei zeigt sich eine ganz Reihe von erwartbaren Ost-West-Unterschieden. Diese betreffen nicht nur die Zusammensetzung der Wählerschaft, sondern auch deren im Westen sehr stark ausgeprägte Konzentration in wenigen Hochburgen. Dennoch zeigen sich innerhalb der Wählerschaft der Linkspartei auch viele Gemeinsamkeiten: Wähler in Ost und West sind sehr stark ideologisch motiviert, nehmen die wirtschaftliche Situation als bedrohlich wahr und empfinden die Gesellschaftsordnung als ungerecht. Stärker als bei früheren Bundestagswahlen stellt sich die Linkspartei damit auch in Ostdeutschland als klassische Arbeiterpartei dar. Sollte diese Entwicklung von Dauer sein, so würde dies offensichtlich die SPD in ihrer Existenz bedrohen. Zugleich könnte damit aber ironischerweise ausgerechnet die Linkspartei für sich in Anspruch nehmen, die bisherige elektorale Ost-West-Spaltung Deutschlands wenn schon nicht zu beenden, dann doch deutlich zu reduzieren. Sie würde dann zur einzig sozialistischen Partei der deutschen Einheit.


Allison, Paul D. (2002). Missing Data. Thousand Oaks, London, New Delhi: Sage.
Arzheimer, Kai (2006). „’Dead Men Walking?’ Party Identification in Germany, 1977-2002“. In: Electoral Studies 25, S. 791–807. doi: 10.1016/ j.electstud.2006.01.004.
Arzheimer, Kai und Jürgen W. Falter (1998). „Annäherung durch Wandel? Das Ergebnis der Bundestagswahl 1998 in Ost-West-Perspektive“. In: Aus Politik und Zeitgeschichte 48.B52, S. 33–43.
— (2002). „Ist der Osten wirklich rot? Das Wahlverhalten bei der Bundestagswahl 2002 in Ost-West-Perspektive“. In: Aus Politik und Zeitgeschichte 52.B49-50, S. 27–35.
— (2005). „Goodbye Lenin? Bundes- und Landtagswahlen seit 1990: eine Ost-West-Perspektive“. In: Wahlen und Wähler. Analysen aus Anlaß der Bundestagswahl 2002. Hrsg. von Jürgen W. Falter, Oscar W. Gabriel und Bernhard Weßels. Wiesbaden: Verlag für Sozialwissenschaften, S. 244–283.
— (2013). „Versöhnen statt spalten? Das Ergebnis der Bundestagswahl 2009 und die Rolle der PDS/Linkspartei in Ost-West-Perspektive“. In: Wahlen und Wähler. Analysen aus Anlass der Bundestagswahl 2009. Hrsg. von Bernhard Weßels, Oscar W. Gabriel und Harald Schoen. Wiesbaden: VS Verlag für Sozialwissenschaften, S. 118–150.
Arzheimer, Kai und Markus Klein (1997). „Die Wähler der REP und der PDS in West- und Ostdeutschland“. In: Jahrbuch Extremismus und Demokratie. Hrsg. von Uwe Backes und Eckhard Jesse. Bd. 9. Baden-Baden: Nomos, S. 39–63.
Arzheimer, Kai und Harald Schoen (2007). „Mehr als eine Erinnerung an das 19. Jahrhundert? Das sozioökonomische und das religiös-konfessionelle Cleavage und Wahlverhalten 1994-2005“. In: Der gesamtdeutsche Wähler. Stabilität und Wandel des Wählerverhaltens im wiedervereinigten Deutschland. Hrsg. von Hans Rattinger, Oscar W. Gabriel und Jürgen W. Falter. Baden-Baden: Nomos, S. 89–112.
Bartus, Tamás (2005). „Estimation of Marginal Effects Using Margeff“. In: The Stata Journal 5.3, S. 309–329.
Bortfeldt, Heinrich (1992). Von der SED zur PDS. Bouvier.
Campbell, Angus, Gerald Gurin und Warren E. Miller (1954). The Voter Decides. Evanston: Harper und Row.
Campbell, Angus u. a. (1960). The American Voter. New York: John Wiley.
Die Linke (2009). Bundestagswahlprogramm der Partei DIE LINKE. Beschluss des Bundestags-Wahlparteitags 2009 der Partei DIE LINKE 20./21. Juni 2009 in Berlin. url: http://die-linke.de/fileadmin/download/wahlen/pdf/Beschluss_Bundestagswahlprogramm_redTB_revMS_final.pdf.
Fiorina, Morris P. (1981). Retrospective Voting in American National Elections. New Haven, London: Yale University Press.
— (2002). „Parties and Partisanship: A 40-Year Retrospective“. In:
Fuchs, Dieter und Hans-Dieter Klingemann (1989). „The Left-Right Scheme“. In: Continuities in Political Action. A Longitudinal Study of Political Orientations in Three Western Democracies. Hrsg. von Kenneth M. Jennings und Jan van Deth. Berlin: de Gruyter, S. 203–234.
Greene, William H. (2003). Econometric Analysis. Upper Saddle River: Prentice-Hall.
Hough, Dan, Michael Koß und Jonathan Olsen (2007). The Left Party in Contemporary German Politics. Houndmills: Palgrave Macmillan.
Kaspar, Hanna und Jürgen W. Falter (2009). „Angenähert oder ausdifferenziert? Das Wahlverhalten in Ost- und Westdeutschland bei der Bundestagswahl 2005“. In: Wahlen und Wähler. Analysen aus Anlass der Bundestagswahl 2005. Hrsg. von Oscar W. Gabriel, Bernhard Weßels und Jürgen W. Falter. Wiesbaden: VS Verlag, S. 202–227.
King, Gary, Michael Tomz und Jason Wittenberg (2000). „Making the Most of Statistical Analysis. Improving Interpretation and Presentation“. In: American Journal of Political Science 44, S. 341–355.
Krewel, Mona, Rüdiger Schmitt-Beck und Ansgar Wolsing (2011). „Geringe Polarisierung, unklare Mehrheiten und starke Personalisierung: Parteien und Wähler im Wahlkampf“. In: Zwischen Langeweile und Extremen: Die Bundestagswahl 2009. Hrsg. von Hans Rattinger u. a. Baden-Baden: Nomos, S. 33–57.
Kropp, Sabine (2010). „The Ubiquity and Strategic Complexity of Grand Coalition in the German Federal System“. In: German Politics 19.3/4, S. 286–311.
Long, J. Scott und Jeremy Freese (2006). Regression Models for Categorical Dependent Variables Using Stata. 2. Aufl. College Station: Stata Press.
Neller, Katja (2006). DDR-Nostalgie. Dimensionen der Orientierungen der Ostdeutschen gegenüber der ehemaligen DDR, ihre Ursachen und politische Konnotationen. VS.
Neller, Katja und S. Isabell Thaidigsmann (2002). „Das Vertretenheitsgefühl der Ostdeutschen durch die PDS. DDR-Nostalgie und andere Erklärungsfaktoren im Vergleich“. In: Politische Vierteljahresschrift 43, S. 420–444.
Niedermayer, Oskar (2001). „Nach der Vereinigung: Der Trend zum fluiden Fünfparteiensystem“. In: Parteiendemokratie in Deutschland. Hrsg. von Oscar W. Gabriel, Oskar Niedermayer und Richard Stöss. 2. Aufl. Bonn: Bundeszentrale für politische Bildung, S. 107–127.
— (2008). „Das fluiden Fünfparteiensystem nach der Bundestagswahl 2005“. In: Die Parteien nach der Bundestagswahl 2005. Hrsg. von Oskar Niedermayer. Wiesbaden: VS Verlag für Sozialwissenschaften, S. 9–35.
— (2009a). Parteimitglieder in Deutschland: Version 1/2009. Arbeitshefte aus dem Otto-Stammer-Zentrum 15. Otto-Stammer-Zentrum, FU Berlin. url: http://www.polsoz.fu-berlin.de/polwiss/forschung/systeme/empsoz/schriften/Arbeitshefte/ahosz15.pdf.
— (2009b). „Regionalisierung des Wahlverhaltens und des Parteiensystems seit 1949“. In: Wahlen und Wähler. Hrsg. von Oscar W. Gabriel, Bernhard Weßels und Jürgen W. Falter. VS Verlag für Sozialwissenschaften, S. 399–420.
O’Loughlin, John (2002). „The Electoral Geography of Weimar Germany: Exploratory Spatial Data Analyses (ESDA) of Protestant Support for the Nazi Party“. In: Political Analysis 10.3, S. 217–243. doi: 10.1093/pan/10.3.217. eprint: http://pan.oxfordjournals.org/content/10/3/217.full.pdf+html. url: http://pan.oxfordjournals.org/content/10/3/217.abstract.
Pappi, Franz Urban (1984). „The West German Party System“. In: Party Politics in Contemporary Western Europe. Hrsg. von Stefano Bartolini und Peter Mair. London: Frank Cass, S. 7–26.
Pappi, Franz Urban und Susumu Shikano (2003). „Schröders knapper Sieg bei der Bundestagswahl 2002“. In: Zeitschrift für Politik 50, S. 1–16.
Pedersen, Mogens N. (1983). „Changing Patterns of Electoral Volatility in European Party Systems, 1948-1977: Explorations in Explanation“. In: Western European Party Systems. Continuity and Change. Hrsg. von Hans Daalder und Peter Mair. Beverly Hills, London, New Delhi: Sage, S. 29–66.
Popkin, Samuel L. (1994). The Reasoning Voter. Communication and Persuasion in Presidential Campaigns. 2. Aufl. Chicago, London: University of Chicago Press.
Rudi, Tatjana und Harald Schoen (2005). „Ein Vergleich von Theorien zur Erklärung des Wählerverhaltens“. In: Handbuch Wahlforschung. Hrsg. von Jürgen W. Falter und Harald Schoen. Wiesbaden: VS Verlag für Sozialwissenschaften, S. 305–323.
Schmitt-Beck, Rüdiger (2011). „Parteibindungen“. In: Zwischen Langeweile und Extremen: Die Bundestagswahl 2009. Hrsg. von Hans Rattinger u. a. Baden-Baden: Nomos, S. 155–164.
Shikano, Susumu (2008). „Die Eigendynamik zur Eindimensionalität des Parteienwettbewerbs: eine Simulationsstudie“. In: Politische Vierteljahresschrift 49.2, S. 229–250.
Venables, William N. und Brian D. Ripley (2002). Modern Applied Statistics with S. 4. Aufl. New York, Berlin, Heidelberg: Springer.

1Im Sinne einer möglichst einfachen Vorgehensweise werden die (wenigen) ungültigen Stimmen ebenfalls dieser Gruppe zugeordnet.

2Die Berliner Wahlkreise decken sich immer noch weitgehend mit der früheren Sektorengrenze. Lediglich der Bezirk Kreuzberg muß als Bestandteil des Wahlkreises „Kreuzberg – Friedrichshain – Prenzlauer Berg Ost“ dem Ostteil zugerechnet werden.

3Idealerweise sollten Wahlergebnisse auf einem möglichst niedrigen Aggregationsniveau, d. h. auf der Ebene der Stimmbezirke analysiert werden. Daten auf der Stimmbezirksebene stehen aber momentan nur für die Bundestagswahl 2009 zur Verfügung. Die darüberliegende Ebene der Wahlkreise ist deutlich weniger gut für entsprechende Analysen geeignet, da es sich hier um vergleichsweise große und teils auch heterogene Einheiten handelt. Hinzu kommt, daß sich seit der Verkleinerung des Bundestages für die Wahl 2002 Veränderungen im Zuschnitt der Wahlkreise ergeben haben: Aufgrund der Bevölkerungsbewegungen haben Thüringen, Sachsen und Sachsen-Anhalt jeweils einen Wahlkreis an Baden-Württemberg, Bayern und Niedersachsen abgegeben. In allen sechs Bundesländern mußten deshalb Wahlkreisgrenzen neu gezogen werden. Für die folgenden Analysen wurden deshalb die in der Regionaldatenbank Genesis hinterlegten Wahlergebnisse verwendet, die auf die Landkreise und kreisfreien Städte umgerechnet sind, deren Grenzen über den Zeitraum von 2002-09 weitgehend stabil waren. In den Stadtstaaten Berlin und Hamburg sowie in einigen Großstädten und besonders großen Landkreisen mit mehreren Bundestagswahlkreisen wurden diese Daten durch die Ergebnisse der jeweiligen Wahlkreise ersetzt. Auf diese Weise ergibt sich ein hybrider Datensatz mit 447 stabilen Gebietseinheiten. Einzelheiten zur Behandlung der Großstädte und -kreise sowie zur Gebietsreform in Sachsen-Anhalt sind im Dataverse zu diesem Kapitel dokumentiert.

4Dabei handelt es sich größtenteils um Westdeutsche. Kandidaturen von Ostdeutschen auf westdeutschen Landeslisten wie etwa die von Sahra Wagenknecht in Nordrhein-Westfalen sind die Ausnahme.

5Interview mit dem Berliner Tagesspiegel vom 17.05.2005

6Lafontaine kandidierte für das Direktmandat im Wahlkreis Saarbrücken, wo er über viele Jahre zunächst Oberbürgermeister und später Ministerpräsident gewesen war, und erhielt dort 26 Prozent der Erststimmen. Sein Einzug in den Bundestag war über den ersten Platz der nordrhein-westfälischen Landesliste abgesichert.

7Bezogen auf die Wahlberechtigten gewann die Partei im Westen insgesamt rund 2,9 Prozentpunkte hinzu.

8Positive räumliche Autokorrelation bedeutet, daß sich benachbarte Einheiten bezüglich der untersuchten Variable ähnlicher sind, als dies bei einer zufälligen Verteilung zu erwarten wäre. Wenn in politikwissenschaftlichen Anwendungen räumliche Autokorrelationen auftreten, so sind, diese fast immer positiv, etwa weil sich benachbarte Einheiten im Rahmen von Diffusionsprozessen gegenseitig beeinflussen oder durch die Zugehörigkeit zu größeren Organisationen gemeinsamen räumlich organisierten Einflüssen ausgesetzt sind.

Moran’s I und vergleichbare Indikatoren sollten mit einer gewissen Vorsicht interpretiert werden, weil ihre Berechnung die Definition einer Gewichtungsmatrix erfordert, die festlegt, welche Gebietseinheiten als potentiell relevante Nachbarn betrachtet werden. In der Regel gibt es dafür eine ganze Reihe gleichermaßen plausibler Spezifikationen. Im Sinne einer möglichst einfachen Vorgehensweise verwenden wir den Kehrwert der Distanz zwischen den Zentroiden der Gebietseinheiten als Gewichtungsfaktor. Gebiete, deren Zentroide mehr als 120 Kilometer Luftlinie entfernt sind, bleiben unberücksichtigt. Innerhalb jeder Zeile wurden die Gewichte so normalisiert, daß sie sich zu eins aufsummieren.

9Dies zeigt sich sowohl in einer Betrachtung lokaler Maße der räumlichen Autokorrelation als auch in einem starken nicht-linearen Zusammenhang zwischen den Residuen und der einfachen räumlichen Entfernung von Saarbrücken (nicht ausgewiesen).

