Conceptual Confusion is not Always a Bad Thing: The Curious Case of European Radical Right Studies


“I want to ride my bicycle
I want to ride my bike
I want to ride my bicycle
I want to ride it where I like.”



Over the course of many years, as a teacher, scholar, and friend, Ruth Zimmerling has impressed on me the importance of precisely defining one’s concepts. After all, if there is no agreement on the intension and extension of a concept, it is impossible “to assess the truth or falsity or, more generally, the correctness or incorrectness, of propositions, hypotheses or theories” (Zimmerling, 2005, p. 15). The statement is almost self-evident: Without precisely defined concepts, the whole endeavour of science becomes pointless, and scholarly discourses are bound to turn into dialogues of the deaf.

In her magisterial monograph, before she moves on to dissect and then reassemble the concepts of power and influence in a bid to clean up the mess that others have left, almost en passant Ruth makes a couple of important observations. First, she notes that in everyday situations, all of use words which lack clear definitions, yet most of the time, we are able to communicate “reasonably well” (Zimmerling, 2005, p. 15). Although “we must do better than just understand each other reasonably well” (Zimmerling, 2005, p. 15) in the realm of science, one unspoken implication is that the difference between scientific language and everyday language is often a gradual one. Second, she coins the notion of the “bicycle concept”: a concept “which is unproblematic as long as one does not stop to think about how exactly it works” (Zimmerling, 2005, p. 15).

While it would be difficult to disagree with Ruth’s plea for conceptual clarity on a general level, I think that, at least for the sake of an argument, it is possible to read these two observations against the grain. Unlike Ruth, I would like to argue that “bicycle concepts” can have their virtues (even if they might sometimes land one “flat on one’s face”): The very shininess of bicycle concepts may stimulate discourses by attracting new scholars to an emerging field, whereas their inherent flexibility and agility facilitates movement across disciplinary boundaries.

In a bid to backup this claim, in this chapter I will look at the development of a subfield of Political Science, namely what I will call for want of a better name “European Radical Right Studies” (henceforth ERRS), that clearly suffers from the problems identified by Ruth and yet has arguably made some progress over the last decades. ERRS presents an extreme case in several ways:

  • It clusters not around one, but several bicycle concepts.
  • It cannot even agree on the most appropriate label for the core bicycle concept that defines the field: is it the “extreme”, “radical”, “far”, “populist”, “anti-immigrant”, or “new” right?
  • It consciously abandoned conceptual reflection in favour of empiricism, then, a decade later, relatively quickly adopted a very specific (and highly useful) set of definitions, both under the influence of a single scholar.

My analysis is largely exploratory, probably affected by selection bias, and relies on messy data. Even worse, I may not have too clear a concept of “progress”. This irony is not entirely lost on me.

“European Radical Right Studies” – a messy field

Extremism and radicalism are venerable concepts in Political Science. Their long and convoluted history has been explored in detail elsewhere (Backes, 1989, 2007; Lipset & Raab, 1971). In postwar Europe, their use was confined to organisations at the very margins of the political spectrum and their supporters – communists of different strands on the one hand and right-wing parties and movements that harked back to the authoritarianism of the interwar period on the other.

However, in the late 1970s and early 1980s, an apparently new party family that was not easily classified as extremist rose to prominence in Western Europe. While some of its putative members were connected to traditional right-wing extremism through their history, ideology, and membership, others were unencumbered by such political baggage (the Danish and the Norwegian Progress Parties in particular, see Svåsand (1998)). More importantly, not a single one of the electorally relevant members of this family campaigned to replace democracy with some authoritarian alternative, and very few openly espoused traditional biological racism or antisemitism (although more covert references are not so unusual). Some of this parties such as the Dutch LPF and PVV or the Norwegian FrP even argue, in sharp contrast to traditional right-wing positions, that they are defenders of the rights of sexual minorities. What ultimately unites these parties (at least since the early 1980s) is their sharp opposition to non-Western immigration on the one hand and their problematic and ambivalent relationship with liberal democracy on the other (Arzheimer, 2008, Chapter 1.2.4): While they claim to be champions of some aspects of European democracy such as majority rule and freedom of speech (at least where it favours their own interests), they are highly sceptical of others (minority rights, representation, deliberation).

The early successes of these parties were sometimes mistaken as a return of interwar right-wing extremism (see Hagtvet (1994) for an example), but more often, they were perceived as worrying in their own right. Both perspectives have certainly contributed to the enormous and perhaps disproportionate (Mudde, 2013, p. 2) scholarly attention the phenomenon has received over the last three decades or so. Mudde’s claim that “more than a hundred scholars from across the globe work on the topic” (Mudde, 2013, p. 2) is an understatement – “several hundreds of scholars” (counting PhD students, PostDocs and established researchers) would be a more realistic assessment. Their work spans various subfields, most notably party and party system research and voting behaviour, but also political communication, political economy, political psychology, and several others. It is also interdisciplinary: while most scholars in the field are political scientists or sociologists by training and affiliation, contributions have also come from economy, psychology, history, and education.

A sizeable part of this research is documented in an extensive bibliography on the field that I maintain. This bibliography began as a list of references that I had perused in my own work from the 1990s up to and including my 2009 monograph on the electorate of these parties. Since 2010, the bibliography has been publically accessible on the internet ( Every six to eight months it is updated, using input from content databases and journals on the one hand and colleagues on the other. At the time of writing (January 2018), the bibliography contains 659 titles: 98 books, 121 chapters, and 439 articles from learned journals.

However, this literature’s disciplinary and conceptual diversity is often seen as problematic. As early as 1989, Uwe Backes bemoaned the “confusion of tongues” in the (much more narrowly defined) field of (German) research into right-wing extremism (Backes, 1989, p. 33). Reflecting on this observation and various other attempts to bring more conceptual clarity to the field, Cas Mudde wrote seven years later: “In 26 definitions of right-wing extremism that are used in the literature, no less then 58 different features are mentioned at least once. Only five features are mentioned, in one form or another, by at least half of the authors…” (Mudde, 1996, p. 229).

In this short contribution, I will focus only on the most obvious conceptual problem, namely the disagreement over a proper label for the field’s “core bicycle concept”. I will argue that this confusion has not prevented the emergence of large body of scholarship on the subject, and, more importantly, that this scholarship displays a high degree of interconnectivity and has not degenerated into a dialogue of the deaf.


The main advantage of using my own bibliography is that is shaped by my attempts to consciously identify a coherent (yet diverse) research community and its outputs. While I hope that the result contains much of the relevant research on the topic, it is important to point out some serious limitations and biases.

First, the bibliography’s substantive focus is on electorally relevant parties and their voters in Western Europe. Social movements and fringe parties, as well as Central and Eastern Europe get some coverage too, but in a much less comprehensive manner. Other geographical regions (North America in particular) are hardly represented at all when it comes to parties and voters, whereas research on potential antecedents and consequences (attitudes towards immigrants and even attitudes of immigrants) may be included, irrespective of the country on which the research was conducted.

Second, the bibliography contains just under ten per cent sources that are written in German, with the rest almost exclusively in English. Literature in other potentially relevant languages (in particular French and Italian) is neglected. For the analyses here, all sources in other languages than English are consciously excluded.

Third, the bibliography leans towards publications in peer-reviewed journals. While this is in line with developments in the field, and in Political Science more general, it still constitutes a source of bias. Finally, what is essentially a one-person endeavour can never compete with comparable commercial or institutional databases and will always be shaped by the authors personal preferences and prejudices.

To offset these problems, I will also make use of a commercial reference database (Web of Science/Social Science Citation Index). These sources have some problems of their own. The bias towards English-language sources and towards (certain) peer-reviewed journals is even more pronounced. Also, the selection of references pertaining to a subfield is largely driven by simple keyword searches, not by human expertise. On the other hand, the commercial nature of these databases means that conditional on these limitations, they provide unrivalled coverage of the literature. Moreover (again, within these limits), they provide the data necessary for bibliometric analyses.


What kind of “Right”?

[phrases-table] Distribution of various phrases in the literature
Any Phrase61
Radical Right*27
Extreme Right*21
Right-Wing Populis*10
Populist Radical Right*7
Far Right*6
Right-Wing Extrem*5
Populist Right*3
New Right*1
Radical Populist Right*0
Right-Wing Radic*0

The use of one of the aforementioned phrases1 in a prominent position (i.e. the title or the abstract (where applicable)) is very simplistic operationalisation of commitment to or at least engagement with a concept. The search strings included a “wildcard” (*) pattern so that they would also find common variations of the respective pattern. Table [phrases-table] shows the result of a host of substring search for common phrases in these positions. More than half of the 659 items in my bibliography use at least one of the phrases listed in the table. While a number of contributions amongst these 61 per cent make use of more than one phrase, the average number of phrases in this group is by no means excessive at just 1.3.2

Two phrases stand out: “Radical Right” and “Extreme Right” collectively show up in the titles and abstracts of nearly half of the items in the bibliography, whereas all other phrases are present in less than ten per cent of all items, respectively. The reminder of the analysis will there focus on the prominence and conceptual quality of the “radical” vs “extreme” distinction.

Extreme Right vs Radical Right over time

[fig-er-rr-timeseries] Prominence of “Extreme Right” and “Radical Right” over time

While the findings in Table [phrases-table] suggest the “Extreme Right” and “Radical Right” are more or less equally prominent in the literature, even a cursory glance at some of the landmark studies in the field suggests that their respective popularity may have waxed and waned over time. “Right-Wing Extremism in Western Europe” (Beyme, 1988) is one of the first edited volumes that reflected on the electoral mobilisation at the right end of the ideological spectrum in 1980s Western Europe. For von Beyme and his contributors, the rise of new or newly transformed parties at the very margins of the European party systems represented nothing less than a “third wave” (after the 1920s/30s and the 1950s/60s) of traditional right-wing extremism, which they addressed using the tools provided by the existing European (and largely German) research into extremist attitudes, ideologies, and behaviours.

Similarly, an influential article by Piero Ignazi in the European Journal for Political Research (Ignazi, 1992) provided a (then) new explanation (backlash against the “New Politics” of the 1960s and 1970s and their proponents) for the rise of these parties but also classified them as “extremist”, i.e. fundamentally opposed to liberal democracy. In this vein, “Extreme Right” became a convenient shorthand for a supposed family of parties to the right of the established Conservative and Christian Democracts (see e.g. Hainsworth (1992, 2008, 2000; Merkl & Weinberg, 1997, 2003)), even if their exact relationship with democracy was less clear-cut than the label extrem[e|ist] would suggest.

Some intellectual justification for this perhaps questionable practice came from a selective reading of Cas Mudde’s classic 1996 article in the European Journal for Political Research on the problems of defining the Extreme Right party family (Mudde, 1996). In this article, after a painstaking analysis of possible criteria for membership and their respective problems and shortcomings, Mudde remarks (paraphrasing a similar observation by von Beyme) that these definitional problems are of limited relevance for the practice of research, because there is almost universal agreement amongst on which parties should be classified as members of the Extreme Right (Mudde, 1996, p. 233). Many authors took this statement as a licence for definition by (collective) fiat.

However, around the same time, two monographs that had a substantial impact on the further development of the field were published: Hans-Georg Betz’s “Radical Right-Wing Populism in Western Europe” (Betz, 1994) and Herbert Kitschelt’s “The Radical Right in Western Europe” (Kitschelt, 1995). Both authors were influenced by Daniel Bell’s classic monograph on the Radical Right in the US (Bell, 2002)3, and both authors highlighted the internal diversity of the party family and how at least some of their members differed from traditional right-wing extremism. These monographs signify the emergence of a second stream in the literature that emphasises the relative newness of the parties under study.

Even so, within both streams scholars analysed the same set of parties and their voters. The question of how separate these streams are will be addressed in the next section, using commercial bibliographical data.