10Zur Berechnung der Prozentwerte wurden die in der Wahlstudie bereitgestellten Repräsentativgewichte verwendet.

11Dies ist ein erster Hinweis darauf, daß sich die Wählerschaft der PDS, die sich früher zu einem erheblichen Teil aus den ehemaligen DDR-Funktionseliten rekrutierte, deutlich verändert hat.

12Die graphische Darstellung basiert auf einer zweidimensionalen Kerneldichteschätzung, d. h. die Einstufungen der Partei auf den beiden Skalen mit je elf diskreten Kategorien werden als Ausdruck einer unterliegenden Verteilung von kontinuierlichen Wahrnehmungen interpretiert. Zu den Details der verwendeten Prozedur siehe Venables und Ripley (2002, S. 130-131).

13Mit Blick auf die ökonomische Dimension versteht sich dies von selbst. Der Abschnitt zur Zuwanderungspolitik im Bundeswahlprogramm 2009 ist zwar recht kurz, aber ebenfalls eindeutig. Zuwanderung wird dort innerhalb des größeren Abschnitts „soziale Gleichstellung“ (2.8) diskutiert. Gefordert werden u. a. ein Ende der „sozialen Ausgrenzung von Migrantinnen und Migranten“, liberalere Nachzugsmöglichkeiten für Familienangehörige und gleichgeschlechtliche Lebenspartner, Sicherung des Asylrechtes, aktives und passives Wahlrecht unabhängig von der Staatsangehörigkeit, erleichterte Einbürgerung von Migranten sowie eine Lockerung der Regelungen zur doppelten Staatsbürgerschaft (Die Linke, 2009, S. 17-18)

14Die Zusammenhänge mit dem Zuwanderungsitem sind noch niedriger. Obwohl die beiden Policy-Dimensionen analytisch voneinander unabhängig sind, sind diese sogar etwas stärker miteinander korreliert (0,20 bis 0,34) als das ökonomische Policy-Item und die allgemeine Links-Rechts-Skala

15Dies erklärt sich möglicherweise daraus, daß das Item zwei Dimensionen (Steuersenkungen/-erhöhungen und Ausbau/Abbau von Sozialleistungen) miteinander verknüpft. Obwohl der tradeoff zwischen beiden in der Fragestellung explizit gemacht wird, kann dies zu Verwirrungen führen, zumal von den Parteien der Linken immer wieder weitere Alternativen (zusätzliche Schulden, Sondersteuern nur für Reiche oder Wirtschaftsunternehmen) ins Spiel gebracht werden.

16Wie verwenden hier den Mittelwert der Werte von CDU und CSU als Gesamtwert für die Union. Fehlt einer dieser Werte, so wird der jeweils andere Meßwert als Gesamtwert für die Union betrachtet.

17Die Werte für Kruskals Streßmaß liegen bei 0,04 (West) und 0,05 (Ost). Selbst mit einer eindimensionalen Lösung ließen sich Streßwerte <0,10 erreichen.

18Hier und im folgenden gewichten wir die Daten mit dem kombinierten Repräsentativgewicht für Ost- und Westdeutschland (IPFWEIGHT_GES). Alle Standardfehler wurden mit der Survey-Option in Stata 11.1 geschätzt. Dabei wurden Ost- und Westdeutschland als Strata und die „virtual sampling points“ als Primary Sampling Units definiert. Die resultierenden Standardfehler sind insofern konservativ, als sie wesentliche Elemente des Designs berücksichtigen. Verbleibende Abhängigkeiten der Residuen, die sich daraus ergeben, daß mehrere sampling points in denselben Wahlkreis bzw. dasselbe Bundesland fallen, werden jedoch nicht modelliert.

19Mit Hilfe eines logistischen Regressionsmodells für ordinale abhängige Variablen (nicht tabellarisch ausgewiesen) läßt sich zeigen, daß statistisch signifikante Ost-West-Unterschiede nur in der mittleren und der obersten Kategorie auftreten.

20Bei Verwendung des Survey-Schätzers in Stata ist die Likelihood-Funktion nicht definiert, so daß keine Anpassungsmaße ausgegeben werden. Ein äquivalentes Modell, das die Gewichtungsvariable, nicht aber die Korrelation der Fehlerterme berücksichtigt, erreicht Pseudo-R2-Werte im Bereich von 0,16 (Cox-Snell) bzw. 0,06/0,10 (McFadden korrigiert/unkorrigiert).

21Alle Wahrscheinlichkeitsaussagen beziehen sich auf die Schätzungen, die sich aus dem Modell ergeben.

22Die Wahrscheinlichkeiten und ihre Konfidenzintervalle wurden mit dem margins-Befehl in Stata 11.1 geschätzt. Die Schätzungen der Standardfehler sind konservativ, weil sie die Design-Effekte berücksichtigen und sich jeweils auf die entsprechenden, teils recht kleinen Sub-Populationen (z. B. Grünen-Anhänger in Ostdeutschland) beziehen. Innerhalb dieser Sub-Populationen wurde das Merkmal „Stärke der Parteibindung“ von 0 bis 4 variiert. Die Verteilung der übrigen Variablen (Bildung, politisches Interesse) entspricht der realen Verteilung in den Subgruppen.

23In den beiden Wellen der GLES wurden Informationen zur Bewertung von Angela Merkel, Frank-Walter Steinmeier, Renate Künast, Guido Westerwelle, Gregor Gysi, Oskar Lafontaine, Karl-Theodor zu Guttenberg erhoben. In der Vorwahluntersuchung wurde zusätzlich die Sympathie gegenüber Horst Seehofer erfragt. Um möglichst viele Fälle verwenden zu können und das Modell nicht zu überfrachten, ignorieren wir die Aussagen zu Seehofer. Die Sympathiewerte von Gysi und Lafontaine sind in beiden Landesteilen recht hoch (im Bereich von 0,6 bis 0,7) miteinander korreliert. Sofern beide Politiker bewertet wurden (was rund 95% aller Fälle betrifft), bilden wir deshalb den Mittelwert beider Beurteilungen. Etwas komplizierter ist die Situation der Union, da die CSU eine selbständige Partei ist und auch so wahrgenommen wird. Wir kombinieren deshalb die Beurteilungen für Merkel und zu Guttenberg zu einer neuen Variable, die in Bayern der Bewertung zu Guttenbergs und außerhalb von Bayern der Bewertung von Merkel entspricht. Fehlende Werte für Merkel bzw. zu Guttenberg werden durch die Werte des jeweils anderen Politikers (sofern vorhanden) ersetzt.

24Angesichts der in Abschnitt 3.2 dokumentierten Unsicherheiten bei der Einordnung der Parteien haben wir sowohl auf die Verwendung der spezifischeren Skalen als auch auf die Berechnung von Distanzen zwischen Befragten und Parteien verzichtet.

25Es wäre naheliegend, weitere sozialstrukturelle Variablen wie die Konfession, die Kirchgangshäufigkeit und die Berufsgruppe mit in das Modell aufzunehmen (Arzheimer und Schoen, 2007). Dies ist jedoch nicht zwingend notwendig, da über die Ideologie und die Parteiidentifikation bereits ein großer Teil der (sozialstrukturell vermittelten) Orientierungen abgedeckt ist. Um das Modell einigermaßen übersichtlich zu halten, verzichten wir deshalb auf diese Variablen.

26Ein äquivalentes (vgl. FN 20) Modell erzielt Pseudo-R2-Werte von 0,769 (Cox-Snell) bzw. 0,385/0,437 (McFadden korrigiert/unkorrigiert).

27Die Punkte in der Grafik zeigen jeweils zehn (West) bzw. 20 (Ost) Prozent der Beobachtungen. Da von den Skalen nur ganzzahlige Werte erfaßt werden, wurden die Positionen zufällig variiert, um die einzelnen Punkte sichtbar zu machen. Die Kurven sind nicht-lineare Dichteschätzer (lowess) mit einer Bandbreite von 0,8 und wurden über die Gesamtheit der ungewichteten Ausgangsdaten berechnet. In den Kurven für die Kandidaten von Union, SPD und FDP zeigen sich in Einklang mit den theoretischen Erwartungen recht deutliche Hinweise auf ein kurvilineares Muster. Beispielsweise scheint Westerwelle sehr rechten ostdeutschen Wählern nicht rechts genug zu sein. Ebenfalls auffällig ist die insgesamt größere Popularität der Spitzenkandidaten der Linkspartei in Ostdeutschland sowie der trotzdem recht steile Verlauf ihrer Popularitätskurven über das ideologische Spektrum. Für die hier gewählte Analysestrategie sollten diese Zusammenhänge unproblematisch sein.

Lakatos reloaded. A reply to Lister


Michael Lister has responded to my critique of his work with what Lakatos (1980) called a “problem shift”. In his original paper, Lister (2007) presented empirical evidence that selective welfare state institutions undermine norms of solidarity, which in turn depresses turnout in general elections. In my comment (Arzheimer 2008), I demonstrate that, while the hypothesis is certainly plausible, the evidence itself is inconclusive at best. From my reading of Lister’s reply, I take it that we are in agreement on this point at least. Our disagreements (or mutual misunderstandings?) relate to three different issues, which I will address in turn.

The Status of Social Norms and the Problem of Ecological Fallacy

Lister (2009) argues that I do “great violence” to his position by oversimplifying his presentation of the causal chain that links a) institutions, b) social norms, c) individual (internalised) norms, d) individual voting behaviour and e) turnout, and also by falsely accusing him of committing an ecological fallacy. In my bid to develop a stylised representation of his argument (Arzheimer 2008: 682), I do indeed simplify his sophisticated theoretical account and omit social norms, and I do so for two good reasons: first, in the reductionist spirit of Almond and Verba (1965), it is not clear how social norms are different from the distribution of individual attitudes; and second, even if we assume that social norms do have an independent ontological status at the macro level, we have no independent measure for them and cannot, therefore, model their relationship with institutions. As for my “false accusation”, Lister (2007) studies a correlation between two aggregate variables, inequality, as a proxy for institutions, and turnout, which he then interprets in terms of individual agency. That approach, by definition, constitutes an ecological fallacy.

A quantitative impasse?

In my comment on Lister’s original article, I point out that institutional variables tend to vary a lot between countries but are often fairly stable over time within countries. As a result, it is often extremely difficult to disentangle institutional and country effects. Responding to this point, Lister (2009) concludes that a “statistical evaluation of this [i.e. his] hypothesis is virtually impossible”, but this is not major problem because his original results “should be seen as ‘indicative’” anyway. I could not disagree more strongly.

First, new statistical methods have been developed that are better suited to dealing with “weak” datasets (see the references in my comment). Second, over the last three decades or so, waves of democratisation have given us many more (varied) observations, which most studies have so far ignored. Moreover, political scientists and welfare economists have begun to tap into the power of experimental surveys that allow us to confront citizens with a whole range of alternative hypothetical welfare-state arrangements. The quantitative analysis of such experiments at the micro level is a useful complement to macro-quantitative information.

Finally, and perhaps most importantly, we should remember that the whole purpose of quantitative analysis is to test competing hypotheses so that the most plausible can be identified, given the data at hand. If one is not interested in this type of reality-check, there is little point in conducting quantitative analysis.

Qualitative vs Quantitative Analysis of the Institutions-Attitudes link – a false dichotomy

In his response to my comment, Lister suggests that the alleged link between institutions and attitudes should be studied by applying “more qualitative techniques, such as case studies”. He then briefly compares the development of the welfare state and levels of social trust in Sweden and the United States.

Following King, Keohane and Verba (1994: 4), qualitative research does not rely on numerical measurement but aims at providing more detailed information on the phenomenon in question. However, the same logic of inference associated with quantitative research and the same general rules of scientific inquiry apply. In this instance, an appropriate qualitative research design might include a detailed analysis of welfare-state legislation (at the macro level), accompanied by some carefully selected in-depth interviews with Swedish and American citizens that illustrate their experience of the welfare state and their more-or-less-trusting relationship with other citizens.

What we get instead is the information that some benefits introduced under President Johnson were effectively abolished under the Reagan and Clinton administrations so that the system remained highly selective, that cuts were made in Sweden during the 1990s without changing the universal nature of the system, that social trust has fallen in the US from 53% in 1963 to 35% in 1999 and that  levels of social trust in Sweden are extremely high at present and were presumably slightly lower in the past.

How is this different from the original quantitative analysis? The institutional information could easily be translated into an ordinal or quasi-interval measure for universalism (from low to extremely low in the United States, and from extremely high to quite high in Sweden), while the information on social trust comes directly from quantitative population surveys such as the World Values Survey.

The dependent variable has thus changed from turnout to social trust, yet we are still left with an opaque aggregate correlation that could have been caused by any number of processes at the micro level. The biggest difference between this new information and the original analysis is that Lister has dropped all control variables and reduced the number of observations from about 130 to 4.

Lakatos famously differentiated between “progressive” and “degenerative” problem shifts. Whether the switch from a time-series cross-sectional design to a case-study design falls into the former or the latter category is a judgement best left to the reader.


Almond, Gabriel A. and Verba, Sidney, 1965: The Civic Culture. Political Attitudes and Democracy in Five Nations. Boston, Little, Brown and Company.

Arzheimer, Kai, 2008: ‘Something old, something new, something borrowed, something true? A comment on Lister’s ‘Institutions, Inequality and Social Norms: Explaining Variations in Participation’’, British Journal of Politics and International Relations, 10,  681-697

King, Gary, Keohane, Robert O. and Verba, Sidney, 1994: Designing Social Inquiry. Scientic Inference in Qualitative Research. Princeton, Princeton University Press.

Lakatos, Imre, 1980: The Methodology of Scientific Research Programs. Cambridge, Cambridge University Press.

Lister, Michael, 2007: ‘Institutions, inequality and social norms: Explaining variations in participation’, British Journal of Politics and International Relations, 9, 20–35.

Something old, something new, something borrowed, something true? A comment on Lister’s ‘Institutions, Inequality and Social Norms: Explaining Variations in Participation’


1 Introduction

During the last 15 years, the (aggregate) analysis of electoral turnout in liberal democracies has become a minor industry. A recent survey of the relevant literature (Geys2006) lists not fewer than 83 empirical studies that relate turnout to a plethora of institutional, political and social factors. Amongst these, population size, the closeness of the respective contest, and (a rather less surprising finding) compulsory voting emerge as the most important independent variables. Michael Lister’s (2007) recent article in this journal is a valuable addition to this discussion, because by focusing on social inequality, he draws our attention to a whole host of societal factors that have by and large been neglected so far. Moreover, Lister’s contribution is one of the few studies that analyses turnout over time and in a cross-national perspective, whereas the majority of the analyses looks at subnational units, often in a cross-sectional perspective.

There are, however, a number of methodological and substantive issues with Lister’s analysis that call the validity of his findings into question: First, Lister’s account of causal relationships is highly problematic, second, the methodology is not appropriate given the low and unequal number of elections per year, third, most variables in the model are constant or near-constant within countries, and forth, even if there is a statistically significant relationship between inequality and turnout, it is trivial. In what follows, I will address these points in turn.

2 Is there an effect of inequality on turnout?

2.1 Causality

Macro level welfare state institutions PICPICMicro level internalised norms/expectations electoral participation (f )(b)(d)(a)(c)(e): +/- ? Squares represent observed variables, ovals represent variables for which there are no data. The solid line is the single observed relationship, dashed lines represent hypothesised, the dotted line represents the confounding effects of inequality on attitudes.

Figure 1:Observed and unobserved variables and relationships in the causal chain

Lister’s central argument is that the institutions of the welfare state shape citizens’ expectations (or norms) and thereby their political behaviour (Lister2007, 25). More specifically, he argues that welfare state institutions which are based on universalist principles provide ‘more support for norms of solidarity’ (Lister2007, 25). These norms encourage electoral participation both directly and indirectly. Means-tested welfare programs, on the other hand, have opposite effects (Lister2007, 25). Building on Coleman’s (1994, 7) framework for sociological explanation, his argument can be reconstructed in a slightly simplified way by employing three causal statements (see Figure 1):

  1. Features of the welfare state (a macro-level variable) affect internalised norms and expectations, i.e. individual attitudes.
  2. These individual attitudes have an impact on an individual’s decision to participate in a national election
  3. These individual decisions constitute turnout, another macro-level variable

Amongst these, only the third statement is unproblematic since it involves a purely mechanical aggregation (given that in liberal democracies, people are normally not prevented from voting in any systematic way). Statements (a) and (b) on the other hand are rather bold claims about the consequences and antecedents of individual variables which can never be proven right or wrong on the basis of macro-data. Ever since Robinson (1950) published his famous paper on ecological correlation, social scientists have struggled with the problem of ecological fallacies, i.e. the impossibility of deriving valid conclusions about individual behaviour from the aggregate measures.

Even the most advanced statistical techniques in the field that aim at making probabilistic statements about the likely strength of relationships between micro-level variables (say race and voting in a two-party competition, see King 1997) rely on information about the distribution of micro-level variables on the macro-level. In the absence of such information on the distribution of individual norms and expectations, nothing can be said about the validity of statements (a) and (b). Moreover, the analysis presented in Lister relies on another unobserved relationship, namely the causal connection between the institutions of the welfare state and inequality between households (d). While the nature of the welfare state’s institutions at any given point t (‘universalistic’ vs. ‘liberal’ or ‘minimal’ arrangements) will arguably have a substantial effect on inequality between households at t, it will hardly completely determine the current level of inequality. Rather, a whole host of other factors including the global and the national economy, the system and level of taxation, the previous level of inequality at points t – 1,t – 2,… as well as the previous nature of welfare state institutions and all sorts of unintended consequences and side-effects of previous policy will affect the current level of inequality, making this a rather crude measure of welfare state arrangements. Finally, over and above serving as an indicator for welfare state arrangements, inequality in itself can easily have a positive or a negative impact on one’s internalised norms and attitudes, thereby either masking or exaggerating the importance of causal effects that work along path (a). On the one hand, very low and falling levels of inequality (which are presumably associated with very high tax rates) could encourage the parties of the centre and the right to mobilise the middle classes, which would ceteris paribus lead to an increase in turnout. On the other hand, a high level of inequality would provide the working class with an incentive to vote in order to achieve a more comprehensive welfare state — this is the logic of the ‘democratic class struggle’ (Anderson and Davidson1943).1

To summarise, while Lister’s article builds on a complex framework involving three aggregate and two micro-level variables, claims about three causal relationships which are crucial for the argument and a fourth which can potentially distort the results are not and cannot be backed by data. Therefore, any conclusions from the analysis are confined to claims about the relationship between inequality and turnout on the aggregate level (i.e. ‘polities which face a high level of inequality will ceteris paribus experience a high/lower turnout than those with a more egalitarian distribution of resources’).

2.2 Data and Modelling

2.2.1 Data

The analysis presented by Lister relies chiefly on a single source which is in the public domain: the ‘Comparative Welfare States Data Set’ (Huber et al.2004) that compiles information for 18 countries from a variety of sources, covering the time-span from 1960-2000. This data set provides information on turnout (VTURN, drawn from Mackie and Rose 1982 and the reports in the European Journal of Political Research) as well as on a number of control variables such as a (chain) index of GDP per capita (RGDPC), the strength of bicameralism (STRBIC), the presence and strength of federalism (FED),2 the proportionality of the electoral system (SINGMEMD), and whether the respective country has a presidential system (PRES). These are merged with information on the focal independent variable income inequality (INEQ), which comes from the University of Texas Inequality Project (2004), and a report on compulsory voting (COMPVOTE) which was compiled by IDEA (Gratschew2001).

While Lister dismisses welfare state spending data very quickly, comparing income inequality across time and countries is full of pitfalls (Atkinson and Brandolini2000). There is no discussion of the quality and particular features of the data from the University of Texas Inequality Project whatsoever, and alternative data sources such as the Luxembourg Income Study (Atkinson2004) that might well be better suited for the research problem at hand are not even considered.

More generally relying on data sets in the public domain has clear advantages in terms of availability and replication, yet it imposes equally clear restrictions on the selection of observations and the time-frame (1963-93), which are not addressed in the text. This notwithstanding, it would have been helpful to discuss the rationale for not including Norway and New Zealand in the analysis, although these countries are covered by the sources, since the inclusion/exclusion of a single country can substantially affect the results of the regression model (see section 2.2.3 below).3

Variableβxβ ×x
Table 1:Mean turnout rate implied by Lister’s model

To replicate Lister’s findings as closely as possible, a data set (available from hdl:1902.1/10558 UNF:3:2UNq+CMPvmjb7Aat9NvpKw==) was constructed in the following way: from the Comparative Welfare States Data Set, the 15 countries under study were selected. For those, all election years between 1963 and 1993 were retained (averaging over the 1974/1982 elections for the UK and Ireland), resulting in 136 observations with non-missing values for turnout, federalism, presidentialism, bicameralism, and GDP per capita.

Best efforts notwithstanding, it proved impossible to reproduce Lister’s findings exactly, although the differences are fairly small.4 More troubling is the fact that at first glance, the coefficients reported by Lister do not seem to sum up. If one plugs in the mean values5 for all the independent variables, the expected turnout is a staggering 161 per cent (see Table 1). However, the magnitude of the coefficients and the constant reveals that the author has converted turnout6 from its original percentage scale to a relative frequency scale and then applied a logit transformation to the new variable to account for the fact that the dependent variable is bound to the interval [0;100].7

Accordingly, lagged turnout (VTURNLAG) was constructed by taking the turnout from the previous election for all election years as outlined in footnote 4 while LOGITVTURN was constructed as ln (-TURNOUT-∕100-)<br /><br /><br />  1-(TURNOUT∕100). Then, a variable was created that reflects Bingham Powell’s assessment of the proportionality of each electoral systems.8 Information on compulsory voting is a dummy variable which takes the value ‘1’ for Australia, Belgium, and Italy and ‘0’ for all other countries. Finally, information on inequality from the University of Texas Inequality Project (2004) was matched to the data set. Since this information is missing for four French, one Italian and two British election years, the number of observations is further reduced to 129 observations.

2.2.2 Inappropriate Methodology

The data constitute a Time-Series-Cross-Sectional (TSCS) or panel arrangement: (very short) time series from n = 15 countries are pooled and analysed jointly. In political science, this design became extremely popular after Beck and Katz (19951996) suggested that the familiar Ordinary Least Squares (OLS) estimator can be applied to this specific data structure as long as the standard errors are ‘panel corrected’ (PCSE) to account for the dependencies amongst observations.

Following Beck and Katz (1995, 636), a generic TSCS model can be written as

y=x ☐ +ε    with  i=1,...,N;t=1,...,T<br /><br /><br />  i,t   i,t    i,t<br /><br /><br /> (1)

where y is the dependent variable, x is a vector of independent variables (including the constant), ε is a random error term, and observations are indexed by unit/country (i) and time (t). With TSCS data, the standard assumptions of regression analysis are likely to be violated as one would expect the ε to be ‘non-spherical’ that is, contemporaneously correlated, heteroscedastic, and serially correlated (Beck and Katz1995, 636), rendering ‘raw’ standard errors invalid and thereby giving rise to confidence intervals that are too narrow and significance tests that are too generous.9

(Positive) contemporaneous correlation, which is a consequence of unobserved factors increasing or decreasing turnout in several countries at the same time, cannot be ruled out completely but is unlike to pose huge problems for the analysis since the unit of observation is the national election.10 On the other hand, heteroscedasticity (the variance of ε is greater in country i than in country i + 1) is very likely to occur in the case of turnout: in countries where voting is compulsory, the variance of ε is bound to be smaller than in other polities.

Finally, the presence of (positive) serial correlation (the impact of some unobserved factor that affects turnout in country i at time t will still be felt at t + 1 and possibly t + 2,t + 3,…) can be taken for granted. Like most analysts of TSCS data, Lister follows the suggestions by Beck and Katz and accounts for this problem by including the lagged dependent variable (LDV).

This leads, however, to an intricate complication. The length of the election period varies both over time and across countries, e.g. it is fixed at four years in the US, comes empirically very close to the same value in the UK and in Italy, and varies between one year and five years in Canada (where there is a moderate upward trend in the duration of the election period). As a consequence, the autocorrelation of ε will also vary across countries and over time. The approach chosen by Lister does not deal with this, and even if it did, estimating a multitude of autocorrelations poses obvious problems, especially given the low N and T. While one could hope that the inclusion of the LDV somehow ameliorates the situation, exactly the same problems apply to the coefficient for the LDV which should again vary with the length of the election period. Therefore, any findings should be interpreted with extra caution.

INEQ-0.024 **-0.024-0.021
Constant-2.130***-2.130*-2.516 **

* p<0.05, ** p<0.01, *** p<0.001.
Instead of standard errors, t-values are given in brackets to maximise the comparability with Lister’s findings. PCSE were estimated using the xtpcse procedure in Stata 10 with the casewise option for the computation of the covariance matrix. The number of replications for the bootstrap is 200. GEE estimates assume a first-order autoregressive process for the errors. GEE standard errors are based on the ‘robust’ (Huber-White) estimate for the variance.

Table 2:A replication of Lister’s model with Panel Corrected and Bootstrapped standard errors

Given the likely presence of heteroscedasticity and autocorrelation, applying the corrections outlined by Beck and Katz seems to be a sensible strategy at first glance. However, the approach by Beck and Katz was developed for balanced panels consisting of say 10 to 40 time periods (Beck and Katz 1995, 640-642; Beck 2007, 97). In the data set compiled for this analysis, T ≥ 10 in only six countries with a maximum of 13 in Denmark while the number of panel waves is just 7 to 9 in five other countries and extremely low (4 to 6) in four other countries. Under these conditions, PCSEs are not guaranteed to perform well (see Shor et al. 2007 for a review of the associated problems).

As a simple safeguard, a non-parametric bootstrapping procedure (Efron and Tibshirani1993) was applied, that is, 200 samples of n = 129 were drawn from the original data set (with replacement), and the analysis was repeated for each of these samples, thereby simulating the process that generated the data. Since each of these samples is slightly different from the others, the parameter estimates will vary, too. This variation generally provides a realistic approximation for the standard error in circumstances where the distributional assumptions might not hold. The results are shown in column 2 of Table 2. Compared with the first column, the t-values are substantially reduced, rendering all independent variables except compulsory voting and the LDV insignificant.

However, the unequal and rather low T suggests an alternative robustness check with an estimator that does not rely on the time-series nature of the data. Amongst the host of estimators available for panel data, Generalised Estimating Equation Models (GEE) have recently gained prominence in political science because they can accommodate complex structures for the correlation of ε and are fairly robust against misspecification (Zorn2001).11 As it turns out, this method yields almost identical point estimates, and again, compulsory voting and the LDV emerge as the only variables with statistically significant effects (column 3). The upshot is that the calculation of PCSE in Lister’s original analysis of the turnout data is not appropriate and leads him to overconfident conclusions.

2.2.3 Unit Effects and Lack of Variation over Time

But there are even more fundamental issues with this analysis of turnout and inequality. First, one must be sure that the units (countries) can be pooled, i.e. that (roughly) the same slope coefficient(s) prevail(s) in all countries. In the turnout data set, a rigorous check of this assumption that would involve the estimation of country-specific models is impossible because the institutional control variables are constant or almost constant within countries.12 Second, one must check for the presence of unit effects, i.e. for country-specific intercepts.13 If units are pooled and unit effects are not accounted for, massive bias can result. For instance, if some variable x has a moderately positive effect on y within two countries A and B, and the average value of x is higher in country B while the overall level of y is higher in country A, a coefficient with a negative sign might be estimated unless country-specific intercepts are included in the model.14 Unfortunately, it is not possible to test for unit effects since the institutional variables do not vary within countries and are therefore perfectly collinear with the country-specific intercepts. Moreover, there are non-trivial linear dependencies between the independent variables: federalism and bicameralism correlate at r = 0.41,15 the correlation between federalism and proportionality is -0.4816 and even inequality and proportionality correlate at -0.44.


Figure 2:Inequality over time

To make things worse, the focal independent variable is ‘sluggish’ (Beck2001Wilson and Butler2007Plümper and Troeger2007), i.e. inequality varies a lot between countries but does not vary much within most countries (see Figure 2). While there are marked upward trends in Australia, Belgium, and the UK, and some apparently random variation in the Netherlands, inequality is largely stable elsewhere. Therefore, roughly 80 per cent of the total variation occurs between countries.

In a similar fashion, the variation of turnout within most countries is very moderate if compared to the variation between countries (see Figure 3). Turnout is consistently close to 100 per cent in Australia and Belgium (where compulsory voting is enforced) and still very high (i.e. above 80 per cent) in Austria, Denmark, Italy, and Sweden,17 whereas the figure for Japan hovers consistently around 70 per cent, and turnout in the US is permanently below or just above 60 per cent. Across the whole sample, only about 10 per cent of the total variation in turnout (or its transformation) occurs within countries while between-country differences account for the lion’s share of the variation.


Figure 3:Turnout over time

There are other issues here. Although the inclusion of the LDV was championed by Beck and Katz, the LDV is likely to cause problems. Estimates will be biased even if the errors are uncorrelated, and inconsistent in the presence of correlation amongst the errors (Ostrom1990, 62-65).18 There is a whole host of alternative dynamic specifications (Wilson and Butler2007, 106), and, as Wilson and Butler demonstrate, these can give wildly different estimates in many cases.

Yet, the most fundamental problem of the analysis at hand is this: like many (if not most) other comparative data sets, the turnout data are plagued by collinearity and a lack of intra-unit variation and are therefore not very informative (Western and Jackman1994).19 With most of the variation in both the dependent and the independent variables occuring between countries, one can be quite sure that polity-level factors have an effect on turnout, but it is not possible to disentangle the relative effects of the various variables that are constant (like federalism), do not vary much (like inequality) or are constant but not included in the model (unit effects). There is something about the US that depresses turnout while there is something about Australia that drives turnout close to its theoretical maximum, but while registration procedures and compulsory voting are highly likely suspects, it is not possible to prove that these factors are decisive.