Anecdotal evidence suggests that most authors working in the field were “users” rather than “producers” of concepts: they were more interested in applying concepts to substantive problems than in developing them further and could switch effortlessly from studying the “Extreme Right” to work on the “Radical Right” (and back), depending on external factors such as the preferences of conference organisers, co-authors, and reviewers. This is readily illustrated by the titles of a series of volumes edited by Peter Merkl and Leonard Weinberg: Against the general trend (see below), they shifted from “Encounters with the Contemporary Radical Right” (Merkl & Weinberg, 1993) to “The Revival of Right-Wing Extremism in the 90s” (Merkl & Weinberg, 1997) and “Right-Wing Extremism in the Twenty-First Century” (Merkl & Weinberg, 2003), while most contributions in the latter books (including those written by the editors) still use the “Radical Right” label.

The publication of Cas Mudde’s 2007 monograph on “Populist Radical Right Parties in Europe” (Mudde, 2007), however, marked a turning point. Mudde is one of the most productive authors in the field in general and had been one of the most nuanced and thoughtful proponents of the “Extreme Right” school in particular. But in the first chapter of this new book, he proposed a new typology that represents a break from his earlier work (in particular from Mudde (1996) and Mudde (2000)) and aims at establishing a clear, hierarchical relationship between the most prominent adjectives used in the field. For Mudde (2007), the most general definition of the Radical Right party family includes only two elements: “nativism” (a combination of nationalism and xenophobia that implies that non-native elements – persons and ideas – represent a fundamental threat to the homogeneous nation state, see Mudde (2007, p. 19)) and “authoritarianism” in the sense of Altemeyer (1981), i.e. a highly conventional, aggressive, but not necessarily anti-democratic view of society. For Mudde, most (but by no means all) Radical Right parties are also “populist” in a very specific sense of the word: By populism, he means a “thin ideology” (Stanley, 2008) that pits “the pure people” against a “corrupt elite” (Mudde, 2007, p. 23) in the broadest sense and must be filled with more specific ideological content. Finally, a small subgroup within the Radical Right, which Mudde labels as the “Extreme Right”, openly oppose democracy Mudde (2007, p. 23).

Mudde’s typology is not without its own problems. (Right-wing) authoritarianism is a notoriously vague concept, and nativist and authoritarian tendencies are also found in established parties, particularly though not exclusively in (non-radical) parties of the right. Where and how does one draw the border? One the other hand, Mudde brings together three of the most prominent terms used in the literature in a way that is internally consistent and, in the view of many applied researchers, adequately captures empirical differences between the very many European right-wing parties. Consequently, the impact of his monograph on the field can hardly be overestimated.4.

Can these intellectual developments be quantified? Figure [fig-er-rr-timeseries] suggests that this is indeed the case. For each year from 1980 to 2017 (inclusive), the square and triangular markers represent the share of publications that use the respective phrase. Because the average number of publications per year is relatively low (1.3), these numbers fluctuate wildly, and so the superimposed non-linear trends better represent developments over time.5 The resulting picture is clear. During the 1980s, both phrases were used rarely, and from the 1990s into the early 2000s, “Extreme Right” had a slight lead over “Radical Right”. Following the publication of Mudde’s 2007 monograph, this began to change rapidly as can be gleaned from the rightmost part of the graph. For the whole decade from 2008 to 2017, the use of “Radical Right” outnumbers the use of “Extreme Right” by a factor of 2.2, and during the second half of that decade, the use of “Extreme Right” has become (increasingly) rare.

Obviously, using a phrase (albeit in a prominent position) does not necessarily imply a serious conceptual commitment, and where it does, the underlying conceptual framework is not necessarily Mudde’s. Nonetheless, it seems safe to assume that the field has overcome its most blatant shortcoming: the inability to agree on a common label for the phenomenon that scholars are studying.


To get a more rounded view of the field by digging into the relationships amongst authors and concepts, it is necessary to make additional use of commercial bibliographic data sources such as the Social Science Citation Index (SSCI).6 Relying on the SSCI implies some serious limitations. First, the index’s coverage is limited to a select group of journals in principle, while in practice criteria for inclusion are somewhat more vague, and the index may also include conference proceedings and other material. Second, whereas the bibliography is consciously limited to the Extreme/Radical Right in Europe, there are no reliable means for limiting the geographical scope of an SSCI query. Third, such querys are entirely key-word based and may return items which belong into the general domain of social science research but are not at all related to the research question at hand.

Bearing these limitations in mind, two separate queries for the phrases “Radical Right” and “Extreme Right” were run on the SSCI for the 1980-2017 period using the “TS” operator, which will return hits on the title, abstract, and keyword fields of the database. As of January 2018, the first query returns 596 hits, and the second query returns 551. By and large, this confirms the findings on their relative importance reported in the previous section. Also in line with these findings is the relatively low degree of overlap in the use of both phrases: A search for “Extreme OR Radical Right” returns 1,015 hits, implying that only 132 items use both phrases. To make the data more comparable, this list was further restricted to journal articles (745) and book chapters (3), excluding a surprisingly large number of book reviews (220) and other documents. A cursory glance at the titles of the remaining items identified about 20 unrelated articles, chiefly from the fields of brain research, motor-perception research, and genetics, which were removed.

[most-cited-in-ssrc] The ten most cited sources in 726 SSCI items
SourceNumber of times cited
Mudde (2007)160
Kitschelt (1995)147
Betz (1994)123
Lubbers et al. (2002)97
Norris (2005)90
Golder (2003)86
R.W. Jackman & Volpert (1996)77
Carter (2005)66
Arzheimer & Carter (2006)65
Brug et al. (2005)65

The key advantage of using these data is that the SSCI records all the sources that each item on the database cites, including (most) titles which are themselves not covered by the SSCI. The analysis presented in Table [most-cited-in-ssrc] also confirms the dominant position of Mudde’s 2007 book: It is the most cited and at the same time the youngest item on the list, surpassing the much older mongraphs by Kitschelt (1995) and Betz (1994). Put differently, roughly every third of the 552 articles published after 2007 cites this book.

Two other monographs are also frequently cited but appreciably less popular than the books by Betz, Kitschelt, and Mudde: Pippa Norris’s comparative study (Norris, 2005), embraces the “Radical Right” moniker and occasionally uses “extreme” either as a synonym or to refer to particularly problematic parties. Her main criterion for inclusion in the party family is the extremity of the parties’ political positions as measured by expert surveys. Elisabeth Carter (Carter, 2005), on the other hand, assumes that the parties covered by her study are all right-wing extremist. She uses three criteria – the party’s position towards immigration, racism and liberal democracy – to further subdivide the Extreme Right.

The five other items are journal articles with a primarily empirical outlook that lack conceptual ambitions. Golder (2003, p. 443) distinguishes between “older, neofascist parties and more recent, populist parties on the extreme right”. While he highlights their diverging electoral fortunes, the discussion of the criteria he employs is rather brief (Golder, 2003, pp. 446–447)R.W. Jackman & Volpert (1996) acknowledge a similar distinction7 but, like Carter (2005), assume that the parties they study are primarily right-wing extremist. Lubbers et al. (2002) and Arzheimer & Carter (2006) take the “Extreme Right” label for granted, whereas Brug et al. (2005) focus on the parties’ anti-immigrant message and use “radical”, “extreme”, and other labels interchangeably.

Collectively, these findings suggest that the “Extreme Right” label lacked a strong proponent, or at least a strong proponent that was widely received in the literature.8 Once someone presented a clear rationale for using the “Radical Right” label instead, many scholars were willing to jump ship.

If this was indeed the case, the literature should display a low degree of separation by the respective labels. One straightforward way of addressing this question is the analysis of co-citation patterns (Small, 1973). “Co-citation” simply means that two publications are both cited by some later source. By definition, a co-citations represent a view on the older literature as it is expressed in a newer publication. Each time two titles E and R, which respectively use the labels “Extreme” and “Radical”, are both cited in some later publication P, this is a small piece of evidence that the literature has not split into two isolated streams.

Because the SSCI aims at recording every source that is cited by the 726 titles and because most of these sources are themselves not included in the dataset, the number of candidate publications for co-citations is very large: 18255. However, the number of possible co-citation dyads is considerably larger. Less than half a per cent of these potential co-citations do exist, but their absolute number is still very large: 743032.9

To get a handle on this unwieldy co-citation network, the twenty publications with the biggest total number of co-citations and their interconnections were extracted. Many of them are familiar, because the most-cited sources from Table [most-cited-in-ssrc] are all included in this group (see Table [top-twenty-co-cited]). From the middle column of this table, it can also be seen that co-citations within the group are frequent. These titles represent something like the intellectual backbone of ERRS.

[top-twenty-co-cited] The twenty most co-cited sources in 726 SSCI items
SourceCo-citations within top 20Total co-citations
Kitschelt (1995)7457700
Mudde (2007)7408864
Lubbers et al. (2002)6005212
Norris (2005)5685077
Golder (2003)5644687
Betz (1994)5426151
R.W. Jackman & Volpert (1996)4774497
Brug et al. (2005)4623523
Arzheimer & Carter (2006)4603551
Knigge (1998)4453487
Carter (2005)3893291
Arzheimer (2009)3763301
Ignazi (2003)3442876
Ivarsflaten (2008)3343221
Ignazi (1992)3313230
Rydgren (2007)3003353
Bale (2003)2973199
Brug et al. (2000)2762602
Meguid (2005)2462600
Bale et al. (2010)1342449

[fig-network] Co-citations within top 20

But which titles are cited together? Figure [fig-network] depicts the top-20 co-citation network. The titles are arranged in groups, with proponents of the “Extreme Right” on the right side of the graph, authors using the “Radical Right” label in the lower-left quadrant, and a small group that is committed to neither label in the upper-left corner. The width of the lines is proportional to the number of co-citations connecting the titles.

The most obvious finding from Figure [fig-network] is that the network is almost complete: Apart from the missing link between Knigge (1998) and Bale et al. (2010), each title is connected to all other texts by co-citations. This already suggests that the field has not split into incompatible schools. Moreover, there are some very strong ties that bridge the supposed intellectual cleavages, e.g. between Kitschelt (1995) and Lubbers et al. (2002), between Lubbers et al. (2002) and Norris (2005), or between Kitschelt (1995) and Golder (2003).

Intuitively, it would seem as if co-citations were chiefly driven by the general prominence of the titles involved, whereas the use of compatible terminology seems to play a minor role. This intuition can be formalised and statistically tested by means of an appropriate regression model.

Co-citations are by definition counts, which renders the use of linear regression questionable, because counts are always non-negative and integer, and errors are unlikely to be distributed normally and with constant variance (Long & Freese, 2014, Chapter 9). The most simple model for count data is based on the Poisson distribution, which has equal (conditional) mean and variance (Zeileis et al., 2008, p. 5). In the present case, there are 190 observations of counts (the dyads formed by the top-20 titles in the co-citation network), with counts ranging from 0 (Knigge (1998)  Bale et al. (2010)) to 5476 (Mudde (2007)  Kitschelt (1995)). The unconditional mean number of co-citations is 695.4, but the unconditional variance is much higher at 651143, a phenomenon known as overdispersion. Overdispersion will result in invalid estimates for the standard errors.10

Regression of the number co-citations within top-20 on external co-citations and use of terminology


For overdispersed data, the negative binomial model (Hilbe, 2014, Chapter 5) is a popular alternative to Poisson regression. The former differs from the latter only insofar as it contains one additional parameter that accounts for the excess variation in the counts but is of no substantive interest, while the interpretation of the regression coefficients does not change.11

In Table [regression-table], for each dyad in the top-20 titles the number of co-citations was regressed on a) the sum of their respective co-citations outside the top-20 as a rough measure of the general popularity and compatibility of the two titles involved and b) the use of identical or diverging terminology. The coefficients refer to the linear-additive parameterisation of the model, which gives a sense of the direction of the effects but not much more. By exponentiating them, the model can be transformed to a multiplicative form, which is somewhat more accessible, but nonlinear: The model constant (2.852) is the natural log of the expected number of co-citations for two titles that use diverging terminology and have no external co-citations (a very unlikely scenario). By exponentiating this number, the expected co-citations can are obtained exp(2.852)≈17.

By the same logic, exponentiating the coefficient for using the same terminology (0.424) yields  1.53: Holding everything else constant, co-citations are 53 per cent more likely if two works use the same terminology. The effect of the sum of external co-citations as a measure of general popularity is also positive, but very small (0.00038). Exponentiating shows that each additional external co-citation is equivalent to an increase of 0.038 per cent in the count of internal co-citation.