No methodology, however advanced, can overcome this basic lack of independent pieces of information. Given this fundamental problem, it is not surprising that the estimates for the effect of inequality (and the other independent variables) are rather unstable and depend on the inclusion/exclusion of certain observations. This can be most easily demonstrated by removing all observations from a given year or country from the sample. For instance, the coefficient for inequality is reduced from -0.024 to -0.018 if the four elections in 1971 are excluded. By contrast, the coefficient goes up to -0.028 if the four observations in 1970 are excluded. The impact of excluding a single country is even more dramatic: if Austria (eight observations), a country with average inequality but high turnout rates, is removed from the sample, the coefficient rises to -0.038. Excluding Sweden (ten observations), a country with low inequality and high turnout, reduces the estimate to -0.016. Even excluding single observations can have a discernible impact on the estimates: without the Australian general election of 1993, the estimate for the coefficient is -0.028, while excluding the Dutch general election of 1971 brings it down to -0.017. In other words, removing a single observation from the sample can result in a change of the estimate that is roughly equivalent to one standard error.


Figure 4:Turnout and inequality

So is there anything at all that can be said about the relationship between inequality and turnout? The short answer turns out to be ‘no, not really’. One very basic approach is to ignore the institutional control variables as well as the potential impact of the GDP and to analyse the bivariate relationship on a per-country basis (see Fisher 2007 for a related bivariate analysis of turnout and the left vote).20 Figure 4 shows the respective scatter plots, with country-specific linear regression lines overlaid. This figure is quite revealing. Leaving aside the very low variation along both the x- and the y-axis in most countries, only five polities — Austria, France, West Germany, the Netherlands, and Sweden — display a clearly negative relationship between inequality and turnout, and even this statement requires qualification. Fitting any sort of trend to four data points (France) is obviously risky, and the variation of inequality is extremely low in Sweden, Austria, and West Germany. Moreover, the clear-cut negative trend in Austria and West Germany hinges on one outlying election respectively, which happens to be the rather unusual first election immediately after unification in Germany. This leaves the Netherlands as the only real example for the negative relationship between inequality and turnout. In all other countries, the relationship is weakly positive or close to nil.



* p<0.05, ** p<0.01, *** p<0.001.
T-values are given in brackets. PCSE were estimated using the xtpcse procedure in Stata 10 with the casewise option for the computation of the covariance matrix. Australia is the reference category.

Table 3:Regression of turnout on inequality, GDP, a LDV, and unit effects with Panel Corrected standard errors

To carry out a more formal test, one could run a final pooled regression of turnout on inequality and control for GDP as well as for unit effects (assuming that the effects of GDP and inequality are constant across countries).21 The results are shown in Table 3.22 As expected, the effect of inequality is positive but insignificant. This holds regardless of the transformation of the dependent variable (models 1-4 vs. models 5-8), the inclusion of the LDV (columns 1, 2, 5, and 6) and the specification of a (common23 ) autoregressive term for the errors (models 2, 4, 6, and 8) . Given the data at hand, the conclusion is that there is no evidence for a negative effect of inequality on turnout.

2.2.4 Trivial Effect Size

Lister’s interpretation of his findings is driven almost exclusively by the statistical significance of the coefficients, and accordingly, much of the discussion in the two preceding sections has focused on the merits of various modelling techniques, the choice of estimators, and the statistical significance of parameter estimates. However, the findings from a statistical model should always be judged in terms of their substantive implications and political relevance (King et al.2000).

After all, statistical significance is nothing but a statement about the conditional probability of observing an estimate of a given magnitude. In itself, such a probabilistic statement is not of substantive interest. First, it is easily possible that politically important effects go undetected because the respective significance test does not have enough power in small samples. Second, provided that the sample size is large, significance tests will pick up tiny deviations from the null hypothesis that are of no political consequence whatsoever. Therefore, statistical significance is entirely distinct from the substantive significance of the underlying factual claim. As a consequence, many significance tests that are routinely carried out are utterly insignificant in terms of their material implications (Gill1999).

The (disputed) statistically significance of inequality would simply imply that it is highly unlikely to come up with an estimate of this magnitude if the true value of this coefficient is exactly zero — not less, but certainly not more. It would not prove that inequality has any real-world consequence on turnout.

As it turns out, the analyses presented by Lister would not support his central argument — the institutions of the welfare state have an indirect impact on turnout — even if they were not conceptually and methodologically flawed and the estimates could be taken at face value. This is because the consequences of the alleged negative effect of inequality reported in Lister (2007) are negligible.

This fact is somewhat disguised by the non-linear transformation of the dependent variable but become readily apparent if one takes a closer look at Table 1 above. If all independent variables are at their mean, the expected transformed turnout is 1.613, that is 83.38 per cent (invlogit (1.613) × 100). If inequality is set to its empirical minimum of 27.08 (Sweden 1979), the expected turnout rate changes to 84.68 per cent (invlogit(1.710) × 100). If, on the other hand, inequality reaches its empirical maximum of 39.53 (Italy 1968), the expected turnout falls to invlogit(1.517) × 100 = 82.01 per cent — hardly a difference of any political relevance, even if it was significant in statistical terms.24

3 Conclusion

Michael Lister’s article makes a useful contribution to the (already very large) discussion on aggregate variables that foster or depress turnout by drawing attention to societal factors. But while the question of whether inequality reduces turnout in the aggregate is a relevant one, his analysis is fraught with methodological problems that call the validity of his findings into question. Firstly, the original article builds on an interesting theoretical argument about the impact of institutions on attitudes, from which a complex causal framework is derived, but Lister’s claims about causal relationship are not backed by data. Therefore, his analysis is confined to the question of whether there is a negative relationship between inequality and turnout in the aggregate. Secondly, no rationale is given for selecting this particular time-frame and sample of countries, and it is difficult to exactly reproduce the findings. Thirdly, Lister applies techniques developed by Beck and Katz to overcome the small-N problem of comparative politics. But since most of the control variables are constant within countries and highly correlated while the focal explanatory variable as well as the dependent variable are ‘sluggish’ (i.e. nearly constant within most countries), there are simply not enough truly independent observations to estimate the model specified by Lister.

By applying simple bootstrapping techniques it can be shown that the t-values reported by Lister are far too large, thereby overestimating the reliability/statistical significance of the parameters. This is confirmed by an alternative robustness test that applies Generalised Estimating Equations to the same data. Moreover, it can be demonstrated that the estimates change considerably if a single year, country or even a single observation is removed from the sample. Simple bivariate regression plots on a country-by-country basis as well as an alternative model that is less demanding than Lister’s specification confirm the assertion that there is no evidence for the supposed universal negative relationship between inequality and turnout.

Some of the problems outlined above can be traced back to Lister’s sole reliance on aggregate data. Over the past two decades, multi-level modelling techniques have been applied in subfields of political science as various as attitude formation (MacKuen and Brown1987), support for European integration (Gabel1998), recycling behaviour (Guerin et al.2001) and the vote for the Extreme Right (Arzheimer and Carter2006), and the joint analysis of individual and macro-data would seem like an obvious remedy for at least some of the issues identified in the first part of this paper. As a case in point, in a recent contribution Anderson and Beramendi (2005) regress individual turnout on a number of individual and polity-level variables and find that income inequality at the macro-level reduces the probability of electoral participation.

But like PCSE, multi-level analysis is no panacea. Even if they are jointly analysed with individual-level information, data on the institutions of modern democracies will often be inherently ‘weak’ (Western and Jackman1994), because both the number of countries and the level of institutional variation within these countries is low, time-series are short, and strong unit effects are likely to prevail. This makes it extremely difficult to identify any causal effect.

Finally, it should be borne in mind that statistical significance is unrelated to substantive relevance. Even if the estimates and standard errors in Lister’s analysis could be taken at face value, they would not support the hypothesis that the institutions of the welfare state have an impact of turnout, because the political relevance of the alleged effects of inequality are negligible.

4 Figures and Tables


Anderson, C.J. and Beramendi, P. (2005) ‘Economic Inequality, Redistribution, and Political Inequality’, Paper prepared for presentation at the conference on ‘Income Inequality, Representation, and Democracy: Europe in Comparative Perspective.’, Maxwell School, Syracuse University, available at http://www.maxwell.syr.edu/moynihan/programs/euc/May6-7_Conference_Papers/Anderson%20and%20Beramendi%20EUC%20Conference%202005.pdf.

Anderson, D. and Davidson, P. (1943) The Democratic Class Struggle (Stanford: Stanford University Press).

Arellano, M. and Bond, S. (1991) ‘Some tests for specification for panel data: Monte carlo evidence and an application to employment equations’, Review of Economic Studies, 58, 277—297.

Arzheimer, K. and Carter, E. (2006) ‘Political opportunity structures and right-wing extremist party success’, European Journal of Political Research, 45, 419—443.

Atkinson, A.B. (2004) ‘The luxembourg income study (lis): Past, present and future’, Socio-Economic Review, 2, 165—190.

Atkinson, A.B. and Brandolini, A. (2000) ‘Promise and pitfalls in the use of ’secondary’ data-sets: Income inequality in OECD countries’, Temi di discussione (Economic working papers) 379, Bank of Italy, Economic Research Department, available at http://ideas.repec.org/p/bdi/wptemi/td_379_00.html.

Beck, N. (2001) ‘Time-series-cross-section data. what have we learned in the past few years?’, Annual Review of Political Science, 4, 271—293.

Beck, N. (2007) ‘From statistical nuisances to serious modeling: Changing how we think about the analysis of time-series-cross-section data’, Political Analysis, 15, 97—100.

Beck, N. and Katz, J.N. (1995) ‘What to do (and not to do) with time-series cross-section data’, American Political Science Review, 89, 634—647.

Beck, N. and Katz, J.N. (1996) ‘Nuisance vs. substance: Specifying and estimating time-series-cross-section models’, Political Analysis, 6, 1—36.

Berk, R.A. (2004) Regression Analysis. A Constructive Critique (Thousand Oaks, London, New Delhi: Sage).

Berk, R.A., Western, B. and Weiss, R.E. (1995) ‘Statistical inference for apparent populations’, Sociological Methodology, 25, 421—458.

Coleman, J.S. (1994) Foundations of Social Theory (Cambridge, London: The Belknap Press of Harvard University Press).

Efron, B. and Tibshirani, R.J. (1993) An Introduction to the Bootstrap (New York: Chapman and Hall).

Fisher, S.D. (2007) ‘(change in) turnout and (change in) the left share of the vote’, Electoral Studies, 26:3, 598—611.

Gabel, M. (1998) ‘Public support for european integration. an empirical test of five theories’, Journal of Politics, 60, 333—354.

Gallagher, M. (1991) ‘Proportionality, disproportionality and electoral systems’, Electoral Studies, 10, 33—51.

Geys, B. (2006) ‘Explaining voter turnout. a review of aggregate-level research’, Electoral Studies, 25, 637—663.

Gill, J. (1999) ‘The insignificance of null hypothesis significance testing’, Political Research Quarterly, 52, 647—674.

Gratschew, M. (2001) Compulsory Voting (http://www.idea.int/vt/compulsory_voting.cfm (04.06.2007): IDEA).

Guerin, D., Crete, J. and Mercier, J. (2001) ‘A multilevel analysis of the determinants of recycling behavior in the european countries’, Social Science Research, 195—218.

Huber, E., Ragin, C., Stephens, J.D., Brady, D. and Beckfield, J. (2004) Comparative Welfare States Data Set (http://www.lisproject.org/publications/welfaredata/welfareaccess.htm (04.06.2007): Northwestern University, University of North Carolina, Duke University and Indiana University).

King, G. (1997) A Solution to the Ecological Inference Problem. Reconstructing Individual Behavior from Aggregate Data (Princeton: Princeton University Press).

King, G., Tomz, M. and Wittenberg, J. (2000) ‘Making the most of statistical analysis. improving interpretation and presentation’, American Journal of Political Science, 44, 341—355.

Lijphart, A. (1984) Democracies. Patterns of Majoritarian and Consensus Government in Twenty-one Countries (Stanford: Stanford University Press).

Lijphart, A. (1999) Patterns of Democracy. Government Forms and Performance in Thirty-Six Countries (New Haven: Yale University Press).

Lister, M. (2007) ‘Institutions, inequality and social norms: Explaining variations in participation’, British Journal of Politics and International Relations, 9, 20—35.

Mackie, T.T. and Rose, R. (1982) The International Almanac of Electoral History (Houndmills, London: Macmillan).

MacKuen, M. and Brown, C. (1987) ‘Political context and attitude change’, The American Political Science Review, 81:2, 471—490.

Ostrom, C.W. (1990) Time Series Analysis. Regression Techniques (Newbury Park, London, New Delhi: Sage).

Plümper, T. and Troeger, V.E. (2007) ‘Efficient estimation of time-invariant and rarely changing variables in finite sample panel analyses with unit fixed effects’, Political Analysis, 15:2, 124—139.

Powell Bingham, G. (1986) ‘American voter turnout in comparative perspective’, The American Political Science Review, 80, 17—43.

Robinson, W.S. (1950) ‘Ecological correlation and the behavior of individuals’, American Sociological Review, 15, 351—357.

Shor, B., Bafumi, J., Keele, L. and Park, D. (2007) ‘A bayesian multilevel modeling approach to time-series cross-sectional data’, Political Analysis, 15:2, 165—181.

University of Texas Inequality Project (2004) Estimated Household Income Inequality Data Se (http://utip.gov.utexas.edu/data/UTIP_UNIDO2001rv3.xls (04.06.2007): University of Texas).

Western, B. and Jackman, S. (1994) ‘Bayesian inference for comparative research’, American Political Science Review, 88, 412—423.

Wilson, S.E. and Butler, D.M. (2007) ‘A lot more to do: The sensitivity of time-series cross-section analyses to simple alternative specifications’, Political Analysis, 15:2, 101—123.

Zorn, C.J.W. (2001) ‘Generalized estimating equation models for correlated data: A review with applications’, American Journal of Political Science, 45:2, 470—490.

*Thanks to Sarah Kirschmann for research assistance and Elisabeth Carter, Harald Schoen, and two anonymous reviewers for their valuable comments and suggestions. Needless to say, the usual disclaimer applies.

1One could even argue that causality works the other way around: a constantly high level of turnout (which is indicative of a mobilised working class) forces the government to maintain a high level of welfare state protection.

2Information on federalism, bicameralism and presidentialism is drawn from Lijphart (19841999), although there is an inconsistency in the data set: Lijphart (1999, 189) codes federalism with integers ranging from ‘1’ (unitary states with no elements of de-centralisation) to ‘5’ (strong federal arrangements). In Huber et al. (2004), this scale reduced to just three integers (0-2= no/weak/strong federalism) in a not entirely transparent way. Particularly confusing is the case of Belgium, which is coded as ‘0’ until 1993 although it is assigned a value of ‘3’ by Lijphart (1999).

3Moreover, the measure for the proportionality of the electoral system (SINGMEMD) (which is apparently not drawn from Powell’s (1986) seminal paper but rather from Lijphart (19841999)) is a static index, while there is now ample evidence that proportionality is best understood as the result of the dynamic interaction between electoral rules on the one hand and the fragmentation and spatial distribution of party support on the other hand (see e.g. Gallagher 1991. At any rate, it is unlikely that an index will have a linear effect. Like federalism and bicameralism, it should probably be replaced by a series of dummy variables.

4First, including a lagged dependent variable (turnout at point t -1) would normally imply that the first wave of the panel is lost, but here it is possible to retain the first wave since turnout was recorded for the elections preceding the cut-off year of 1963. Second, the data on compulsory voting (Gratschew2001) are somewhat ambiguous with regard to Austria, the Netherlands, and Italy. Finally, the UK (1974) and Ireland (1982) held two general elections in a single year, and various solutions for dealing with this problem are conceivable.

5Calculated for those 129 election years for which complete information is available and treating Australia, Belgium, and Italy as having compulsory voting.

6Apparently, no such transformation was applied to the lagged dependent variable.

7In practice, the predicted values are well-behaved even without the transformation, and predictions based on the original and the transformed values are extremely highly correlated (r = 0.99). Whatever the transformation’s utility, while a discussion of the procedure and the rationale for its application could be relegated to an appendix, the fact that the variable was transformed should be mentioned in the article.

8Bingham Powell’s index refers to French presidential elections and classifies them as very proportional, which seems rather odd. Nonetheless, the coding scheme discussed by Lister (2007, 30) suggests that this variable was used in the original analysis. My replication data set also includes the alternative variable provided by Huber et al. (2004), but the substantive conclusions are the same, regardless of which operationalisation is chosen.

9Non-spherical errors also render the parameter estimates inefficient, but this is usually considered a minor problem (Beck and Katz1995, 636).

10While there are historical events like the oil price crises that might affect turnout in all countries in a given year, it is not easy to conceive of an error process that affects say the second election in the period of study in each or even most countries in the same way. Yet, such effects are not entirely implausible. For instance, in five countries, the second election under study was held in the eventful year of 1968.

11Another panel estimator with desirable properties was proposed by Arellano and Bond (1991). However, the Arellano-Bond estimator involves first differences of the independent variables and can therefore not deal with those regressors that are (almost) constant within panels. An Arellano-Bond regression of turnout on the dynamic variables (GDP and inequality) is therefore not fully comparable to the results in Table 2 but demonstrates again that inequality has no significant effect (not shown as a table).

12In Sweden, STRBIC changes from ‘weak bicameralism’ to ‘no second chamber or second chamber with very weak powers’ after a constitutional reform in 1970.

13A wider definition of unit effects would include country-specific slopes and error variances, see Wilson and Butler 2007, 104.

14See the figures in Wilson and Butler for graphical examples. The inclusion of the LDV in the model does not necessarily capture unit effects (Wilson and Butler2007, 107).

15Goodman and Kruskal’s γ = 0.53.

16γ = -0.64.

17West Germany is another high-turnout country. The rather low value in 1990 is actually a combined figure for both West and East Germany.

18The inclusion of the LDV also changes the interpretation of the coefficients for the independent variables because the impact of x will cumulate over time (Ostrom1990, 72-74). The situation is even more complicated here because the lagged endogenous variable was apparently not transformed. See footnote 24 for an explanation.

19A more general and almost philosophical question is whether ‘apparent populations’ should be treated as samples at all. See the exchange initiated by Berk et al. (1995) and the monograph by Berk (2004, 42-56) for a critical assessment.

20To ease interpretation, actual turnout was plotted against inequality. Analyses using the transformed variable (LOGITVTURN) lead to essentially the same conclusions.

21See Plümper and Troeger (2007) for an interesting new approach that aims at giving biased but relatively efficient estimates for the effects of slowly changing or time-invariant variables in the presence of unit-effects.

22The four French elections and the German election of 1990 were removed from the sample for reasons stated above.

23As explained above, it seems unwise to estimate panel-specific autoregressive terms.

24The issue is actually slightly more complicated because (a) Lister’s specification includes a LDV, which implies that the present effect indirectly affects future levels of turnout and (b) because the dependent is transformed in a non-linear fashion while the LDV is retained in its original scale, thereby creating a complex linear-non-linear feedback loop. As a consequence, turnout would rapidly (within 10 elections) converge towards 100 per cent if the process starts from the mean values in Table 1 and is otherwise left alone. However, this convergence depends on the initial level of turnout. Setting inequality to its maximum and thereby reducing the initial level of turnout to 82 per cent is sufficient to trigger a convergence towards a 0 per cent turnout rate, again within the course of ten elections. The decision over whether such a specification make sense substantially is left to the reader.

Fringe Parties


Fringe Parties

There is no universally accepted definition of what constitutes a “fringe party”. “Fringe party” is mostly used by journalists, politicians and political scientists as a pejorative term to demarcate the boundary between “reasonable politics” and the “lunatic fringe”, a label famously applied by Theodore Roosevelt in his Autobiography to describe “the foolish fanatics always to be found in such a [reform] movement and always discrediting it” (Roosevelt 1922, 206). Consequently, some political scientists have argued that the term should best be replaced by more neutral expressions, such as “marginal parties”, “non-established parties” or “non-mainstream parties”.

It is, however, possible to derive a set of common and interrelated characteristics of fringe parties from the way the phrase is used in political language: Fringe parties do usually attract only minor segments of the electorate, they are small in terms of party membership, their leadership does not (longer) belong to the established elite groups of their respective political system, and their party ideology does either violate the political consensus or is simply considered irrelevant by most voters.

Put differently, fringe parties are not part of their countries political mainstream, and they are not normally electorally relevant parties. This statement does, however, require two qualifications: First, most new parties (e.g. the Green parties) started out as fringe groups but became both electorally relevant and accepted by the more established parties and the majority of the citizenry over time. Second, some parties remain isolated and outside the political mainstream although they attract relatively large segments of the electorate (e.g. some Communist parties and some members of the Extreme Right party family).

Moreover, the ideological marginality of a party is not only conditional on time, but also more generally conditional on political context. Within the boundaries of a liberal-democratic regime, parties that promote a dictatorship of the proletariat or biological racism are clearly beyond the pale because their ideology contravenes the system’s most basic norms and values. Left- and right-wing extremist groups which aim to abolish or radically transform liberal democracy are therefore amongst the most prominent fringe parties in Western democracies. Within the context of a stable authoritarian system, however, a nascent grouping of democrats would well be considered a fringe party while the dominant non-democratic parties define the political mainstream.

Most fringe parties are, however, marginal not because they harbour extremist views but rather because they tend to campaign for a single issue which is not – at least not in itself – important enough to secure them sufficient levels of political support. Examples from Western Democracies include, but are not limited to:

Religious parties. Historically, religious conflicts have had an impact on the formation of European party systems during the 19th century. In post-war Western Europe, the Christian-Democratic party family has been rather successful electorally, and Christian values have had an impact on the party ideologies of many other Western parties. Today there is, however, a number of tiny Christian parties that represent fundamentalist and/or evangelical views and try to distance themselves from both mainstream churches and Christian-Democratic parties. Moreover, an even smaller number of non-Christian (mostly Islamic), spiritual and New Age parties exists in Western countries. So far, they have had no electoral success whatsoever. In other countries where religious cleavages are more prominent (e.g. India or Israel), religious parties can be much more relevant and would not automatically be considered as part of the “fringe”.

Regional and ethnic parties. In many countries, ethnic and regional cleavages are simply not salient enough to sustain a single-issue party, rendering attempts to mobilize political support on the basis of some long-forgotten territorial unit futile. However, where they exist, ethnic parties are sometimes well-integrated into the political system like the Swedish People’s Party in Finland. They might even enjoy special privileges like the parties of the Danish and Sorbian minorities in Germany, which are exempted from the five percent electoral threshold. Therefore, it would be difficult to portray these parties and their constituencies as being “on the fringe” in any meaningful way.

In other countries, regionalist or separatist movements may have started out as fringe parties. But during the revival of regionalism after the Second World War, they became relevant political players that cannot be ignored by mainstream parties. This would include many of the regional parties in Spain, the Scottish National Party or the various regionalist movements and Italy that merged to form the Lega Nord. Similarly, many regional and ethnic parties in India are too relevant to be considered as genuine fringe parties.

Social groups, specific interests and frivolous parties. There is a host of rather colourful parties that claim to speak for large segments of society like women, the elderly, or families with children. Normally, the interests of these groups are fairly well represented by mainstream parties of the left and of the right who cannot afford to ignore these groups. Consequently, women’s/feminist parties, family leagues and “grey” parties usually fail to attract relevant numbers of voters.

For more specific and concentrated interests like hunting, farming or even car driving, the incentive structure is slightly different because demands from these groups are more easily ignored by the existing major parties. In most countries, however, agrarian and similar parties were either absorbed into mainstream parties or linger at (or beyond) the border of political irrelevance. The French Hunting, Fishing, Nature, Tradition Party is a case in point.

Other parties might campaign for a single political issue which is less obviously linked to a social group but nonetheless seen as marginal by most voters. An example would be the host of tiny and ineffectual eurosceptic groups in generally europhile countries like Germany. It is, however, worth pointing out again that both the Green parties and the anti-immigration parties of the Extreme Right parties began their ascendancy as marginal single-issue movements.

Finally, there is a bewildering host of frivolous parties that exist to make fun of “real” fringe or mainstream parties, either to get access to state funding or just for the fun of it. Examples include beer-lovers parties in several post-soviet states, Canada, Germany, Norway and Poland, parties that allude to grand (and often fictional) political and religious ideas (“Imperial British Conservative Party”, “Scottish Jacobite Party”, “Church of Militant Elvis Party”), parties that exist to challenge political correctness and the establishment (the “Anarchist Pogo Party of Germany” and the “PARTY”, which campaigns for rebuilding the Berlin wall), or the many British groups that play with the word party (“Mongolian Barbecue Great Place to Party”).

Kai Arzheimer, University of Mainz


Capoccia, Giovanni, “Anti-System Parties. A Conceptual Reassessment.” Journal of Theoretical Politics 14, no. 1 (January 2002): 9-36.

Hainsworth, Paul, The Extreme Right in Western Europe. New York, London: Routledge, 2008.

Jolly, Seth Kincaid, “The europhile fringe? Regionalist party support for European integration.” European Union Politics 8, no.1 (March 2007): 109-130.

Kitschelt, Herbert, “Left-Libertarian Parties: Explaining Innovation in Competitive Party Systems.” World Politics 40, No. 2 (January 1988): 194-234.

Norris, Pippa, “Preaching to the Converted? Pluralism, Participation and Party Websites.” Party Politics 9, no. 1 (January 2004): 21-45.

Pedersen, Mogens N., “Towards a New Typology of Party Lifespans and Minor Parties” Scandinavian Political Studies 5, no. 1 (January 1982): 1-16.

Reiter, Howard L., “Party Decline in the West. A Skeptic’s View.” Journal of Theoretical Politics 1, no. 3 (July 1989): 325-348.

Roosevelt, Theodore, An Autobiography. New York: Charles Scribner’s Sons, 1922.

Voter Behaviour


Between the early 1940s and the late 1960s, four basic models of voter behavior have been proposed on which almost all studies of electoral behavior draw. These models describe how humans react to environmental factors and choose between different courses of action. Homo sociologicus (more or less implicitly) forms the basis of the approaches to voting behavior laid out in the first three parts of this entry. In contrast rational voter theory explicitly invokes homo oeconomicus through deductive reasoning. A closer examination reveals, however, that these seemingly very different approaches are in fact complementary and can be regarded as aspects of an overarching model. In the past few years this line of reasoning has become increasingly present both in social-psychological as well as rational choice writings.

The Micro-Sociological Model

The micro-sociological model was developed in the early 1940s by Paul F. Lazarsfeld and his colleagues and its formulation in The People’s Choice first appeared at the end of World War II. A milestone of modern electoral research, itwas also criticized for its methodological and empirical deficiencies and these critiques informed the design of the follow-up study Voting.

The motivating question for Lazarsfeld and his colleagues can be found in the sub-title of The People’s Choice; how do voters develop concrete vote intentions over the course of an election? Lazarsfeld et al. investigated this question by conducting an intensive study of Erie County, Ohio during the 1940 presidential election. They interviewed a representative sample up to seven times over the course of the campaign with regard to vote intention, their evaluation of the candidates and assessment of the major political issues. By doing so the researchers sought to determine how each individual voter developed their political attitudes over time and the impact of the campaign on this process.

Lazarsfeld et al. rapidly determined that socio-structural variables, above all socio-economic status and religious affiliation, strongly influenced vote intention for both major American parties. Taken together with a voter‘s living situation (urban or rural), the researchers constructed an “Index of Political Predisposition” with an extremely accurate predictive capability. Blue-collar workers and Catholics disproportionately trended toward the Democrats while Protestants and middle class voters predominantly supported the Republicans, with the interaction of both variables strengthening these effects.

With muted reservations, the authors concluded that the political preferences of their respondents were largely socially determined. For many voters, party choice was fixed months before the election and new information was used selectively to reinforce rather than challenge or update prior opinions. These findings were far removed from the ideal of the responsible democratic citizen, painstakingly informing themselves about the various parties and candidates before coming to a decision based on sober reflection.

As they conceded, however, Lazarsfeld et al. could only tentatively explain why socio-structural variables influenced vote choice so strongly, despite the relative anonymity of individual members to these large, impersonal structures. The authors argued implicitly that socio-structural variables could be viewed as indicators of membership in a mostly homogenous social environment of friends, family, neighbors and colleagues with similar political views. This web of interactions is then capable of reinforcing wavering individual opinions through social pressure. In these circumstances so-called Opinion Leaders played an important role by intensively informing themselves about political events through the media and then passing their observations on to less interested or educated citizens. To describe this relationship Lazarsfeld et al. formed their famous „two-step-flow“ hypothesis of political communication.

The Columbia-Group’s emphasis on the immediate social environment disposed them to observe an interesting phenomenon: If a voter‘s social environment is not homogenous and they belong to multiple social groups with incompatible political norms, conflicting behavioral expectations (cross pressures) should develop. To explain non-voting or party-switching, two phenomena that electoral researchers have always been pre-occupied with, Lazarsfeld et al. were forced to rely above all on cross pressures in the immediate social environment.

The Macro-Sociological Model

In contrast to the Columbia study, the macro-sociological approach focuses its explanations on processes at the level of the entire society. In Germany this approach was initially forwarded by M. Rainer Lepsius who was primarily occupied with “social-moral milieus”, a key characteristic of German society in the Imperial and Weimar periods. Internationally Lepsius had little impact, while even within the German literature his approach was soon displaced by a competing macro-sociological model that argued from the outset with abstract categories, was tailored to explain a larger area (Western Europe) and was easily portable to other contexts. This model was the “cleavage” theory of Seymour Martin Lipset and Stein Rokkan, originally formulated in the comprehensive introductory chapter of their work Party Systems and Voter Alignments.