[fig-margins-terminology] Effect of external co-citations and use of terminology on predicted number of co-citations within top 20

One must, however, keep in mind that each of these 20 titles has thousands of external co-citations, and that the variation in this count is in the thousands, too. Because of this wide variation in the range and distribution of the independent variables on the one hand, and because of the non-linear nature of the model on the other, plotting the expected counts against a range of plausible values12 for the number of co-citations gives a much clearer idea of what these findings mean in substantive terms.

As can be seen in Figure [fig-margins-terminology], the expected number of co-citations is largely unaffected by the question of terminology for works that have between 6,000 and 8,000 external co-citations. From this point on, the expected number of co-citations grows somewhat more quickly for dyads that share the same terminology. However, over the whole range of 6,000 to 12,000 external co-citations, the confidence intervals overlap and so this difference is not statistically significant.13 Put differently: Unless two titles have a very high number of external co-citations, the probability of them being both cited in a third work does not depend on the terminology they use, and even for the (few) heavily cited works, the evidence is insufficient to reject the null hypothesis that terminology makes no difference.

While the analysis is confined to the relationships between just 20 titles, it can be argued that these titles matter most, because they form the core of ERRS. If we cannot find separation here, that does not necessarily mean that it does not happen elsewhere, but if happens elsewhere, that is much less relevant.


For better or worse, European Radical Right studies have thrived over the last three decades, although for a long time, the subfield could not even agree on the name of its core concept. While the analysis in this chapter is seriously limited by a number of shortcomings – choice of a perhaps unusual subfield, focus on a mere label and purely quantitative and somewhat mechanistic methods of analyses, idiosyncratic underlying assumptions – this suggests that “bicycle concepts” may have their virtues: They are inherently flexible and can attract new scholars to a field.

Importantly, however, many scholars in the field recognised the need for a clear(er) concept and were all to willing to replace their favourite bicycle with something that was at least slightly more stable and much better understood. This suggests that “bicycle concepts” are most useful in emerging domains of research. Once a subfield becomes established, we must indeed “do better than just understand each other reasonably well” (Zimmerling, 2005, p. 15).

Altemeyer, B. (1981). Right-wing authoritarianism. Winnipeg: The University of Manitoba Press.

Arzheimer, K. (2008). Die wähler der extremen rechten 1980-2002. Wiesbaden: VS Verlag für Sozialwissenschaften.

Arzheimer, K. (2009). Contextual factors and the extreme right vote in western europe, 1980-2002. American Journal of Political Science 53(2): 259–275.

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

Backes, U. (1989). Politischer extremismus in demokratischen verfassungsstaaten. elemente einer normativen rahmentheorie. Opladen: Westdeutscher Verlag.

Backes, U. (2007). Meaning and forms of political extremism in past and present. Central European Political Studies Review 9(4): 242–262.

Bale, T. (2003). Cinderella and her ugly sisters: The mainstream and extreme right in europe’s bipolarising party systems. West European Politics 26: 67–90.

Bale, T., Green-Pedersen, C., Krouwel, A. & Luther, K.R. (2010). If you can’t beat them, join them? Explaining social democratic responses to the challenge from the populist radical right in western europe. Political Studies 58(3): 410–426.

Bell, D. (ed.). (2002). The radical right. third edition. with a new introduction by david plotke. New Brunswick: Transaction Publishers.

Betz, H.-G. (1994). Radical right-wing populism in western europe. Houndmills, London: Macmillan.

Beyme, K. (ed.). (1988). Right-wing extremism in western europe. London: Frank Cass.

Brug, W. van der, Fennema, M. & Tillie, J. (2000). Anti-immigrant parties in europe: Ideological or protest vote? European Journal of Political Research 37(1): 77–102.

Brug, W. van der, Fennema, M. & Tillie, J. (2005). Why some anti-immigrant parties fail and others succeed. a two-step model of aggregate electoral support. Comparative Political Studies 38: 537–573.

Carter, E. (2005). The extreme right in western europe. Manchester, New York: Manchester University Press.

Golder, M. (2003). Explaining variation in the success of extreme right parties in western europe. Comparative Political Studies 36(4): 432–466.

Hagtvet, B. (1994). Right-wing extremism in europe. Journal of Peace Research 31(3): 241–246.

Hainsworth, P. (2008). The extreme right in western europe. Abingdon, New York: Routledge.

Hainsworth, P. (ed.). (1992). The extreme right in europe and the uSA. London: Pinter.

Hainsworth, P. (ed.). (2000). The politics of the extreme right. from the margins to the mainstream. London, New York: Pinter.

Hilbe, J.M. (2014). Modeling count data. Cambridge: Cambridge University Press.

Ignazi, P. (1992). The silent counter-revolution. hypotheses on the emergence of extreme right-wing parties in europe. European Journal of Political Research 22: 3–34.

Ignazi, P. (2003). Extreme right parties in western europe. Oxford u.a.: Oxford University Press.

Ivarsflaten, E. (2008). What unites right-wing populists in western europe? Re-examining grievance mobilization models in seven successful cases. Comparative Political Studies 41(1): 3–23.

Jackman, R.W. & Volpert, K. (1996). Conditions favouring parties of the extreme right in western europe. British Journal of Political Science 26: 501–521.

Kitschelt, H. (1995). The radical right in western europe. a comparative analysis. Ann Arbor: The University of Michigan Press.

Knigge, P. (1998). The ecological correlates of right-wing extremism in western europe. European Journal of Political Research 34: 249–279.

Lipset, S.M. & Raab, E. (1971). The politics of unreason. right-wing extremism in america, 1790-1970. London: Heinemann.

Long, J.S. & Freese, J. (2014). Regression models for categorical dependent variables using stata (3rd ed.). College Station: Stata Press.

Lubbers, M., Gijsberts, M. & Scheepers, P. (2002). Extreme right-wing voting in western europe. European Journal of Political Research 41: 345–378.

Meguid, B.M. (2005). Competition between unequals: The role of mainstream party strategy in niche party success. American Political Science Review 99(3): 347–359.

Merkl, P.H. & Weinberg, L. (eds.). (1993). Encounters with the contemporary radical right. Boulder: Westview.

Merkl, P.H. & Weinberg, L. (eds.). (1997). The revival of right-wing extremism in the 90s. London, Portland: Frank Cass.

Merkl, P.H. & Weinberg, L. (eds.). (2003). Right-wing extremism in the twenty-first century. London: Frank Cass.

Mudde, C. (1996). The war of words. defining the extreme right party family. West European Politics 19: 225–248.

Mudde, C. (2000). The ideology of the extreme right. Manchester, New York: Manchester University Press.

Mudde, C. (2007). Populist radical right parties in europe. Cambridge: Cambridge University Press.

Mudde, C. (2013). Three decades of populist radical right parties in western europe: So what? European Journal of Political Research 52(1): 1–19.

Norris, P. (2005). Radical right. voters and parties in the regulated market. Cambridge, New York: Cambridge University Press.

Rydgren, J. (2007). The sociology of the radical right. Annual Review of Sociology 33(1): 241–262.

Small, H. (1973). Co-citation in the scientific literature. a new measure of the relationship between two documents. Journal of the Association for Information Science and Technology 24(4): 265–269.

Stanley, B. (2008). The thin ideology of populism. Journal of Political Ideologies 13(1): 95–110.

Svåsand, L. (1998). Scandinavian right-wing radicalism. In H.-G. Betz & S. Immerfall (eds.), The new politics of the right. neo-populist parties and movements in established democracies. New York: St. Martin’s Press.

Zeileis, A., Kleiber, C. & Jackman, S. (2008). Regression models for count data in r. Journal of Statistical Software 27(8): 1–25.

Zimmerling, R. (2005). Influence and power. variations on a messy theme. Dordrecht: Springer.

  1. Searching for phrases is more interesting than just searching for adjectives. It allows for distinguishing between e.g. “the radicalisation of the Extreme Right” and “the more extreme elements of the Radical Right”.
  2. This number partly reflects overlap that results from the way the search was conducted, e.g. “Populist Radical Right” is a subset of “Radical Right”.
  3. First published in 1961/63.
  4. As early as 2009, Political Studies Review (Volume 7, Issue 3) devoted a Review Symposium comprising of eight articles to the book. A decade after its publication, it had already been cited 2,085 times, according to Google Scholar. By way of comparison, the two leading mongraphs of the previous decade, Betz (1994) and Kitschelt (1995), had been cited 1950 times and 1599 times (all numbers as of January 2018).
  5. The smoother used is linear local regression, with a fraction of 75% of the data points included.
  6. The Social Science Citation Index is part of the Web of Science package, which was taken over in 2016 by Clarivate Analytics as part of a larger deal but was previously owned and developed by Thomson Reuters and was originally developed by Eugene Garfield, the pioneer of citation analysis.
  7. In a sense, their analysis is the template for Golder’s study.
  8. Somewhat ironically, Mudde (2000) (which builds on Mudde (1996)) also has a conceptual chapter which seeks to define the “Extreme Right” party family, but neither title was (or is) as widely cited as Mudde (2007).
  9. Its commercial origin not withstanding, the SSCI citation data are by no means perfect. The SSCI records authors with their last name and their initials, but last names change and the use of middle initials is often inconsistent, so that the same person may appear as two or more separate authors. In the case of very common last names and first names that start with the same letter, the opposite can happen. Moreover, cited sources are abbreviated in the SSCI, but sometimes there are slight variations in the way these abbreviations are formed, leading to a comparable inflation in the number of titles. This means that the analyses may slightly overestimate the fragmentation of the field.
  10. Overdispersion in the raw data does not necessarily result in conditional overdispersion (i.e. excess variation of the residuals around the model-implied mean), and even conditional overdispersion is sometimes only “apparent”, i.e. a result of model misspecification (Hilbe, 2014, Chapter 3). Here, however, the variance is more than 900 times bigger than the mean, suggesting that overdispersion is a real and serious problem.
  11. This parameter is often called α. Results in Table [regression-table] were obtained using the glm.nb function in R, which uses a somewhat unusual parameterisation and estimates θ = 1/α.
  12. The x-axis in Figure [fig-margins-terminology] ranges from the first to the ninth decentile of the empirical distribution.
  13. To simplify the analysis, the 190 dyads were treated as independent observations when in reality they are formed by permutations of just twenty titles. If this mutual dependency was factored in, the confidence intervals would be even wider.

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 (, 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:

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:

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.

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 ( or Michigan’s Survey Research Center (, 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, 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 ( 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 (, 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  (  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:
// 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: 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.
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.

Political Interest Furthers Partisanship in England, Scotland, and Wales


1. Introduction

In the field of public opinion and voting behaviour, few concepts have stimulated as many analyses and have aroused as much debate as that of party identification. The idea of a durable “identification” with a political party has been a staple in American electoral studies since the 1940s and is at the core of the Michigan model of electoral choice (Campbell et al., 1960). But with the proliferation of Michigan-inspired voting studies during the late 1960s and early 1970s (see Miller, 1994) scholars tried to apply the concept to political systems other than the US and began to couch their analysis of voting behaviour in Western Europe in terms of stabilizing party attachments and disruptive short-term factors. Roughly about the same time, the concept came under criticism both in the US and abroad.

A first strand of this critique accepts the basic premises of the Michigan model but holds that their empirical findings were dependent on the unusually quiet and stable political setting of the 1950s. According to this school, the political unrest of the 1960s and 1970s greatly reduced the prevalence and electoral importance of partisanship (Abramson, 1976; Nie et al., 1976). A second and related line of argument questions the validity of the indicators, which is of particular importance for the application outside the US. Critics objected that in political systems other than the US indicators for party identification merely presented another measure for vote intentions (Schleth & Weede, 1971; Thomassen, 1976; LeDuc, 1981; Küchler, 1986). Other challenges focus on the fact that the original Michigan indicator was based on a single continuum spanned by the two relevant American parties, whereas the vast majority of European polities feature multi-party systems (Katz, 1979).