By “cleavage” Lipset and Rokkan mean a social “fault line”, a sustained social conflict pitting (at least) two large groups with conflicting social interests (and defined by their social characteristics) against one another. According to Lipset and Rokkan, European social conflicts can be systematically ordered and divided into four groups:

  1. Conflicts between the national Center and the subordinate Periphery,

  2. Conflicts between the State and the Catholic Church;

  3. Conflicts between Urban and Rural territories; and

  4. Conflicts between Labor and Capital.

These four conflicts ultimately go back to processes of modernization. The first two refer predominantly to the cultural sphere and hearken back to the development of modern nation-states, while the latter conflicts are above all economically motivated and consequences of the Industrial Revolution.

For Lipset and Rokkan, social conflicts become politically relevant if a specific set of conditions is fulfilled:

  1. The conflict must remain virulent over a long period and play a central role in the life of the affected individuals.

  2. Social mobility must be low, so that one typically remains a lifelong member of the relevant social group.

  3. Those affected by the conflict must have the motivation and opportunity to ensure that their interests are incorporated into formal associations.

  4. The leaders of these “pressure groups” must found their own party or agree to some form of coalition with a pre-existing party.

  5. This party must have an opportunity within the electoral system to cross the threshold of parliamentary representation.

Under these conditions social conflicts achieve a sort of political reification. The parties that develop are understood as the agents of social groups and are treated as such by group members. The format of the party system that develops, such as the number of parties or polarization between them is determined by the number of relevant social cleavages and whether these fault lines run parallel to or overlap with one another. So long as the system of social conflicts remains stable, for example when parties negotiate a lasting compromise that is also acceptable to their represented social groups, the party system will remain fundamentally stable.

Lipset and Rokkan’s unpacking of the relationship between social structures and the party system is highly internally consistent and constitutes a powerful analytical frame, in that prior findings on voting behavior are easily integrated into a cleavage theory. An obvious deficiency in their model, however, is the failure to consider the individual level and the role of communication. Lipset and Rokkan are silent on why individual voters usually behave empirically as elites expect them to.

It is possible to close the micro-level gap and integrate Lipset and Rokkan’s macro-sociological model with the complimentary micro-sociological findings of the Lazarsfeld-Group and Lepsius’ work at the meso-level. But even this combined approach has a major deficiency, as it only poorly explain moments of political change. For relatively short-term fluctuations in the relative strength of political parties, which lead to relatively frequent changes in the size and composition of governing parties or coalitions, the picture of an ideal-typical homo sociologicus blindly adhering to the norms of his reference groups is unsatisfying. The socio-psychological model represents a solution to this problem and its findings are highly complementary with the previous sociological models.

The Socio-Psychological Model

Ten years after The People’s Choice Angus Campbell and his associates at the Survey Research Center published their first major election study. The Voter Decides was distinct from the Lazarsfeld-Group’s work in two respects. First, Campbell and his co-authors conducted a random sample covering the entire United States, as opposed to prior regionally limited inquiries. Second, Campbell et al. initially explained voting behavior exclusively through psychological variables, specifically the evaluation of candidates, their positions on the major political issues and their “party identification”, the degree of attachment to a political party. Initially all three psychological variables were considered equally important. Sociological variables, of primary importance to Lazarsfeld et al., were held in The Voter Decides as exogenous and remained unconsidered.

Campbell et al. were initially and correctly criticized for almost fully ignoring the social context of vote choice and for relying on variables so temporally and empirically connected with the act of voting that the model risked tautology. Campbell et al. reacted to these critiques by extending their inquiries to include the broader social context, a social-psychological model of behavior that would become identified with the University of Michigan as the “Ann-Arbor-Model”. Put forward in the American Voter, this new model demonstrated its effectiveness by using surveys from the 1952 and 1956 elections. The response from the scientific community was overwhelming: The American Voter became one of the most influential monographs in the history of electoral research and the Ann-Arbor-Model dominated the study of voting behavior in western democracies for many years after its publication.

The American Voter deviated from the Michigan-Group’s past work in two respects. First, party identification was now taken as a long-term stable variable, causally prior to individual evaluations of candidates and political issues. Second, psychological variables were no longer taken as given, but rather were seen as influenced by a voter’s sociological background. This included the experiences of an individual’s reference groups with the various parties and an integral role for the strengthening or weakening of opinions through a voter‘s immediate social circumstances. The Ann-Arbor-Model can therefore be considered an extension of sociological theories of voting behavior.

It is often overlooked, however, that the American Voter describes a wide range of potential influencing factors, in later years were seen as alternatives to the social-psychological model. These factors include for example the institutional context, economic situation or personality structure of voters.

Using the famous image of the “causality funnel” Campbell et al. summarized the relationship of all these varied factors. Individual vote decision is understood as the result of a complex process principally traced far back into a voter’s past. At the moment of the vote decision itself, only the previously identified psychological variables are of interest. The further one moves back into a voter’s history, the more potential influences have to be taken into account to explain the final behavior. The causality funnel therefore expands into the past until a complexity of factors is reached that the researcher can no longer untangle.

Despite its evident advantages, a dispute arose over the next several years concerning the Ann-Arbor-Model’s portability outside of the American context. Especially problematic for the Michigan-Group was the central concept of party identification: The idea of a “psychological membership” seemed too dependent on the peculiarities of the American system, particularly the (relatively) stable two-party system, the organizational weakness of the major parties and the absence of historically-based ideological conflicts.

In an influential contribution Russell Dalton, Paul Allen Beck and Scott C. Flanagan showed that long-term stable party identification did not imply a “psychological party membership”. The characteristic coalitions between social groups on the one hand and ideological parties on the other in Europe, as set out by Lipset und Rokkan, could instead be considered the functional equivalent of party identificationdescribed in the American Voter.

The Rational Voter Model

The theory of rational voting goes back to Anthony Downs’ pioneering study An Economic Theory of Democracy. Building on prior work by Kenneth Arrow, Joseph Schumpeter, Herbert Simon and others, Downs applied the tenets of neo-classical economics to voting behavior and provided the impetus for a new and fruitful research agenda in political science.

Downs’ concept differed from previous approaches in two respects. First, Downs’ approach to electoral research was primarily theoretical: Downs engaged in no empirical studies, but rather limited himself to deductively deriving axiomatic propositions that could be empirically tested. Second, the approach founded by Downs was based much more strongly on formal modeling than earlier approaches. While the sociological and socio-psychological approaches tied into experiences from everyday political life and seemed intuitively plausible despite their abstractions, the rational choice approach initially struck many researchers as all too artificial and unrealistic. That the contributions of rational choice theorists were often presented through systems of equations only strengthened this impression.

The starting point for Downs is the assumption that politicians and voters behave as rational actors in a market, in which political power (in the form of votes) is exchanged for the realization of political objectives. The “rationality” of actors is therefore understood in a formal sense that has nothing to do with “reasonableness” in the commonly accepted sense, but rather is solely related to the decision between alternative actions.

According to the model, rational actors possess stable and transitive preferences, which gives them the ability to select from a set of alternatives to maximize their benefits. “Benefits” are not limited to economic gains for an actor, but rather any result that is in line with their preferences. “Stable” means nothing more than that the preferences of actors remain constant during the period in question; “transitive” means that there are no contradictory or cycling preferences. A rational actor that prefers a government formed by Party A to one formed by Party B, and prefers Party B to Party C, must therefore prefer Party A to Party C given a choice between the two.

If one assumes that political programs can be placed as positions on an ideological Left-Right continuum, a rational voter will choose the party that stands nearest to their “ideal point” on that continuum (the point where their benefits are maximized). At the same time, parties will formulate their political programs with an eye toward maximizing their vote total. As the preferences of actors are seen as stable, changes in behavior are only explained through structural changes, such through the entry of a Party D.

How the “benefits” of actors persist and how their preferences come about are not discussed in rational choice models. As the preferences of actors are as a rule constructed from observations of their behavior, the rational voter approach is fundamentally tautological: The rationality postulate is an axiom rather than an empirically tested hypothesis.

But this tautological structure is precisely the greatest strength of the rational choice approach. It makes it possible to connect psychological or sociological models to the rational choice approach by treating them as mechanisms for the construction of preference sets. Thus it is possible to hold exogenous the complex and often-idiosyncratic backgrounds of individual voters in order to focus on the influence the structural features of a situation have on decision-making.

Downs himself recognized some of the complications that arose from exporting market behavior to electoral research. The most famous of these problems is the so-called „paradox of voting“: Independent of the electoral system, in a mass-democracy with millions of voters the probability of any one voter casting the winning vote is impossibly small. There is therefore effectively no relationship between a voter’s behavior and the victory of their preferred party. While it is unlikely that an actor can draw some personal (instrumental) benefit from casting their ballot, participation certainly entails costs. A voter has to spend time and/or money to build a picture about the intentions of the various parties (information costs). Furthermore, the acts of voting or voter registration often take time, which means foregoing other material or immaterial benefits (opportunity costs).

The net benefit of voting is therefore always negative and rational individuals should not choose to vote. This conclusion contradicts actual voter turnout, which consistently reached 70-80% in many democratic states in the 20th Century. Many solutions have been suggested to solve the paradox of voting, though all have their complications.

The uncertainty inherent in the act of voting is not limited to whether one’s own behavior can affect the outcome. For example, a voter cannot be certain whether she will receive their desired outcome for the actual costs spent; her party could lose the election. Uncertainty also exists regarding a party’s future actions. Even when parties intend to honor promises made during the election, changes in the general political situation could cause them to depart from their programs.

Voters in mass-democracies find themselves thus in a “low-cost situation” (Kirchgässner). Rational choice explanations typically do not delve further into the intricacies of behavior in low-cost situations, as it is already highly irrational under these circumstances for rational actors to put effort into collecting information or engaging in a cost-benefit analysis. Instead moral and expressive patterns dominate behavior in these situations; decisions are made on the basis of everyday information, group norms or fundamental ideological beliefs, which function as information shortcuts (Popkin). Rational voters therefore often behave as the sociological or social-psychological models would predict. Such considerations were already evident in Downs’ approach and stand at the center of the research agenda for some newer theories of rational voting.

Kai Arzheimer / Jürgen W. Falter

See also: Lazarsfeld, Paul F., Party Identification,People’s Choice, The (book), Survey Research Center , Voting (book), Two – Step Flow Model of Communication

Further Readings and References

Berelson, B., Lazarsfeld, P. F., & McPhee, W. N. (1954). Voting. A Study of Opinion Formation in a Presidential Campaign. Chicago: Chicago University Press.

Brennan, G., & Lomasky, L. (1993). Democracy and Decision. The Pure Theory of Electoral Preference. Cambridge, New York: Cambridge University Press.

Campbell, A., Converse, P. E., Miller, W. E., & Stokes, D. E. (1960), The American Voter. New York: John Wiley.

Dalton, R. J., Beck, P. A., & Flanagan, S. C. (1984). Electoral Change in Advanced Industrial Democracies. In R. J. Dalton & S. C. Flanagan & P. A. Beck (Eds.), Electoral Change in Advanced Industrial Democracies: Realignment or Dealignment (pp. 3-22). Princeton: Princeton University Press.

Downs, A. (1957). An Economic Theory of Democracy. New York: Harper.

Kirchgässner, G. (1992). Towards a Theory of Low-cost Decisions. European Journal of Political Economy, 8, 305-320.

Lazarsfeld, P. F., Berelson, B., & Gaudet, H. (1944). The People’s Choice. How the Voter Makes up His Mind in a Presidential Campaign. Chicago: Columbia University Press.

Lipset, S. M. & Rokkan, S. (1967), Cleavage Structures, Party Systems, and Voter Alignments: An Introduction, in: Lipset, S. M. & Rokkan, S. (Eds.), Party Systems and Voter Alignments: Cross-National Perspectives (pp. 1-64). New York, London: Collier-Macmillan

Popkin, Samuel L. (1994). The Reasoning Voter. Communication and Persuasion in Presidential Campaigns (2 ed.). Chicago, London: University of Chicago Press.


Bread and butter à la française: forecasts of the French legislative vote from regional economic conditions



Electoral forecasting is a longstanding sub-discipline of psephological research, but one which has concentrated principally on Anglo-Saxon systems, and in particular two-party systems with a straight incumbent / opposition trade-off in votes. Multiparty systems with more complex interrelationships between party vote shares have remained a minority field in forecasting work, and certainly theoretical and methodological advances in forecasting overall vote outcomes are much less common.

The simple association between economic conditions and level of incumbent support, either through election or public opinion polling, is the cornerstone of the vote-popularity (VP) function literature in its economic guise (Lewis-Beck 1988), and its applicability has been shown to be clearer in single-party government (two-party) systems than in multiparty coalition settings (Powell and Whitten 1993). Multiparty systems present a more complicated – confused even – situation of accountability for economic downturns, through coalition responsibility obscuring a target for dissatisfied voters. Moreover, in the French case, the semi-presidential “bicephalous” executive has prompted the question ‘Who’s the chef’ (Lewis-Beck 1997) with regard to allocation of economic responsibility by voters between Prime Minister and President.

However, a more conceptual problem with regard to multiparty systems is how satisfactorily to construct an economic forecasting model that allows prediction of a set of inter-related votes across multiple parties. From the Anglo-Saxon roots of vote forecasting, the zero-sum incumbent / opposition approach which lends itself well to simple linear regression modelling is inapplicable to a model which wishes to go beyond a simple bloc prediction or restrict itself to a single party prediction. This paper sets out to provide a constrained ‘seemingly unrelated regression’ (SUR) model predicting all votes in French legislative elections between 1981 and 2002, and uses this to forecast the 2007 legislative election result.

The French case

Amongst non-Anglo Saxon cases, France has increasingly become a country of interest to electoral forecasters. There is a growing tradition of election forecasting of incumbent vote (e.g. Jérôme et al 1999; Jérôme et al 2003; Auberger and Dubois 2005), as well as estimates of third-party performance (Jérôme and Jérôme-Speziari 2003; Auberger 2008; Evans and Ivaldi 2008) and second-order elections (Jérôme and Jérôme-Speziari 2004; Auberger 2005). The majority of this work provides discrete estimates for individual parties or blocs. There is little attempt to find a single model predicting distribution of the totality of vote across all parties, however. Given the archetypal multipartism present in the French system, approaches based upon the two-party adversarial models used in the United States and the United Kingdom (over-) simplify the relative party scores because of their underlying zero-sum logic.

Moreover, in common with all elections, but of specific interest in France, it is important to take into account not just voting but also non-voting. In France, abstention is regarded as ‘wasted’, as it is not factored into the valid vote upon which outcomes such as progression to second rounds are based. However, its effect in denying political parties the support of their core electorate has been perceived as a significant influence on election outcomes, not least in the 2002 presidential election where the Socialist candidate failed to reach the run-off stage due to relatively poor turnout amongst grassroots Socialist supporters (Shields 2005). Abstention takes on even greater importance in legislative elections which follow presidential races, as losing candidates in the presidential election generally see their parties lose even more support in the subsequent parliamentary race.1

In many ways, electoral patterns in France since the mid-1970s have encouraged the use of government / opposition models, given the establishment of the so-called ‘bipolar quadrille’ of Left and Right blocs, and the alternation between the two at ever legislative election from 1981 to 2007 – ‘hyper-alternance’, as it has been termed (Evans and Ivaldi 2002). However, the importance of third parties in disrupting this stable two-bloc pattern, from the late 1980s ‘til 2002, means that a unified model providing a set of predictions across all party groups is increasingly important. Even in the years of the moderate two-bloc bipolar quadrille, vote distribution amongst the two Right-wing parties (RPR and UDF) and the Socialists and Communists could not be assumed to be a simple sum of the two bloc-members on either side. Furthermore, the rise of the Extreme Right FN, and particularly of its presidential candidate Jean-Marie Le Pen, upset any premise whatsoever based upon a straight Left / Right contrast.2 Indeed, looking again at the 2002 presidential elections, the failure of the Socialist candidate to reach the second round due to high abstention was compounded by a strong showing by Le Pen, with the consequent unexpected progression of the Extreme Right candidate precisely into the ballottage.

As a case for third-party prediction, then, France and the FN / Le Pen have become compelling. To what extent has this party’s performance been predictable, and was the 2002 result an aberration or consistent with a possible model of Extreme Right voting? Given the small number of cases with which to construct a model, the notions of outliers is somewhat problematic in analytical terms, but focus on attempting to predict the FN and Le Pen’s performance in 2007 suggested that in fact the success would be repeated. The majority of predictions for 2007 saw no reason to expect anything other than a strong performance by the Extreme Right (Lewis-Beck et al, 2008). The majority of predictions were consequently wrong, with Le Pen and then the FN suffering a relative electoral rout. Models predicated upon conditions of success in 2002 fell apart in 2007.

However, despite the real-world necessity of focusing on this single candidate / party’s vote, methodologically there are issues with simply attempting to predict a single party’s vote in a multiparty setting. Whilst it may in practice be possible to predict a single party’s vote3, this does not take into account provenance of the vote or consider how the other parties’ votes relate to the outcome of the party in question. Parties do not contest elections in isolation, with votes varying purely in relation to their own electoral performance: performance of other parties must have an effect. If one wishes to look at the entirety of the vote distribution, and how this can be predicted, multiple equations may provide counter-intuitive results such as vote scores summing to more than 100%. A constrained model producing scores for all party alternatives is a more appropriate approach.

Before moving to the modelling process itself, we can briefly state the explanatory variables which we include in the model. Our predictors, which we operationalise in the following section, are from the foundations of economic forecast modelling. In France, evidence has been found of relatively sophisticated economic voting, with voters able to allocate responsibility for economic management amongst incumbents (Lewis-Beck, 1997). More broadly, the state of the national economy “matters” in politics (Auberger, 2004) and indeed influences second-order as well as legislative elections (Jérôme and Jérôme-Speziari, 2004). Our predictors present no surprises. First, we use unemployment as a key indicator of economic stability, and one closely associated with shifts in incumbent party support. Previous research in France has shown clearly shown decreases in incumbent party support when unemployment increases (Jérôme et al, 1999; Auberger and Dubois, 2005). Equally, a large literature exists linking Extreme Right party voting with this variable (Jackman and Volpert 1996, Lubbers et al. 2001, Golder 2003).4 Secondly, we look at the effect of GDP as a broader independent measure of economic success and stability. Again, this is a staple of the classic VP-function literature, though is less analysed in Extreme Right voting studies, not least because of a less clear link between a broad macro-economic indicator and extremist support than with unemployment, which taps into Extreme Right appeals to direct social precariousness. In French legislative election forecasts, the role of national economic growth is less common as a predictor, but where used, has been demonstrated as relating positively with incumbent support (Auberger and Dubois, 2005). Lastly, we also include a political variable, Left incumbency which we combine in an interaction with unemployment, to explore whether there is any difference in the way voters penalise, or indeed reward, for changes in economic fortunes under different parties – an economic ‘issue ownership’ effect which has been found in previous research using France as a case (Whitten and Palmer 1999).

Data and methods1

Even if we considerably simplify the number and structures of choices voters face in France, there are at least six different alternatives available to them: they may abstain (the reference category) or vote for the Communists, the Moderate Left (Socialist Party and allies, including the Parti Radical de Gauche), the Moderate Right (Gaullist RPR / UMP, UDF and Nouveau Centre), the Extreme Right (the Front National and splinter parties like the Mouvement National Républicain), or any other party (chiefly the Extreme Left groupuscules and the Greens, but also including regional parties and spoilt / blank ballots).5

Principally, then, we have six interrelated vote shares at the departmental level. This setup renders linear regression – still the workhorse of political science research – unfeasible. Linear regression requires a single unbounded dependent variable with errors that are distributed identically and independently. Vote shares, by contrast, are bounded to the interval [0;100]. Even if we ignore this boundary problem and try to model the six shares separately, vote shares are dependent by definition because they must always sum up to 100: if one party (say the Front National) does particularly well, the pool of voters available for other parties becomes smaller or turnout increases.

Two solutions to these problems have been suggested in the literature: Katz and King (1999) propose a complex likelihood estimator for multiparty data that is exact. Unfortunately, their method requires defining and programming one’s own likelihood functions and becomes computationally extremely demanding for more than three parties. A more viable three-step strategy was outlined by Tomz et al. (2002). First, an arbitrary option is chosen as the reference category, and log-ratios are calculated, which removes the boundary problem. For instance, in Paris (75), the abstention rate was 38.9 per cent in 1988. In the same election, 9.9 per cent of those eligible to vote voted for the FN, 32.3 per cent for the moderate left, and so on. Therefore, the log-ratio for the FN = ln(9.9/38.9) = -1.37, and the log-ratio for the Moderate Left is ln(32.3/38.9) = -0.19.

Second, the effects of the independent variables on the set of transformed variables are estimated via seemingly unrelated regression (SUR), a multi-equation version of linear regression. SUR allows for errors that are correlated across equations. Third, the predictions from this model are converted back to the original scale (percentages of those eligible to vote). While the Katz-King method is more accurate in theory, the extensive analyses by Tomz et al. show that their simpler approach performs equally well in practice.

There is, however, one additional complication. While Tomz et al. model single elections, the structure of our data is time-series cross-sectional, i.e. we have (rather short) time-series of results for a “panel” of 96 departments, each with its own socio-structural composition and tradition that distinguishes it from the other districts. To incorporate this structure into our model, we specify a fixed effect for each department that reflects its voting history, political tradition and other more or less constant (over the course of two and a half decades) factors which affect the outcome of a given election. This implies that our model focuses on variation over time within departments while time-invariant differences between departments (that cannot possibly be the result of our dynamic independent variables) are taken for granted.

Another implication of the time-series aspect of our data is that errors within a department will most probably track over time. A large part of this variation will be soaked up by the fixed effects, but as an additional safeguard, we specify robust (Huber-White) estimators for the variance that account for this dependency, thereby yielding conservative standard errors.6

Our dependent variables are the vote shares and turnout rates at the departmental level from the first rounds of the legislative elections from 1981 to 2002 in Metropolitan France.7 For 1981 to 1997, these figures were compiled by Caramani (2000). Results for 2002 and 2007 (for checking the accuracy of our predictions) were taken from the Centre de Données Socio-Politiques’s website (http://cdsp.sciences-po.fr/AfficheElec.php). Unemployment figures on the departmental level come from quarterly figures collected by INSEE (http://www.bdm.insee.fr/bdm2/do/accueil/AccueilAppli). Information on GDP growth (which is collected at the national level and therefore constant across departments in any given election year) comes from Eurostat structural indicators (http://epp.eurostat.ec.europa.eu). Data are not lagged, i.e. we use unemployment rates and GDP growth for the year in which the respective election was held. We tested for possible cross-correlation between unemployment and GDP growth, and found no evidence of this.8


[Table 1 about here]

We first examine the overall model fit statistics with a nesting to examine the degree to which political tradition at departmental level accounts for party bloc support, and the improvement in fit that the economic and political indicators provide (Table 1). Overall, it is clear that the additional forecasting variables provide a substantial improvement in fit, as evidenced by the shrinking of the Bayesian Information Criterion (BIC). However, the differential change in R2 from the fixed effects baseline indicates varying political traditions as well as responsiveness to economic variation. Broadly speaking, Left parties, and particularly the Communists with almost half the variation in their vote accounted for by departmental effect, enjoy stronger baseline support in traditional Left-oriented areas (and conversely consistently low support in non-leftist departments). The Moderate Right sees around one-third of variation accounted for by this. Conversely, the Extreme Right and other parties are not determined by localised consistency, with much smaller R2.

The move to the full model shows a much different pattern across the party blocs. The Moderate blocs and the other parties see a relative improvement in fit once economic indicators are taken into account,9 but not to the same extent as the Extreme Right, which sees economic indicators improve the model fit very substantially. In accounting for Extreme Right support, then, macro-economic indicators matter more than any longer term departmental tradition. At the other end of the spectrum, the Extreme Left Communists see a much more muted improvement, particularly in comparison to the fixed effects baseline. Political tradition in its ‘bastions’ has historically mattered more for the PCF than it does for other parties – something which has been noted elsewhere in the literature on Communist support in France and elsewhere (Bell and Criddle, 1994).

[Table 2 about here]

We now turn to the parameter estimates of the regression model to interpret the above changes in more detail (Table 2). There are both intuitive and less clear effects. Recall, firstly, that all reported estimates in SUR are contrasts, in this case with the abstention reference category. Consequently, parameters should not be read as absolute values for each party grouping, but rather relative effects. Rises in unemployment see a drop in support for all the mainstream parties, relative to abstention. Conversely, with only support for the Extreme Right and oOther parties increasesing. Equally, given the contrast category in the logit, unemployment appears to reduce participation. Abstention and votes for marginal and extreme parties benefits, then, from downturns in the jobs market. Similarly, Aa rise in national GDP will increase participation, benefiting both the mainstream parties and as well as the Extreme Right party, but ERP – less easily explicable – and disfavouring marginal parties. Unlike any other party, then, the FN in France appears to benefit from both economic upturns and downturns, but dependent upon the area of the economy.

The Left incumbency effect, as might be expected, penalises Left-wing parties, but (in particular the Communists. When the Left has been in power, supporters of these parties increasingly abstain, in net terms.) Bbut it also appears to penalise the Moderate Right in a similar way. We can only speculate that this may be a general participatory malaise which depresses turnout for all mainstream parties in the ‘aftermath’ of a Left-wing government – in the cases we include in this model, Left-wing governments generally performed singularly badly in economic terms., Apparently, however, although the mainstream Right opposition appears not to have inspired particular confidence as an alternative!and thus support as a replacement.10

Lastly, we turn to the interaction effect between unemployment and Left-wing incumbency effectively indicates the ‘premium’ that the latter imposes on the former’s effect. The interaction does not exhibit what we might expect it to. If we hypothesise ‘issue ownership’ of unemployment as a policy-area associated with the Left, we might expect support for Moderate Left parties to decline more heavily when they have been in government and unemployment has increased. Under Socialist governments, the electoral penalty for rises in unemployment on mainstream parties is less pronouncedower than under Right-wing governments (i.e. when the interaction effect is effectively zero, given the Right-wing reference in the incumbency variable). The same is true for Right-wing support as well, as indicated by the identical parameter estimate., and Cconversely, the Extreme Right and other parties do less well, losing more support to abstention when unemployment rises under a Left-wing government. In that respect, there does not seem to be evidence of ‘issue ownership’ of unemployment by Left governments which sees them punished more strongly than Right governments by rising unemployment.

[Figure 1 about here]
  • Maybe cut the interpretation of the parameters a bit, point out that the parameter-based interpretation of a multi-nomial model can be confusing
  • We focus on the median growth level of 1.9 per cent to keep things simple and because growth does not seem to make too much of a difference for most choices at most levels of unemployment and incumbency.
  • One of the scripts in the replication archive is easily adjusted to estimate vote shares for different growth levels.
  • That being said, a decline in the economy seems to boost the prospects of the “other” choices (though empirical support for this effect is thin) and reducing support for the FN, while strong growth seems to improve the prospects of the moderate left (if the economy is doing well, we can even afford a socialist government!) and, to a lesser degree, the prospects of the Communists.
  • Both the major left and the major right pay a price for governing (there must be a reference for this or a similar phrase which I cannot remember): ceteris paribus, support for them is lower when they are in government cf. top-left and top-centre graphs
  • This effect reflects on the electoral fortunes of the extreme parties within the blocks: the FN does better under a leftist government, while the Communists are better off under a right-wing government (though this effect is diminished when unemployment is very high). Cf. bottom-left and bottom-centre.
  • The moderate left is punished severely for high unemployment, even if they are not in government, whereas the moderate right can get away with moderate increases in unemployment. Even if unemployment is very high, the moderate right performs better than the moderate left (cf. top left vs. top centre)
  • Unemployment boosts FN voting, especially under a leftist government. (bottom left)
  • The prospects for “other”parties increase with unemployment; dramatically so if there is a moderate right government. (bottom right)
  • Non-voting increases (i.e. turnout goes down) with unemployment under a right-wing government. Non-voting goes down (i.e. turnout goes up) with higher unemployment rates if there is an incumbent left government (top-right). The FN, communists and other parties would benefit from this increase in turnout.
[Table 3 about here]

We move now to the predictive efficacy of the model for the 2007 legislative elections (Table 3). Recall that the vote share indicated includes abstention and spoilt/blank ballots. We therefore provide a table which includes not only the notional vote shares for 2007 as a proportion of registered voters under our predictive model, but also predicted and observed election outcomes as a proportion of votes cast.11 At first glance, the prediction is disappointing. All mainstream party scores are significantly over-estimated, with a concomitant under-estimation of the apparently ‘protest’ or ‘dissatisfaction’ votes. In terms of precision, then, our model does not pass with flying colours. That said, there are some positive aspects. Firstly, the model does correctly predict the winner of the election. Whilst this is less important for a relative majority system such as France than a two-horse race, it still passes the test of ‘spotting the winner’ – not always the case with ex ante forecasts (Lewis-Beck et al 2008).

Similarly the model actually under-estimates the Extreme Right score. As we noted in the overview of the French context, a model able to predict with some degree of accuracy the Extreme Right vote in France has since 2002 been an increasing focus of the forecasting literature. However, precisely given the exaggerated success – in qualitative terms – given to Le Pen’s performance in the 2002 presidentials, ex ante predictions generally exaggerated – in quantitative terms – the likely Le Pen / FN score in 2007. Yet, a simple economic model plus incumbency control in fact gives a very respectable estimate.

Comparing votes as a proportion of registered voters and as a proportion of votes cast,12 a clear failing in the model is that it fails to predict that, even in terms of first-round votes, the Moderate Right grouping wins an absolute majority of votes. Clearly, this does not factor in seat distribution, which follows from the run-off results (although in the 2007 National Assembly election, the Moderate Right UMP group finished with 317 seats or 54.9% of the 577 total – very close to the first-round proportions).13 However, in a system dominated by the notion of majoritarianism, and the lowest 50%+1 mandate of its president, the psychological effect of surpassing 50% of votes cast at the first round is significant, and one which we would hope to pick up in a robust forecasting model.