The third and most fundamental line of criticism addresses the meaning and theoretical status of party identification. The original Michigan model as outlined in the American Voter was firmly grounded in then-contemporary assumptions about human behaviour that have been subsumed under the label of a “homo sociologicus of sociological empiricism” (Lindenberg, 1985, p.102): Voters were perceived as actors who “form opinions about everything”, are “easily influenced by others”, and “act directly on the basis of … [their] opinions” (Lindenberg, 1985, p.102). Consequentially, the attachment to one of the major parties would have a powerful direct impact on one’s political behaviour and would operate as a “perceptual screen” that biases the reception and the processing of political information (Campbell et al., 1960, p.133).

According to the classic Michigan model of political behaviour, citizens do not apply an explicit decision rule to maximise their utility from voting but rather react in a semi-conscious and stochastic way to political stimuli. This runs against the grain of a competing “model of man” which holds that citizens act rationally, i.e. actively choose the political alternative that is optimal given their general preferences. Starting out from Downs’ instrumental interpretation of party identification (Downs, 1957), authors such as Samuel L. Popkin and Morris P. Fiorina developed a competing view of partisanship. According to them, party identification is nothing but a cost-saving device for voters in a situation where it would be irrational to gather too much political information (Popkin et al., 1976; Fiorina, 1977; Fiorina, 1981).

While the question of whether humans act rationally in politics or merely follow their attitudinal dispositions is obviously a deep and complex one, in the case of party identification both competing views have observable implications that lend themselves to an empirical test. While the Michigan school never claimed that party identification is perfectly stable (Campbell et al., 1960, pp. 135, 165), it follows from their presentation in the American Voter and subsequent accounts that one’s identification is normally acquired early in life and then retained for years if not decades unless the political landscape changes dramatically. On the other hand, the “revisionist school” claims that party identification is nothing but a “running tally of retrospective evaluations of party promises and performance” (Fiorina, 1981, p. 84) that is frequently updated and changes accordingly as new political information comes in (see Milazzo et al. 2012 for a more subtle development of this idea in the British context).

On the basis of panel data that track the party identifications of individuals over time, it should be fairly straightforward to establish which of the competing views more adequately reflects political realities: if many electors change their identifications over a comparatively short period of four or five years, that would constitute rather strong evidence of rational updating in the light of new information. If, on the other hand, identifications are more or less stable over this period, this would prima facie support the classic view, although proponents of the revisionist school could still argue that political circumstances have simply not changed enough to trigger significant updating.

The issue is further complicated by random measurement error: respondents will get tired and bored during the course of a lengthy interview and give an incorrect or random answer to get over with the procedure, and interviewers will mishear or mistype their utterances. Therefore, not all apparent changes in the answers to a survey question reflect a true shift in the underlying attitude. Put differently, the correlation between measurements at time t and time t+1 will be attenuated.i Therefore, results from panel analyses of partisan change and stability that do not account for random measurement error will be seriously biased in favour of the revisionist school (Green et al., 2002).

As regards party identification in Britain, to our best knowledge only the analyses by Clarke and his colleagues (Clarke et al. 2004, Clarke et al. 2009, Clarke & McCutcheon 2009) deal adequately with this problem. Their main finding is that even if measurement error is taken into account, the “substantial latent-level dynamics in party identification in recent years … [are] inconsistent with the high levels of partisan stability as argued by Green et al. and other proponents of Michigan-style partisanship” (Clarke et al., 2004, p. 193; see also Clarke et al. 2009, pp. 327330; Clarke & McCutcheon 2009).

While we agree with Clarke et al. that party identification in Britain is not completely stable, we disagree with some of their interpretations of their findings. More importantly, we argue that their model should be extended to include political interest as a key factor that can explain stability of partisan ties. This reflects the crucial role political interest plays in another strand of the literature that harks back to Dalton’s (1984) seminal contribution on cognitive mobilisation and partisan dealignment. Although both strands of the literature should obviously complement each other, they are by and large disjoint. Our own model aims to narrow this gap.

In the remainder of this paper, we first review very briefly the existing evidence on the status of party identification in Britain. Then, we present findings from a model that differs from the specification chosen by Clarke et al. by introducing one additional variable and by defining the analytical sample in a slightly more restrictive way. These two innocuous modifications lead us to substantively different conclusions that are much more in line with the classic model of party identification. Finally, we discuss how our findings relate to the broader discussion on the nature and stability of party identification.

2. Previous Research

The Michigan-inspired concept of party identification (labelled as “partisan self-image”) was introduced by Butler and Stokes (1969; 1974) to the analysis of public opinion and voting behaviour in Britain. While building on the American concept, the authors found considerable Anglo-American differences. Though highly stable in absolute terms, partisan self-images in Britain turned out to be much more likely than their American counterparts to travel in tandem with vote choice over time. Given these findings, some scholars discarded the concept arguing that in Britain party identification is not as independent from voting behaviour as implied by the Michigan school. Others related the Anglo-American differences to features of the political institutions or even to measurement problems and concluded that the concept could be applied to Britain in principle (LeDuc, 1981; Mughan, 1981; but also see Crewe et al., 1977, p. 141; Cain & Ferejohn, 1981). Following the latter account, many scholars used the concept of party identification to analyse public opinion and voting behaviour in Britain.

Subsequent analyses, however, yielded more and more evidence that seemed at odds with the traditional concept. Beginning in the 1970s, Britain underwent a period of partisan dealignment, i.e. the strength of party attachments decreased considerably (Crewe et al., 1977). What is more, the decline in intensity resulted from party attachments’ responsiveness to retrospective evaluations of party performance on salient political issues and party leader evaluations (Clarke & Stewart, 1984; see also Johnston & Pattie, 1996). Expanding on this line of research, Clarke and colleagues (e.g. Clarke et al. 1997, 1998) argue in a number of contributions that the directional component of British partisanship also responds to short-term forces like performance evaluations and party leader images. Summarising these findings, Clarke et al. conclude that this level of aggregate and individual volatility cannot be reconciled with the classic model. Therefore, party identification in the sense of the Michigan model should be replaced by the Fiorina-inspired concept of “valenced partisanship” (Clarke et al., 2004, p. 211; Clarke et al. 2009; Whiteley et al. 2013).ii Specifically, employing fractional cointegration methods (e.g., Box-Steffensmeier & Smith 1996), Clarke and his colleagues suggest that party leader images and personal economic evaluations exert effects on aggregate-level party support in Britain (Clarke & Lebo 2003; see also Lebo & Young 2009; Pickup 2009).

But this conclusion has not gone uncontested. Several scholars claimed that the apparent instability of party identification in Britain is the result of an inappropriate survey instrument. Since the traditional indicator prompts respondents with a list of party labels and does not offer them an explicit non-identity option, respondents who lack the sense of durable attachment that is implied by the concept will in all likelihood answer the question on the basis of their present (but not necessarily stable) party preference. This will result in an inflated figure for the percentage of partisans in Britain as well as in a deflated estimate of partisan stability.

Cross-sectional analyses have largely supported the view that the traditional indicator has problems (Brynin & Sanders, 1997; Bartle, 1999; Bartle, 2003; Blais et al., 2001; Sanders et al., 2002; see also Sanders, 2003; see for a similar argument on the effects of question order Heath & Pierce, 1992). When it comes to the stability issue, however, the findings are somewhat mixed. Comparing the traditional and a revised measure (the so-called “supporter question”), Clarke et al. can demonstrate that both indicators result in identical rates of stable responses. At the same time, the revised indicator leads to considerably lower rates of inter-party change than the traditional instrument (Clarke et al., 2004, pp. 196199).

A related methodological objection against the notion that party identification in the UK is merely a “running tally” was raised by Donald Green and his colleagues. Using a simple dummy indicator specification for party identification, Green et al. show that party attachments in Britain turn out to be very stable over time if one controls for random measurement error. What is more, they demonstrate that short-term forces do not seem to affect party identification in Britain if random error is accounted for. Accordingly, Green et al. conclude that evidence in favour of the running tally account of party identification is based on methodologically flawed analyses (Schickler & Green, 1997; Green et al., 2002).

Clarke and his colleagues (e.g. Clarke & McCutcheon 2009, pp. 711-714), however, argue that Green et al.’s approach to measurement error is methodologically flawed and suggest to take an alternative route based on Mixed Markov Latent Class (MMLC) analysis a more adequate method for dealing with categorical latent variables that estimates probabilities for transition from one latent group to another and obtain very different results. Their analyses of panel data collected between 1963 and 2006 reveals considerable change in party identification at the latent-variable level even when allowing for measurement error. Testing four different specifications, they find that a Mixed Markov Latent class model with time homogeneous measurement error rates fits the data for Britain (and also Canada and the USA) best. In such a model, the measurement of partisanship is assumed to be affected by random error that is homogeneous (constant) over time. Moreover, one group of respondents is assumed to have perfectly stable partisan orientations (the “stayers”), whereas members of a second group change their orientations randomly from one wave to the next (the “movers”). The model is related to Converse’s “black-white” model but differs in one important aspect: the assumption that these switches occur with equal probability is relaxed. In Britain, the estimated size of the “mover” group varies between 29 and 37 per cent (Clarke & McCutcheon 2009, pp. 721). Therefore, they suggest discarding the original Michigan concept in favour of the running tally model (Clarke et al., 2004, pp. 194195; Clarke et al., 1999, pp. 97-101; Clarke et al. 2009, pp. 327330).

Clarke et al.’s works are an important methodological and substantive step forward, although we think that they do not represent the final word on the “Michigan vs Running Tally” debate. To be sure, the probabilities for retaining one’s party identification do not equal unity. Perfect stability, however, is not implied by the Michigan model. Rather, this would be a feature of the “unmoved mover” caricature of the original concept (Converse, 2006, p. 5).iii The finding that about two thirds of the respondents in the UK are “stayers” with perfectly stable partisan orientation, combined with the fact that estimated retention rates amongst the “movers” are also relatively high could easily be read as an endorsement of the Michigan model.

However, the debate on the relative merits of both models has become somewhat stale. In reality, electorates are comprised of stable partisans, party switchers, and apartisans, and investigating the scale, the sources and the consequences of this unobserved heterogeneity seems to open up promising new avenues for research (Neundorf, Stegmueller & Scotto 2011). In this article, we suggest adding political interest as an explanatory variable into Clarke et al.’s model to re-examine the over-time stability of party identification in Britain in the late 1990s. This seemingly innocuous modification leads to an important new insight that helps us to discriminate more clearly between the traditional interpretation of partisanship and the revisionist one. In addition, it has some bearing on a related conflict within the traditionalist camp.

The rationale is simple: From the ideas of Downs, Fiorina and Popkin it follows that citizens who are interested in politics and hence exposed to political information will update their “running tally” more frequently than their compatriots with lower levels of political interest. Being interested in politics should therefore have a positive effect on the probability of being in the “mover” class. Moreover, since party identification is merely a function of prior information, interest in politics will raise levels of exposure to political information and should hence be associated with higher levels of partisanship.iv

From the point of view of the traditionalists, expectations regarding the role of political interest are less clear-cut. On the one hand, starting from ideas originally formulated in the American Voter, the thrust of Dalton’s (1984) seminal “cognitive mobilisation” thesis is that low-interest citizens will rely on partisan cues, whereas cognitively mobilised citizens have no need for these cost-saving devices. Cognitive mobilisation will therefore undermine partisan ties and contribute to dealignmant (e.g. Dalton 2014).v For individual respondents, this implies that political interest should once more increase the probability of being the mover group. Moreover, it follows that political interest should have a negative effect on the likelihood of identifying with a political party.

On the other hand, Campbell et al. (1960, p. 133) have famously argued that party identification acts as a perceptual screen that affects the way political information is selected, processed, and stored (see also Bartels 2002). This view is in line with modern theories of motivated reasoning, which suggest that voters with high levels of interest in politics and high levels of political knowledge already stored in their long-term memory are more likely to select information that bolsters their preferences and to counter-argue information that runs contrary to their existing preferences (confirmation and disconfirmation bias). In effect, additional information is likely to support, rather than undermine, existing preferences. Among low-interest voters, this effect is smaller (e.g., Taber and Lodge 2006; see somewhat relatedly Converse, 1962; Zaller, 1992). Political interest should therefore be associated with higher levels of partisanship (e.g. Albright 2009), as well as with a higher probability of being in the stayer class.

To summarise, the three different theoretical perspectives lead to diverging expectations regarding the effect of political interest:

– Table 1 about here –

3. Data and Model

It is not our aim to fully replicate the various analyses by Clarke and his colleagues or to refute their findings. Rather, we want to demonstrate how adding a single variable to their model can shed some new light on the heterogeneity of partisanship stability in Britain and beyond. We therefore focus on data that were collected for the British Election Panel Survey from 1997 to 2000. There are three reasons for re-analysing these data that were collected more than a decade ago. First, this was the last face-to-face panel study of the BES, which began to employ internet (access) panels from 2005 on. While these are cost-efficient and can offer comparable data quality, the bulk of the data analysed comes from traditional surveys, and we would rather avoid mode effects. More specifically, we are worried that “professional respondents” who participate in many surveys may be more interested in politics in the first place and will develop more consistent attitudes over time, which would bias the results in favour of the traditionalist camp. However, an additional analysis of the 2005-2009 internet panel (documented in the online appendix) leads to essentially identical findings.

Second, the 1997-2000 panel (just like the 2005-2009 study) is a “truncated” panel, as far as party identification is concerned, because the last measurement was taken well before the beginning of the election campaign in the following year. Hence, it will not be contaminated by any campaign effects that may prompt voters to revert to their initial identification (e.g. Finkel 1993), while still giving us the benefits of analysing four measurements that span almost the complete life of a parliament.

Third, and most importantly, Clarke & McCutcheon (2009, p. 721) flag up the 1997-2000 panel as the UK’s least stable in modern times. This is confirmed by our own analyses of the 1963-2009 BES panel data: While 80 per cent of the respondents reported the same identification in the last wave in 2000 as they did when they were first interviewed in 1997 (compared to 74 per cent in the 1992-95 and 84 per cent in the 2005-08 periods), the year-to-year changes (documented in the online appendix) are higher than most of those that have been previously observed. To summarise, re-analysing the 1997-2000 data stacks the odds against finding stable identifications of traditionalist lore and is therefore a sensible modeling choice.

The volatility of the 1997-2000 period does not come as a surprise. After all, “New Labour” won the 1997 election by a landslide following the abolition of the old Clause IV (that called for nationalisation) in 1995 and an “Americanised” campaign, which played down ideological differences between the parties. Subsequently, Labour implemented a flurry of constitutional changes, but carried on with many of the previous government’s fiscal and economic policies and established a leadership style that was widely seen as presidential and tried to appeal to the broadest possible constituency (Bevir & Rhodes, 2006). Therefore, the early Blair years should provide a litmus test for the hypothesis of (largely) immutable identifications: if identifications resist the momentum of these events, if they remain stable although Britain has, according to Clarke et al., entered an era of valence politics, this would constitute strong evidence in favour of the classic model.

Unlike Clarke et al., we analyse the 1064 respondents from the smaller nations (882 from Scotland and 182 from Wales separately). Since the 1970s, both Scotland and Wales have featured party systems that are clearly distinct from the national party system, with Labour as the dominant party (until 2014/15) and comparatively strong nationalist parties that represent the centre-periphery cleavage. Devolution of the Scottish Parliament and the Welsh Assembly in 1999 has made these differences even more salient (Bohrer & Krutz, 2005; Lynch, 2007). In our view, pooling respondents from the three nations would lead to more unobserved heterogeneity instead of controlling for sources of heterogeneity, as the presence of the nationalist parties (taking seats in the Scottish Parliament and the Welsh assembly from 1999) will alter the meaning and possibly the stability of the “other” category. Moreover, the stability of mainstream party identifications might be lower in the periphery because respondents may well hold diverging party preferences and loyalties at the national and the sub-national level (see for evidence on the US and Canada Niemi et al., 1987; Stewart & Clarke, 1998).

The second important difference between the analysis by Clarke et al. and our approach is the specification of the model: As outlined in the previous section, we introduce self-stated political interest as an additional variable, which is correlated with both the mover/stayer property and the initial latent identification. The model’s parameters were estimated with MPlus 7.11 assuming that data are missing at random (MAR). Apart from considering political interest, our model is identical to the Mixed Markov model championed by Clarke et al.: We distinguish between Labour identifiers, Conservative identifiers, and all others. Like Clarke & McCutcheon, we further assume that reliability of the indicator is stable across the 1997-2000 period but place no restrictions on the transition matrices within the mover group.

4. Findings

As outlined in the previous section, we set up the model separately for a) England and b) the devolved nations. To establish a baseline, we first estimate the model without considering political interest and its correlations with the latent classes. In line with our expectations, restricting the sample to English respondents results in a higher estimate for the proportion of stayers (75 per cent vs. 66 per cent). Like Clarke et al., we note that some (about one third) of those placed in the mover category do not actually change their identification, bringing the total number of respondents with perfectly stable party identification to 83 per cent. This finding alone is a powerful re-assertion of the idea of largely stable party alignments.

In 1997, 40 per cent of the respondents were Labour identifiers, and 32 per cent identified with the Conservatives. Amongst the initial Labour identifiers, only one quarter are movers. For the Conservatives, the rate is even lower (one fifth), but for the “other” group, it approaches one third. Across all respondents, the probability of retaining one’s identification from one wave to the next varied between 0.94 and 0.99 for Labour and between 0.86 (between 1997 and 1998) and 0.97 for the Tories. For the heterogeneous “other” group, the retention rate is markedly lower (0.74) in 1998 and somewhat lower (0.92) in 1999 and 2000.

For Scotland and Wales combined, the estimate for the stayer group is also somewhat higher than 70 per cent and thus tends to exceed the figure reported by Clarke & McCutcheon. While this difference is small and within the margin of error, the discrepancy underlines that it is worthwhile to disaggregate the data. Once more, about one third of the “movers “(most of them Labour identifiers) retain their initial identification over the four waves, bringing the total number of stable partisans close to 80 per cent. Across all respondents, retention rates are slightly lower than in England and vary between 87 per cent (Labour from 1998 to 1999, the year of the first elections to the Scottish Parliament and Welsh Assembly) to 97 per cent (the Conservatives 1999 to 2000).

Next, we introduce political interest as an additional variable. Interest was measured in every wave using a five-point scale. While we could have constructed another latent variable from these four measurements (see, e.g., Prior 2010), this would have made the estimation prohibitively expensive in terms of computation, and would have led to a further proliferation of parameters. Instead of trying to purge random measurement error from this variable, we simply use the single measurement from the first wave as a rough indicator for general political interest. This is actually a conservative modelling strategy: Any random noise in the 1997 measurement (which could be due to higher levels of interest during the campaign) will dilute the relationship between political interest on the one hand and partisanship and its stability on the other.

  • Table 2 about here –

Including political interest in the model involves estimating three additional parameters – one for membership in the mover group, and two for initially identifying with Labour or the Conservatives, respectively. The change in the Bayesian Information Criterion (BIC) shows that this additional complexity is well-warranted: The BIC (which aims to strike a balance between model fit and parsimony) drops from 11,024 to 10,996 in England and from 4,386 to 4,378 in Scotland/Wales, which indicates a modest improvement over the pure Mixed Markov model (see Table 2).

  • Table 3 about here –

In both regions, including political interest in the model has hardly any effect on the estimated sizes of the mover/stayer classes. However, higher levels of political interest seem to reduce the probability of being a mover. The effects are sizable at about -0.2, meaning that each one-point increase in interest will reduce the odds of being a mover by roughly 18 per cent, but are not statistically significant by conventional standards (p=0.06 in England, 0.09 in Scotland/Wales). Hence, the findings (narrowly) fail to support the “motivated reasoning” perspective on partisan stability, but also clearly contradict the “running tally” and “cognitive mobilisation” perspectives, which predict a positive coefficient.

Regarding the effect of political interest on partisanship, the results are less equivocal. Political interest has a statistically significant and substantial positive effects on both identification with the Labour and the Conservative party (see Table 3). They are equivalent to changes in the odds ranging from 19 to 43 per cent. The somewhat weaker effects in the devolved nations are compatible with a lower number of apartisans in the “other” category due to the relevance of the nationalist parties.

  • Table 4 about here –

Moreover, there is a strong correlation between membership in the mover/stayer groups and the direction of the (initial) party identification. In England, stayers are considerably more likely to identify with one of the two major parties (see Table 4) than movers. From the relative size of the groups it follows that only about 19 per cent (a quarter of 75 per cent) of the respondents remain in the “other” group over the course of the survey. In Scotland and Wales, however, stayers predominantly identify with none or one of the nationalist parties, whereas more than two thirds of the movers identified with Labour in 1997 but may have changed their allegiance further down the line.

  • Table 5 about here –

Taken together, the results imply that overall retention rates are high in some, but not in all circumstances as Table 5 reveals. In particular, we can see that Labour experiences a significantly lower rate of retention outside England while the situation is reversed for the Tories and to a lesser extent for the “other” parties, both of which see a higher levels of loyalty among supporters in Scotland and Wales than in Englandvi These findings suggest that some of the late-1990s support for Labour in Scotland and Wales may have been rather instrumental in nature, and that devolution may have facilitated and accelerated the emergence of subnational party systems in the devolved nations.

5. Conclusion

In many ways, the discussion between proponents of the Michigan model (that has been updated and revised many times since its first inception in the 1950s) and their revisionist critics could be described as a dialogue of the deaf. This is in part because both models will lead to very similar empirical findings under most circumstances. In this contribution, we have tried to bring some fresh air to this otherwise stale debate by taking a closer look at the role of interest in politics in affecting the prevalence and stability of party attachments. Relying on this perspective, we derived three models with clearly distinguishable predictions: the “running tally” model, the “cognitive mobilisation” model, and the “motivated reasoning” model, which is well in line with the traditional notion of party identification.

The evidence gleaned from the 1997-2000 BES panel survey suggests that in Britain the traditional model is better suited to describe the role of interest in politics in affecting the prevalence and stability of party attachments than its contenders. Rather than providing evidence for a frequent “updating” of identifications amongst those who are interested in politics, our results for England and Scotland and Wales support the classic view that a change in party identification is a comparatively rare event amongst both high- and low-interest citizens, and it is even rarer among the former than among the latter. Our findings thus support the classic notion of party identification (Campbell et al., 1960; see also Bartels 2002), which implies that party attachments serve as perceptual screen. In particular, they suggest that the affective nature of party identification is conducive to motivated reasoning and thereby lends considerable stability to party attachments. In line with recent findings in political psychology (e.g., Taber and Lodge, 2006), these self-stabilizing effects of party attachments are particularly strong among high-interest citizens. This finding in turn fits nicely into Zaller’s (1992) RAS model which posits that high levels of political awareness is valuable in identifying and refuting information that contradicts existing predispositions.

In accordance with the latter model, our findings also have implications for the dynamics of party identification at the aggregate level, the so-called macro-partisanship (e.g., MacKuen et al. 1989; Clarke et al. 2001; Clarke & Lebo 2003). They suggest that the dynamics in the aggregate-level distribution are primarily driven by voters who are not heavily interested in politics. To be sure, this finding on voter heterogeneity does not imply that these dynamics are indicative of some kind of irrationality (see on this debate, e.g., Page and Shapiro 1992). However, it contradicts the notion that it is highly involved voters who cause shifts in macro-partisanship, a notion that would appear to be desirable from a democratic accountability perspective. Thus our findings can be seen as supportive of the well-known paradox whereby individuals with less than ideal citizenship traits actually seem to make important contributions to the functioning of democratic political systems (Berelson et al. 1954: 316; Neuman 1986).