The overall efficacy of the model in predicting 2007 voting levels in the election is clearly far from perfect. To try to identify issues with the model, we run some basic diagnostic tests on the model data (Appendix 1). Looking at outliers in each model, a small number of outliers stand out with high studentized residuals. In particular the Moderate Left contrast in Corse du Sud (Southern Corsica) has consistently high residuals in all years except 1988 and 1993, due to the localised party system including Corsican nationalist and local mainstream party variants. Due to the low leverage of the residuals, however, the removal of this department has little effect on the model. Similarly, for the ‘other’ parties, there are a number of underestimations in the 1980s elections for the Central South-West region and some North-East border areas. In many cases, this is due to so-called ‘majority’ candidates (i.e. candidates who explicitly support the President, and who were therefore supported by the official presidential party) who would in reality be more appropriately seen as Moderate Left, or in the case of the North-East, independent Right-wing candidates. For the Communists in 2002, a degree of under-estimation is present in a number of southern or Midi departments due to the disproportionately strong support from the working class, to an extent significantly beyond the political tradition of these areas.

Looking at studentized residuals for the Moderate Left and Communists party groups in 2007, 50 of the 96 departments were significantly over-estimated (cut-off for residual > 2.0) for the Moderate Left, and a massive 90 out of the 96 departments were over-estimated for the Communists. Overall, then, case diagnostics reveal little that indicates why the 2007 forecasts erred from the pattern set by 1981 – 2002, beyond the Left in toto performing much worse than previous patterns would suggest.


Given the demands of using a constrained model in a complex multiparty setting with across-time variation in coalitions and party affiliation, the SUR approach provides a relatively robust model of French legislative election outcomes. Out-of-sample forecasting of the elections between 1981 and 2002 are on average quite strong, and concomitantly the parameter estimates for these models are generally consistent with what one would expect from the economic and VP-function literature. Inevitably, however, the panel nature of the data mean that changes in electoral coalition, mutual desistment pacts specific to individual elections and localised micro-party systems introduce bias into the model which it is difficult to correct. In an attempt to identify broad patterns economic effects on party support across time, such specificities will always explicitly influence predictions at the meso-level, and implicitly do so at the macro-level.

Of greater concern, perhaps, is the errant ex ante forecast of the 2007 result. Again, real life intervenes. In the French case, the key intervention which affects the legislative election outcome is the preceding presidential election. For four of our seven elections, a presidential race will have largely resolved the legislative outcome one month before the latter election. Yet, in 2007, there is good reason to believe that the resolution will have been more influential than ever. What had been expected to be a close presidential race, up until a relatively short time before the election, materialised into a comprehensive Right-wing victory over an unconvincing and divided Left, and the inevitable victory of Nicolas Sarkozy over the Socialist Ségolène Royal in the run-off. As a result, the legislative elections effectively became a confirmatory election to install President Sarkozy’s government, rather than a true electoral competition.

Similarly, the Extreme Right presidential candidate, Jean-Marie Le Pen, who had surpassed expectations in 2002, experienced a major downturn in electoral fortune in the 2007 presidentials. As with the mainstream rivalry, the legislative elections simply confirmed the electoral collapse of the Extreme Right with one of the worst national election scores ever for the FN. In a mirror image of the ‘unforeseeable’ success of Le Pen in 2002, so the crushing defeat was equally unforeseeable, at least in most of the predictive models prior to the election (Lewis-Beck et al 2008). Yet, as our model has shown, within a constrained model accounting for all party votes, the result was far from unforeseeable, and indeed this model could have been run a significant period before the legislative elections.14 Whilst the forecast is not precise, the prediction is not significantly outside the standard error of the model. To that extent, then, the 2007 performance was within the realms of prediction.

Whilst some additional refinement of party codings could potentially improve the model forecasts incrementally, the six-fold categorisation of the dependent variable does set the bar extremely high for a forecast model across 26 years of an extremely dynamic system. Undoubtedly, the SUR approach we have adopted is more robust than a simple linear regression approach for multiparty data. Yet, somewhat perversely, the evolution which prompted us to use this technique – namely, the rise in relevance in 2002 of a range of third parties and their presidential candidates – precisely seems to have been reversed in 2007. Looking back to Table 3 and the actual results, it is clear that the mainstream Moderate Left and Moderate Right blocs account for the vast proportion of the valid vote, with a noticeable shift back towards a two-bloc and indeed two-party logic (Grunberg and Haegel, 2008). In that sense, the more traditional incumbent / opposition approaches to electoral forecasting may well be increasingly appropriate once more for the French case. Whether a simplistic but effective forecasting outweighs a more nuanced but unstable approach is not a debate which can be resolved here. However, the latter approach which we have used here undoubtedly reveals a greater coherence and predictability to all types of party in the French system than perhaps previous research has been able.


Andersen, R. and J. Evans (2005) ‘The stability of French political space, 1988-2002’ in French Politics, 3:3, 282-301.

Anderson, C. (2000) ‘Economic voting and political context: a comparative perspective’, Electoral Studies, 19: 151-170.

Arzheimer, K. and Carter, E. (2006) ‘Political Opportunity Structures and Right-Wing Extremist Party Success’, European Journal of Political Research, 45, 419–443.

Auberger, A. (2004) ‘Les fonctions de vote: un survol de la littérature’, l’Actualité ÉconomiqueRevue d’Analyse Économique, 80: 95–107.

Auberger, A. (2005) “Forecasts of the 2004 French European election”, Swiss Political Science Review, 11, 61-78.

Auberger, A. (2008) “The National Front Vote and Turnout in the French Presidential Elections”, French Politics, 6:1, 94-100.

Auberger, A. and E. Dubois (2005) “The influence of local and national economic conditions on French Legislative elections”, Public Choice, 125, 363-383.

Bell, D. and B. Criddle (1994) The French Communist Party in the Fifth Republic, Oxford, Clarendon.

Caramani, D. (2000) Elections in Western Europe since 1815 : Electoral Results by Constituencies, Basingstoke, Macmillan.

Dubois, É. and C. Fauvelle-Aymar (2004) ‘Vote functions in France and the 2002 election forecast’, in M.S. Lewis-Beck (ed.) The French Voter: Before and After the 2002 Elections, Basingstoke, Palgrave Macmillan, 205–230.

Evans, J. and G. Ivaldi (2002) ‘Les dynamiques électorales de l’extrême-droite européenne’ in Revue Politique et Parlementaire, 1019, July-August, 67-83.

Evans, J. and G. Ivaldi (2008) ‘Forecasting the Extreme Right vote in France (1984-2007)’, French Politics, 6:2, 137-151.

Golder, M. (2003) ‘Explaining Variation in the Success of Extreme Right Parties in Western Europe’, Comparative Political Studies, 36, 432–466.

Grunberg, G. and F. Haegel (2008) La France vers le bipartisme?, Paris, les Nouveaux Débats (Presses de Sciences Po).

Grunberg, G. and E. Schweisguth (2003) ‘French political space: two, three or four blocs?’, French Politics, 1:3, 331-348.

Jackman, R. W. and K. Volpert (1996) ‘Conditions favouring parties of the Extreme Right in Western Europe’, British Journal of Political Science, 26, 501–521.

Jérôme B., V. Jérôme and M. Lewis-Beck (1999) “Polls fail in France: Forecasts of the 1997 legislative election”, International Journal of Forecasting 15: 163–174.

Jérôme, B. and V. Jérôme-Speziari (2003) “A Le Pen vote function for the 2002 Presidential election: a way to reduce uncertainty”, French Politics, 1:2, 247-251.

Jérôme, B. and V. Jérôme-Speziari (2004) “The 2004 French regional elections: politico-economic factors of a nationalized local ballot”, French Politics, 3:2, 142-163

Jérôme, B., V. Jérôme-Speziari and M. Lewis-Beck (2003) “Reordering the French election calendar: forecasting the consequences for 2002”, European Journal of Political Research, 42, 425-440.

Katz, J. and G. King (1999) “A statistical model for multiparty electoral data”, in American Political Science Review, 93:1, 15–32.

Knigge, P. (1998) “The ecological correlates of right-wing extremism in Western Europe”, European Journal of Political Research, 34:2, 249-79.

Lewis-Beck, M. (1988) Economic and Elections, Ann Arbor: University of Michigan Press.

Lewis-Beck, M. (1997) “Who’s the chef? Economic voting under a dual executive”, European Journal of Political Research, 31: 315-325.

Lewis-Beck, M., E. Bélanger and C. Fauvelle-Aymar (2008) “Forecasting the 2007 French Presidential Election: Ségolène Royal and the Iowa Model”, French Politics (2008) 6, 106–115.

Lubbers, M. and Scheepers, P. (2001) ‘Explaining the trend in Extreme Right-wing voting. Germany 1989-1998’, European Sociological Review, 17, 431–449.

Nannestad, Peter, and Martin Paldam (1994) ‘The VP-function – a survey of the literature on vote and popularity functions after 25 years’, Public Choice 79: 213-45.

Powell, G. Bingham and Guy Whitten (1993) ‘A cross-national analysis of economic voting: taking account of the political context’, American Journal of Political Science, 37: 391-414.

Shields, J. (2005) “Political Representation in France: a crisis of democracy”, Parliamentary Affairs 59:1,118-137.

Tomz, M., J. Tucker and J. Wittenberg (2002) “An easy and accurate regression model for multiparty electoral data”, Political Analysis, 10:1, 66–83.

Whitten, G. and H. Palmer (1999) ‘Cross-national analyses of economic voting’, Electoral Studies, 18: 49-67.

Table 1 Nested model comparisons, fixed effects and full model

forecasts (1981-2002)







Moderate Left: Fixed Effects






Moderate Left: Full Model





Moderate Right: Fixed Effects






Moderate Right: Full Model





Communists: Fixed Effects






Communists: Full Model





Extreme Right: Fixed Effects






Extreme Right: Full Model





Other: Fixed Effects






Other: Full Model





Table 2 Parameter estimates for full model forecasts (1981-2002)






Unemployment (U)

-0.03 (.01)

-0.10 (.01)

-0.06 (.01)

0.53 (.03)

0.63 (.09)

Left incumbency (I)

-1.69 (.10)

-0.66 (.12)

-0.36 (.10)

3.14 (.35)

5.24 (1.04)


0.00 (.01)

0.12 (.01)

0.05 (.01)

0.29 (.03)

-0.79 (.09)


0.14 (.01)

0.08 (.01)

0.08 (.01)

-0.15 (.03)

-0.46 (.10)







Root MSE






– All contrasts with abstention baseline

– Departmental dummies not shown

– All coefficients significant at .05 or lower, except Comm ΔGDP

Table 3 Predicted and observed vote outcomes in 2007 (% of total

registered voters and of votes cast)

% registered voters

% votes cast










Moderate Left





Moderate Right





Extreme Right













Figure 1: Unemployment and Voting in France

Appendix 1 Outlier Studentized residuals for Moderate Left, Communist and

Other parties

Moderate Left


| département year stud.resid |


41. | Alpes-Maritimes 2002 -2.19866 |

261. | Jura 1986 2.316615 |

293. | Haute-Loire 2002 -2.025205 |

352. | Haute-Marne 1986 2.060687 |

355. | Haute-Marne 1997 -2.564131 |


408. | Oise 1986 2.182928 |

426. | Pas-de-Calais 2002 -2.22702 |

464. | Haut-Rhin 1986 2.41543 |

534. | Yvelines 1986 2.006661 |

573. | Var 2002 -2.169649 |


659. | Corse-du-Sud 1981 3.205233 |

660. | Corse-du-Sud 1986 4.251734 |

663. | Corse-du-Sud 1997 -2.248766 |

664. | Corse-du-Sud 2002 -4.478615 |




| département year stud.resid |


13. | Aisne 2002 -3.365818 |

23. | Alpes-de-Haute-Provence 1986 2.262079 |

27. | Alpes-de-Haute-Provence 2002 -2.581943 |

50. | Ardennes 1981 2.240427 |

55. | Ardennes 2002 -4.792705 |


76. | Aude 2002 -2.099659 |

111. | Charente 2002 -2.66234 |

181. | Eure 2002 -2.27876 |

202. | Gard 2002 -2.175194 |

230. | Hérault 2002 -2.377525 |


377. | Meuse 2002 -2.454502 |

426. | Pas-de-Calais 2002 -2.64686 |

573. | Var 2002 -3.137405 |

580. | Vaucluse 2002 -2.145657 |

597. | Haute-Vienne 1986 2.517245 |


608. | Vosges 2002 -2.402605 |



| département year stud.resid |
43. | Ardèche 1981 -4.204402 |
52. | Ardennes 1988 -4.296376 |
73. | Aude 1988 -3.726954 |
80. | Aveyron 1988 -4.251042 |
82. | Aveyron 1997 2.003402 |
108. | Charente 1988 -3.450763 |
122. | Cher 1988 -3.783795 |
129. | Corrèze 1988 -3.623167 |
148. | Creuse 1981 -3.935467 |
157. | Dordogne 1988 -3.538015 |
178. | Eure 1988 -3.736878 |
239. | Indre 1981 -4.06781 |
269. | Landes 1988 -3.845304 |
288. | Haute-Loire 1981 -3.965096 |
309. | Lot 1981 -4.078404 |
360. | Mayenne 1988 -3.709984 |
374. | Meuse 1988 -3.643376 |
416. | Orne 1988 -3.594169 |
465. | Haut-Rhin 1988 -3.40009 |
479. | Haute-Saône 1988 -4.197225 |
500. | Savoie 1988 -3.453629 |
540. | Deux-Sèvres 1981 -4.398553 |
556. | Tarn 1988 -4.052777 |
584. | Vendée 1988 -3.667911 |
605. | Vosges 1988 -3.662599 |
666. | Haute-Corse 1981 -4.380981 |

1 Replication data are available from the authors’ dataverse at XXXXX


Table 1 Nested model comparisons, fixed effects and full model

forecasts (1981-2002)







Moderate Left: Fixed Effects






Moderate Left: Full Model





Moderate Right: Fixed Effects






Moderate Right: Full Model





Communists: Fixed Effects






Communists: Full Model





Extreme Right: Fixed Effects






Extreme Right: Full Model





Other: Fixed Effects






Other: Full Model





Table 2 Parameter estimates for full model forecasts (1981-2002)






Unemployment (U)

-0.03 (.01)

-0.10 (.01)

-0.06 (.01)

0.53 (.03)

0.63 (.09)

Left incumbency (I)

-1.69 (.10)

-0.66 (.12)

-0.36 (.10)

3.14 (.35)

5.24 (1.04)


0.00 (.01)

0.12 (.01)

0.05 (.01)

0.29 (.03)

-0.79 (.09)


0.14 (.01)

0.08 (.01)

0.08 (.01)

-0.15 (.03)

-0.46 (.10)







Root MSE






– All contrasts with abstention baseline

– Departmental dummies not shown

– All coefficients significant at .05 or lower, except Comm ΔGDP

Table 3 Predicted and observed vote outcomes in 2007 (% of total

registered voters and of votes cast)

% registered voters

% votes cast










Moderate Left





Moderate Right





Extreme Right