As is the case for most empirical studies of political behaviour, this paper is subject to several limitations.vii First, we studied the stability of party attachments in a specific period in time. The 1997-2000 period is probably atypical in that citizens were provided with much information that could make them switch party allegiances. While this characteristic made a good test case for the hypotheses it also limits the generalizability of our findings, although our additional analysis of the 2005-2009 data leads to essentially identical results. Also, we did not take into account the durability of changes in party attachments. With data from multi-wave panel surveys covering longer time periods, future research may also be able to distinguish short-term fluctuations from more permanent changes in party attachments and hence explore whether political interest affects the durability of shifts in party attachments. Moreover, we have to keep in mind that we simplified our model by treating political interest not as latent variable. Yet, this strategy is likely to have diluted the impact of political interest on the stability of party attachments. Utilizing more sophisticated techniques would probably yield evidence that supports our conclusions even more strongly. So, we are confident that our evidence lends considerable support to the classic notion of party identification in Great Britain.


Abramson, P.R. (1976) Generational Change and the Decline of Party Identification in America: 1952–1974. American Political Science Review, 70(2), pp. 469–478.

Albright, J.J. (2009) Does Political Knowledge Erode Party Attachments? A Review of the Cognitive Mobilization Thesis. Electoral Studies, 28(2), pp. 248–260.

Bartels, L. M. (2002) Beyond the Running Tally: Partisan Bias in Political Perceptions. Political Behavior, 24(2), pp. 117-150.

Bartle, J. (1999) Improving the Measurement of Party Identification in Britain. British Elections & Parties Review, 9, pp. 119–135.

Bartle, J. (2003) Measuring Party Identification: An Exploratory Study with Focus Groups. Electoral Studies, 22(2), pp. 217–237.

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

Bevir, M. & Rhodes, R.A.W. (2006) Prime Ministers, Presidentialism and Westminster Smokescreens. Political Studies, 54(4), pp. 671–690.

Blais, A., Gidengil, E., Nadeau, R. & Nevitte, N. (2001) Measuring Party Identification: Britain, Canada, and the United States. Political Behavior, 23(1), pp. 5–22.

Bohrer, R.E. & Krutz, G.S. (2005) The Devolved Party Systems of the United Kingdom: Sub-national Variations from the National Model. Party Politics, 11(6), pp. 654–673.

Box-Steffensmeier, J.M. & Smith, R.M. (1996) The Dynamics of Aggregate Partisanship. American Political Science Review, 90, pp. 567-580.

Brody, R.A. & Rothenberg, L.S. (1988) The Instability of Partisanship: An Analysis of the 1980 Presidential Election. British Journal of Political Science, 18(Oct 88), pp. 445–465.

Brynin, M. & Sanders, D. (1997) Party Identification, Political Preferences and Material Conditions: Evidence from the British Household Panel Survey, 1991–1992. Party Politics, 3(1), pp. 53–77.

Butler, D. & Stokes, D. (1969) Political Change in Britain: Forces Shaping Electoral Choice (London: Macmillan).

Butler, D. & Stokes, D. (1974) Political Change in Britain: The Evolution of Electoral Choice (London: Macmillan).

Cain, B.E. & Ferejohn, J. (1981) Party Identification in the United States and Great Britain. Comparative Political Studies, 14(1), pp. 31–47.

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

Clarke, H.D. & McCutcheon, A.L. (2009) The Dynamics of Party Identification Reconsidered. Public Opinion Quarterly, 73(4), pp. 704–728.

Clarke, H.D. & Lebo, M. (2003) Fractional (Co)integration and Governing Party Support in Britain. British Journal of Political Science, 33, 283-301.

Clarke, H.D. & Stewart, M.C. (1984) Dealignment of Degree: Partisan Change in Britain, 1974–1983. Journal of Politics, 46(3), pp. 689–718.

Clarke, H.D., Stewart, M.C. & Whiteley, P. (1997) Tory Trends: Party Identification and the Dynamics of Conservative Support since 1992, British Journal of Political Science, 27(2), pp. 299–319.

Clarke, H.D., Stewart, M.C. & Whiteley, P. (1998) New Models for New Labour: The Political Economy of Labour Party Support, January 1992–April 1997. American Political Science Review, 92(3), pp. 559–575.

Clarke, H.D., Stewart, M.C. & Whiteley, P. (1999) New Labour’s New Partisans: The Dynamics of Party Identification in Britain since 1992. British Elections & Parties Review, 9(1), pp. 87–104.

Clarke, H.D., Stewart, M.C. & Whiteley, P. (2001) The Dynamics of Partisanship in Britain: Evidence and Implications for Critical Election Theory. British Elections & Parties Review, 11(1), pp. 66-83.

Clarke, H.D., Sanders, D., Stewart, M. & Whiteley, P. (2004) Political Choice in Britain (Oxford: Oxford University Press).

Clarke, H.D., Sanders, D., Stewart, M. & Whiteley, P. (2009) Performance Politics and the British Voter (Cambridge: Cambridge University Press).

Converse, P.E. (1962) Information Flow and the Stability of Partisan Attitudes. Public Opinion Quarterly, 26(4), pp. 578–599.

Converse, P.E. (2006) Researching Electoral Politics. American Political Science Review, 100(4), pp. 1–8.

Crewe, I., Särlvik, B. & Alt, J. (1977) Partisan Dealignment in Britain 1964–1974. British Journal of Political Science, 7(2), pp. 129–190.

Dalton, R.J. (1984) Cognitive Mobilization and Partisan Dealignment in Advanced Industrial Democracies. Journal of Politics, 46(1), pp. 264–284.

Dalton, R.J. (2014) Interpreting Partisan Dealignment in Germany. German Politics, 23(1-2), pp. 134-144.

Dalton, R.J. (2000) The Decline of Party Identifications, in: R.J. Dalton & M.P. Wattenberg (eds) Parties without Partisans. Political Change in Advanced Industrial Democracies (Oxford: Oxford University Press), pp. 19–36.

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

Finkel, S.E. (1993) Reexamining the “Minimal Effects” Model in Recent Presidential Campaigns. Journal of Politics, 55(1), pp. 1–21.

Fiorina, M.P. (1977) An Outline for a Model of Party Choice. American Journal of Political Science, 21(3), pp. 601–625.

Fiorina, M.P. (1981) Retrospective Voting in American National Elections (New Haven:Yale University Press).

Goodman, L.A. (1961) Statistical Methods for the Mover-Stayer Model. Journal of the American Statistical Association, 56(296), pp. 841–868.

Green, D., Palmquist, B. & Schickler, E. (2002) Partisan Hearts and Minds. Political Parties and the Social Identities of Voters (New Haven: Yale University Press).

Greene, S. (2004) Social Identity Theory and Party Identification. Social Science Quarterly, 85(1), pp. 136-153.

Heath, A. & Pierce, R. (1992) It was Party Identification all along: Question Order Effects on Reports of Party Identification in Britain. Electoral Studies, 11(2), pp. 93–105.

Johnston, R.J. & Pattie, C.J. (1996) The Strength of Party Identification Among the British Electorate: An Exploration. Electoral Studies, 15(3), pp. 295–309.

Kalton, G., Kasprzyk, D. & McMillen, D.B. (1989) Nonsampling Error in Panel Surveys, in: D. Kasprzyk, G. Duncan, G. Kalton & M.P. Singh (eds) Panel Surveys (New York: Wiley), pp. 249–270.

Katz, R.S. (1979) The Dimensionality of Party Identification: Cross-National Perspectives. Comparative Politics, 11(2), pp. 147–163.

Küchler, M. (1986) Wahl- und Surveyforschung. Politische Vierteljahresschrift, 27(Special Issue 17), pp. 194–208.

Langeheine, R., Pannekoek, J. & van de Pol, F. (1996) Bootstrapping Goodness-of-Fit Measures in Categorical Data Analysis. Sociological Methods & Research, 24(4), pp. 492–516.

Lebo, M. J. & Young, E. (2009) The Comparative Dynamics of Party Support in Great Britain: Conservatives, Labour and the Liberal Democrats. Journal of Elections, Public Opinion and Parties, 19 (1), pp. 73-103.

LeDuc, L. (1981) The Dynamic Properties of Party Identification: a Four-Nation Comparison. European Journal of Political Research, 9(3), pp. 257–268.

Lindenberg, S. (1985) An Assessment of the New Political Economy: Its Potential for the Social Sciences and Sociology in Particular. Sociological Theory, 3(1), pp. 99–114.

Lynch, P. (2007) Party System Change in Britain: Multi-Party Politics in a Multi-Level Polity. British Politics, 2(3), pp. 323–346.

MacKuen, M.B., Erikson, R.S. & Stimson, J.A. (1989) Macropartisanship. American Political Science Review, 83 (4), 1125-1142.

Milazzo, C., Adams, J. & Green, J. (2012) Are Voter Decision Rules Endogenous to Parties’ Policy Strategies? A Model with Applications to Elite Depolarization in Post-Thatcher Britain. Journal of Politics, 74(1), pp. 262–276.

Miller, W. E. (1994) An Organizational History of the Intellectual Origins of the American National Election Studies. European Journal of Political Research, 25(3), pp. 247–265.

Mughan, A. (1981) The Cross-National Validity of Party Identification: Great Britain and the United States Compared. Political Studies, 29(3), pp. 365–375.

Neuman, W. R. (1986) The Paradox of Mass Politics: Knowledge and Opinion in the American Electorate. Cambridge: Harvard University Press.

Neundorf, A., Stegmueller, D., & Scotto, T. J. (2011) The Individual-Level Dynamics of Bounded Partisanship. Public Opinion Quarterly, 75(3), pp. 458–482.

Nie, N. H., Verba, S. & Petrocik, J.R. (1976) The Changing American Voter (Cambridge: Harvard University Press).

Niemi, R.G., Wright, S.G. & Powell, L.W. (1987) Multiple Party Identifiers and the Measurement of Party Identification. Journal of Politics, 49(4), pp. 1093–1103.

Page, B.I. & Shapiro, R.Y. (1992) The Rational Public: Fifty Years of Trends in Americans’ Policy Preferences. Chicago: University of Chicago Press.

Pickup, M. (2009) Testing for Fractional Integration in Public Opinion in the Presence of Structural Breaks: A Comment on Lebo and Young. Journal of Elections, Public Opinion and Parties, 19 (1), pp. 105-116.

Popkin, S., Gorman, J.W., Phillips, C. & Smith, J.A. (1976) Comment: What Have You Done for Me Lately? Toward An Investment Theory of Voting. American Political Science Review, 70(3), pp. 779–805.

Prior, M. (2010) You’ve Either Got It or You Don’t? The Stability of Political Interest over the Life Cycle. Journal of Politics 72, pp. 747-766.

Rapoport, R.B. (1982) Sex Differences in Attitude Expressions: A Generational Explanation. Public Opinion Quarterly, 46(1), pp. 86–96.

Richardson, B.M. (1991) European Party Loyalties Revisited. American Political Science Review, 85(3), pp. 751–775.

Sanders, D. (2003) Party Identification, Economic Perceptions, and Voting in British General Elections, 1974–1997. Electoral Studies, 22(2), pp. 239–263.

Sanders, D., Burton, J. & Kneeshaw, J. (2002) Identifying the True Party Identifiers: A Question Wording Experiment. Party Politics, 8(2), pp. 193–205.

Schickler, E. & Green, D.P. (1997) The Stability of Party Identification in Western Democracies: Results from Eight Panel Surveys. Comparative Political Studies, 30(4), pp. 450–483.

Schleth, U. & Weede, E. (1971) Causal Models of West-German Voting Behavior. Sozialwissenschaftliches Jahrbuch für Politik, 2, pp. 73–97.

Stewart, M.C. & Clarke, H.D. (1998) The Dynamics of Party Identification in Federal Systems: The Canadian Case. American Journal of Political Science, 42(1), pp. 97–116.

Taber, C. S. & Lodge, M. (2006) Motivated Skepticism in the Evaluation of Political Beliefs. American Journal of Political Science 50, pp. 755-769.

Thomassen, J. (1976) Party Identification as a Cross-National Concept: Its Meaning in the Netherlands, in: I. Budge, I. Crewe & D. Farlie (eds) Party Identification and Beyond. Representations of Voting and Party Competition (London: Wiley), pp. 63–102.

Whiteley, P. F., Clarke, H. D., Sanders, D. & Stewart, M. C. (2013) Affluence, Austerity and Electoral Change in Britain (Cambridge: Cambridge University Press).

Zaller, J.R. (1992) The Nature and Origins of Mass Opinion (Cambridge: Cambridge University Press).

i While it is possible that random measurement error masks some true changes at the micro-level, the correlation of measurements at occasions t and t+1 will underestimate stability in the aggregate.

ii Similarly, Richardson (1991) claimed that in Britain partisanship did not resemble affective-laden identifications but rather cognitive partisanship.

iii Campbell et al (1960, p.135) proposed non-recursive rather than recursive effects between party identification and short-term attitudes, though they found party identification to be the predominant factor in the 1950s.

iv As Fiorina (1981: 90) suggests that party identification may also have non-political roots, this relationship might be attenuated.

v Dalton’s original measure of of cognitive mobilisation is an additive index of political interest and levels of formal education. We focus on political interest because it is more closely related to Fiorina’s idea of updating one’s identification based on the influx of new information, and also because average levels of formal education have risen sharply in Britain so that educational attainment is now closely linked to birth cohort membership.

vi It should be noted that the retention rates reported here refer only to those respondents who (net of any measurement error) were deemed to hold a perfectly stable identification across all four years. While this is a more conservative measure than looking just at consistency over the two ‘end’ points it does mean we exclude those too young to be included in all waves which may lead to some under-estimation of volatility levels, given the higher rates of switching commonly found among younger voters.

vii More recently, social identity theory has become quite popular in addressing party identification (e.g., Greene 2004), but we obviously have to rely on the traditional BES indicator..


Table 1: Theoretical Perspectives on the Expected Effects of Political Interest




Running tally”


Cognitive Mobilisation”

Motivated Reasoning”



Table 2: Model Fit


Free Parameters


Adjusted BIC


Mixed Markov




Mixed Markov + Interest




Scotland & Wales

Mixed Markov




Mixed Markov + Interest




Table 3 Effects of Political Interest






On Mover




On Lab-Identification




On Con-Identification




Scotland + Wales

On Mover




On Lab-Identification




On Con-Identification




Table 4: Mover/Stayer Property and Initial Party Identification (row percentages)

















Scotland + Wales











Table 5: Estimated Retention Rates 1997 → 2000 (per cent)









Scotland + Wales




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.

The AfD: Finally a Successful Right-Wing Populist Eurosceptic Party for Germany?


Germany is unusual amongst West European countries because all relevant parties (with the possible exception of the Left party) are unwavering supporters of European integration. Moreover, while the Radical Right is now a permanent feature a of many European democracies, the electoral successes of Germany’s Radical Right parties have been very modest and confined to the subnational level.

However, in 2013, only months before the General Election, a new party was formed that campaigned for a dissolution of the Eurozone and a radical re-configuration of German foreign policy. This new “Alternative for Germany” (Alternative für Deutschland or AfD for short) came tantalisingly close to the electoral threshold of five per cent. Nine months on, the party polled seven per cent in the 2014 European parliamentary election and was eventually admitted to the European Conservatives and Reformists group (ECR), which further soured the relationship between German Chancellor Angela Merkel and British Prime Minister David Cameron. In three (eastern) state parliamentary elections held in August/September 2014, the AfD did even better, capturing between 9.7 (Saxony) and 12.2 (Brandenburg) per cent of the vote.

The AfD has been described as eurosceptic and right-wing populist1 by its political rivals and by the mainstream media. If this description was correct, it would signal a qualitative shift in the structure of party competition in Germany. Moreover, due to such a party’s blackmail potential vis-a-vis the moderate right, this would constitute a massive shock to the German party system, with considerable implications for Germany’s future position on European integration and for German immigration policies. It is, however not at all clear if and to what degree such a classification of the AfD is warranted, as those terms are used rather indiscriminately in mediated discourses (Bale, Kessel, and Taggart 2011).

The aim of this article is therefore simply to assess the AfD based on categories derived from the rich comparative literature on the Radical Right and on euroscepticism. Since the AfD is a very young party with no parliamentary record, the primary source of evidence the party’s 2014 European manifesto. Additional information is drawn from material on the party’s website and Facebook presence.

The remainder of this article is organised as follows: The next section briefly reviews the concepts that will be used in the analyses. The third section provides some background information on euroscepticism and right-wing radicalism in Germany, and on the short career of the AfD. The fourth section presents an in-depth analysis of the AfD’s manifesto and other texts produced by the party. The final section summarises the main findings and puts them into perspective.


Radical right-wing populism

In the early 1980s, a new group of right-wing parties emerged in Western Europe. These parties differed significantly and systematically from mainstream parties of the right and were therefore portrayed as a new party family in the scholarly literature. While there is little disagreement as to which parties belong to this new family, research on these parties and their voters has been plagued by the twin questions of what exactly sets these parties apart from the mainstream right, and what adjectives (“radical”, “populist”, “extreme”, “anti-immigrant” …) best capture these differences.

More recently, Mudde (2007) has proposed a new scheme for classifying right-wing parties outside the mainstream that has won international acclaim because it accommodates a wide range of parties while identifying important differences between them. According to Mudde (2007, 19), the lowest common denominator for the party family is “nativism”, an ideology that combines nationalism and xenophobia. Nativism is a broad concept that subsumes racism, ethnocentrism, and anti-immigrant sentiment. Nativism holds that non-native elements (persons, ideas, or policies) present a threat to the nation state, which should be as homogeneous as possible.

However, traces of nativism may be found within the manifestos of mainstream parties. Following Mudde (2007, 21–23), to qualify as a Radical Right, a party additionally needs to display authoritarian tendencies, i.e. an aggressive stance towards political enemies and a preference for a strictly ordered society, strong leadership, and severe punishments for offenders. With authoritarianism comes a political bent that is not necessarily anti-democratic per se but goes against the grain of some of the fundamental values and principles of liberal democracy (Mudde 2007, 25–26) such as tolerance, pluralism, and the protection of minorities and their rights.

Within the Radical Right, Mudde then identifies a subgroup of parties that is also populist in nature. By populism, Mudde (2007, 23) means not just a style of political communication but rather a “thin ideology” (see Stanley 2008) that pits the “pure people” against a corrupt elite and puts majority rule above human rights and constitutional checks and balances. This Populist Radical Right is arguably the most electorally successful subtype within the larger party family.

Finally, a small (and not necessarily populist) subgroup of the Radical Right is actually anti-democratic. Borrowing from the long-standing practice in Germany, Mudde labels these parties as “Extreme Right”.

While Mudde’s system of definitions may not have ended the debate about terminology in the field, it obviously provides a useful tool for assessing new parties, and more generally for discussing developments in Germany within a wider European context.

Euroscepticism and ideology

Euroscepticism broadly refers to a negative stance towards European integration. As a field of scientific inquiry, it only took off in the late 1990s (see Vasilopoulou 2013) when the “Post-Maastricht Blues” (Eichenberg and Dalton 2007) kicked in. Mudde (2012) distinguishes between two main strands in this literature: the “North Carolina school”, which clusters around the Chapel Hill dataset, and the “Sussex school”, which chiefly relies on case studies of party manifestos. Whereas the “North Carolina school” aims at quantifying degrees of euroscepticism, the Sussex group introduced a qualitative distinction between “hard” and “soft” euroscepticism. “Hard” euroscepticism refers to a principled rejection of European integration that is ultimately incompatible with EU membership. “Soft” euroscepticism is not opposed to integration as such, but rejects the current state of European politics as well as the trajectory towards an “ever closer union” (Szcerbiak and Taggart 2008, 1:7–8).

Mudde (2012, 194) notes that in the face of low salience of Europe and euroscepticism, it can be surprisingly difficult to determine whether a party is soft eurosceptic, hard eurosceptic, or not eurosceptic at all. However, in the case of the AfD, which emerged as an anti-Euro party and drew up their first full-length manifesto in the context of the 2014 European election, there is clearly no lack of salience. The hard/soft distinction is therefore a useful template for the analysis of the AfD’s ideology.

But it is not entirely obvious how euroscepticism relates to broader ideologies. In most West European polities, parties position themselves within a two-dimensional (Kitschelt 1995; Benoit and Laver 2006) or perhaps even three-dimensional (Bakker, Jolly, and Polk 2012) space. The everyday language of politics, however, still relies on the traditional left-right-dichotomy, and most West Europeans are quite happy to place themselves on a unidimensional left-right-scale (Lo, Proksch, and Gschwend 2014). While the precise meaning of “left” and “right” may vary across time and space (Huber and Inglehart 1995), left-right-placement usually reflects the perceived distance between voters and parties as well as value-based preferences (Knutsen 1997). The focal value on the right-hand side of this spectrum is inequality (that can be the result of some “natural order”, mandated by tradition and authority, or the product of some market mechanism). The most important value on the left-hand side is equality (see e.g. Bobbio (1997)), realised through state regulation and redistribution of resource that may or may not infringe on liberty and property rights.

Data reported in Ray (1999)’s (1999) early seminal contribution suggest that eurosceptic parties were by and large located at both ends of the political spectrum during the 1990s. The analysis by Marks et al. (2006), which draws on more recent data, confirms precisely such a curvilinear pattern, at least for Western Europe. But even at the very (right) extreme of the political spectrum, euroscepticism is neither omnipresent nor universally “hard” (Vasilopoulou 2013). Therefore, a new party such as the AfD must be very carefully evaluated before it can be classified as right-wing, populist, eurosceptic, or all of the above.

Conditions for radical right success: Demand, context, and supply

Support for Radical Right parties varies considerably across time and political systems. The burgeoning literature has identified three groups of factors that can help to make sense of this variation: “Demand-side” variables refer to individual features such us gender, formal education, class, and, most importantly, political disaffection and anti-immigration attitudes (Brug and Fennema 2007) that make voters more or less susceptible to right-wing mobilisation. Their effect is moderated by contextual conditions, which include institutional factors (e.g. the electoral system or the degree of political centralisation), socio-economic conditions (e.g. the unemployment rate or the annual number of new asylum applications), and political variables such as the salience of the immigration issue for other parties and the media, or the willingness of the elites to co-operate with the Radical Right (Arzheimer 2009).

Demand-side and contextual variables collectively form the external environment to which a Radical Right party has to adapt, at least in the short term.2 A third group of variables, however, is more or less under the control of the party. Such “supply-side” factors include the party’s policy proposals, candidates for office, and general public appearance.3 Amongst these, past research has highlighted the availability of a “charismatic leader” as a precondition for Radical Right success, although this hypothesis is highly contested (Brug and Mughan 2007).

More recently, David Art (2011) has developed a more nuanced account that stresses the importance of party activists in general. In a nutshell, Art argues that the trajectory of Radical Right parties hinges on the nascent party’s ability to attract a sufficient number of the right type of activists, which in turn depends on historical legacies and the initial reaction of mainstream political actors to the new party (Art 2011, 31). To have a chance of electoral success, a new party needs many “moderates”, i.e. nationalists who credibly subscribe to the rules of liberal democracy and steer clear of biological racism and neo-nazism. Ideally, these moderates should also have high social-economic status (SES) and a degree of political experience outside the Radical Right (Art 2011, 33). Conversely, the emerging party should try to curb the number of “opportunist” members without strong political convictions and to avoid attracting any “extremist” activists who are openly hostile to parliamentary democracy: The latter group is prone to infighting over (highly fragmented) political principles, unwilling to temper their political demands in order to appeal to more moderate voters, and provide an easy target for any attempts to ostracise the new party. Moreover, they are often unexperienced (Art 2011, 35–40).

Whilst this article is chiefly concerned with the question of whether the AfD can at all be classified as Radical Right, Art’s theory of radical right mobilisation provides a useful template for the next section and puts the AfD’s electoral appeal in perspective.

Euroscepticism and the radical right in Germany

The lack of successful eurosceptic and right-wing parties

German MPs generally support European integration and even subscribe to a “deep core belief” of “the EU as a good thing” (Kropp 2010, 140). Almost all parliamentary parties in Germany are staunch proponents of European political integration. On the right, both the FDP and the CDU/CSU have supported and shaped European integration from its inception in the 1950s, although the small Bavarian CSU has been occasionally been more critical of the commission and some European policies than its sister party. On the left, the SPD and the Greens have taken a similar stance, both in opposition and in government (Wimmel and Edwards 2011, 295–296).

Only the Left party have voted consistently against the treaties of Maastricht, Nice, and Lisbon, because they reject the “neo-liberal” Single European Market, the monetary and budgetary policies mandated by the Stability and Growth Pact, and the “militaristic” Common Foreign and Security Policy. But even the Left party have declared themselves pro-European in principle at these occasions (Wimmel and Edwards 2011, 306–308), which makes them soft eurosceptics.

Otherwise, German euroscepticism has been confined to a number of unsuccessful single-issue fringe groups such as the Pro-Deutschmark party, and to the country’s three Radical Right parties (Lees 2008): the Republicans, the DVU, and the NPD. While all three had occasional successes in state elections and the Republicans even were represented in the 1989-1994 EP, their support proved fickle, and neither of them has ever won representation in the Bundestag. Compared to other West European countries, this weakness of the Radical Right appears anomalous and makes Germany a large negative outlier in a statistical model of radical right voting in Western Europe that controls for demand-side and contextual factors (Arzheimer 2009).

Art’s assessment of Radical Right mobilisation in Germany (Art 2011, 190–208), however, provides a plausible explanation for this lack of right-wing success. German elites have stigmatised National Socialism and criminalised the use its symbols very early on whilst offering nationalist a home in the mainstream centre-right. This strongly discouraged ‘moderates’ and ‘opportunists’ from joining the NPD, DVU, or the Republicans. These parties in turn have always had a fixation with the past (Ignazi 1992; Kitschelt 1995), which ruled them out as serious political players and made it easy to create and maintain a cordon sanitaire between them and the main parties.

Against this backdrop, the meteoric rise of the AfD and its ability to steer clear of any Nazi connotations is a very unusual4 and significant development. Arguably, this success was only possible because the party was formed by “moderates” with very high SES, considerable civic skills, and some political experience (detailed in the next seciont), and rests on the party’s continuing ability to ward off “extremists”, or at least activists that are perceived as too extreme in the German context. This makes the question of the AfD’s classification all the more pertinent.

The making of the “Alternative for Germany”

The AfD began its political life in September 2012 when a group of disaffected CDU members including Konrad Adam (born 1942), Alexander Gauland (born 1941), and Bernd Lucke (born 1962) founded a political action group called “Wahlalternative 2013” (an electoral alternative for the 2013 General election). While none of them played a leading role in the CDU, all three had been party members for several decades and were reasonably prominent figures: Adam and Gauland are well-known conservative journalists, while Lucke is a professor of economics who has been instrumental in organising two petitions by academic economists against the various bailout packages.

However, the AfD should not be considered a splinter party from the CDU, because the founding members were recruited from a broader centre-right background: Other signatories included 28 university professors (almost all of them economists), entrepreneurs and managers, and a former state party chair of the FDP (the Liberal party). The Wahlalternative’s short manifesto5 demanded that Germany should not guarantee any foreign sovereign debt, that all members of the Eurozone should be free to re-introduce national currencies or to join new currency unions, and that any further transfer of German sovereignty should be subject to a referendum.

Initially, the Wahlalternative was organised as a pressure group that supported the “Federation of Independent Voters”, a fledgling umbrella organisation for community-based, voter associations that are often dominated by the owners of small local business. In January 2013, both organisations jointly drew up a slate of candidates for the state election in Lower Saxony. However, the list polled just over one per cent of the vote, much less than the five per cent required for parliamentary representation. Subsequently, the two groups parted ways, and in February 2013 the Wahlalternative’s leadership formally founded the AfD as a political party, with the stated intention to run in the upcoming federal election on September 22. Adam, Lucke, and Frauke Petry (born 1975), a chemist and entrepreneur from the eastern state of Saxony, were elected to jointly lead the party.

By July, the party had drawn up a short manifesto that focused on monetary and fiscal policies, had set up branches in all 16 Länder, and had attracted more than 10,000 members.6 In the end, the AfD garnered 4.7 per cent of the vote, the best result for any party competing for the first time since 1953. While the AfD narrowly missed the electoral threshold, this result was widely seen as a remarkable achievement that gained them a foothold in the political system and gave them access to state funding.

Over the following six months, the party focused on broadening their programmatic profile and shedding the image of the single-issue party. During this time, it became clear that there was considerable potential for conflict within the party. In some state-level branches, the leadership resigned or was ousted over allegations of financial, political, or personal misconduct.7 More importantly, it became clear that various factions (conservatives, liberals, right-leaning Christian Democrats and perhaps even Christian fundamentalists) were warring for influence within the party. In January/February 2014, a party conference that was supposed to select the candidates for the European election had to be suspended for a week, because the delegates could not agree on a slate. In March, another party conference rejected a change to the statutes that would have made it possible for Lucke to become sole party leader.8 Lucke barely managed to take control of a debate on the party’s position on homosexuality (started by himself) and struggled to enforce a party line that stops short of open populism and hard euroscepticisim.

The most visible split within the party concerned the question of its future membership in a political group in the European Parliament. While Lucke was adamant that the AfD should join the ECR, some of the party’s rank-and-file and the party’s youth organisation “Young Alternative” would rather have worked with the Europe of Freedom and Democracy (EFD) group. Things came to a head when the Young Alternative invited Nigel Farage to give a lecture in Cologne.9 Lucke intervened but could neither forestall the event, nor was he successful in reprimanding Marcus Pretzell, one of the organisers and also a member of the party’s executive committee and a candidate for the EP election.

Electorally, none of this did the party any harm. In the polls, support for the AfD had been consistently in the range of six to eight per cent. In the actual election, they won 7.1 per cent of the vote, which entitled them to seven seats in the European Parliament – as many as the Left party and more than the CSU or the FDP have won. The list of elected candidates includes only two women and reflects the bourgeois background of the party leadership.10 On June 12, the seven joined the ECR, making it the third-largest faction in the EP.

An analysis of the “Alternative”

A quantitative analysis of the AfD’s 2013 European election manifesto

Even for European elections, German parties tend to formulate detailed manifestos that cover a lot of policy domains. The AfD is no exception to that rule. The 2014 European election manifesto is the party’s first full-length policy document and therefore very well suited for assessing the party’s official policy positions. To provide context for its analysis, the manifestos of the main parties, the leftist Pirates, and the right-wing extremist NPD were analysed, too.

At 4,894 words, the AfD’s manifesto is close to the median length of 5,852 words. A number of function words (articles, conjunctions, prepositions etc.) were removed from the files. Running headers or footers, tables of contents and adverts were also discarded, but preambles and prefaces by the party leaders were retained. Because German is an inflected language, the “Snowball” stemming algorithm was applied to prepare the texts for quantitative analysis. Stemming aims at reducing words to their roots by removing suffixes and affixes so that different inflected forms are grouped together as a single item. While stemming is less accurate than full lemmatisation (determining the dictionary form of inflected words), it can be carried out quickly and efficiently to reduce the complexity of a text, and the loss in precision does not matter much in practical applications (Grimmer and Stewart 2013, 272).

Although a number of very common German words had been discarded in the first step, some stems such as “Europ” and “EU” appear very frequently in all manifestos and are thus not useful for discriminating between parties. Therefore, the one per cent most frequent stems were removed. Following Grimmer and Stewart (2013, 273), very rare stems that collectively make up one per cent of the total corpus as well as stems that were exclusively used by a single party (typically the party name) were also disregarded.















human rights














citizens (female)












citizens (female)



member state












(German) people





Table 1: The Five Most Frequent words/stems in Nine Election Manifestos

The remaining 4,430 stems give a very clear impression of the AfD’s priorities. Amongst the 15 most frequent concepts in the AfD manifesto are “member states”, “Eurozone”, “ECB”, and “institutions”. None of these words is amongst the top priorities of any other party. However, the analysis also reveals some similarities. “Competition” features prominently in the AfD’s manifesto, but also crops up frequently in the respective platforms of the FDP and the CDU. “Work” is a common concern of the Left party and the SPD, and both Christian Democratic parties frequently talk about “jobs”. Even looking at just the top five words most frequently used by each party gives a good idea of what they stand for (see Table 1).

The observation that the usage of certain words conveys information on ideological proximity and distance between parties has been formalised by Slapin and Proksch (2008), who derive a statistical model that links word frequency to an underlying left-right dimension. Slapin and Proksch (2008) also develop an estimation procedure they call “wordfish”, which recovers ideological positions from political texts and ideological content of words while controlling for differences in the wordiness of political documents and the global distribution of words. Unlike the related “wordscore” method (Laver, Benoit, and Garry 2003), “wordfish” does not require anchor texts and is thus ideally suited for uncovering the positions of new parties relative to a set of more familiar political actors.

For the present analysis, the words and nine parties were simultaneously scaled using version 1.3 of Slapin and Proksch’s wordfish package for the R statistical system. The algorithm converged quickly on the point estimates. 95 per cent confidence intervals were generated by a parametric bootstrap procedure (a method that does not rely on a normal distribution of the estimates) using 500 draws (Slapin and Proksch 2008, 710).11 Again, there were no convergence problems. The words12 most closely tied to left ideology are “Kürzungspolitik” (austerity policies), “erwerbslos” (unemployed), “Altersarmut” (pensioner poverty), “Migrantinnen” (an inclusive and neutral term for migrants), “Sozialcharta” (social charter), “EU-Politik” (EU politics), “Profit” (profit, a more derogatory term than “Gewinn”), “Rüstungsproduktion” (production of arms), “unbefristed” (open-ended, as in open-ended contract), and “nationalistisch” (nationalistic).

The most right-wing words are “fremd” (foreign or strange), “Volk” (the (German) people, in a very emphatic sense), “verhängnisvoll” (fatal or ominous), “einerseits” (on the one hand), “bürgerfern” (removed or insulated from the interests of ordinary citizens), “Asylbewerber” (asylum seekers), “Gender” (as in gender mainstreaming or similar bugbears of the right), “gängeln” (to boss around someone, typically used with reference to the behaviour of bureaucrats), “Bolognaprozess” (the implementation of the Bolgna accord in German Higher Education), and “schleichend” (creeping, typically referring to slow but sinister political change). As this vocabulary reflects both the socio-cultural and the economic dimension of the left-right dichotomy, the scaling displays a high degree of face validity.

Scaling the AfD EP 2014 Manifesto

Scaling of German election manifestos for the 2014 EP election

Estimates for the party positions are very precise (Figure 1). For most parties, the width of the 95 per cent confidence interval is 0.10 points or less on a scale that ranges from -1.69 to 1.26. Crucially, the interval for the AfD is one of the narrowest. The positioning of the parties themselves will be instantly familiar to any student of German politics. The political spectrum is spanned by the Left party on the one hand and the NPD on the other. The Social Democrats (SPD) and the Christian Democrats (CDU) appear in their familiar centre-left and centre-right positions, with the Greens positioned to the left of the SPD and the FDP to the right of the CDU. The Pirates are located between the Greens and the SPD, which again seems plausible.

Both the CSU and the AfD appear to the right of the FDP, slightly closer to the NPD than to the CDU. The confidence intervals for their positions overlap, which implies that they are statistically indistinguishable.13 Lucke has repeatedly claimed that his party is neither left nor right14 and even stated that the AfD represents a new breed of party (“Partei neuen Typs”) at their founding conference15 – a very awkward pun on the Stalinisation of East Germany’s Socialist Unity party in the 1940s. But their manifesto places them firmly at the far right of the political spectrum.

The position of the CSU is perhaps more surprising, because the Bavarian Christian Democrats have been a fixture of German politics since 1945. But the party has nonetheless been described as anti-immigration and (borderline) right-wing populist in the literature (Lubbers, Gijsberts, and Scheepers 2002; Falkenhagen 2013). Former leader Franz-Josef Strauß famously declared that “there must be no democratic party right of the CSU” (Raschke and Tils 2013, 253, my emphasis). More recently, the party has also steered an ambiguous course towards the EU.16 While the content of their 2014 manifesto may already reflect concerns about the emerging competition from the AfD, the document is nonetheless in line with the CSU’s traditional position at the very margin of the established party system.

Is the 2013 manifesto radical, populist, and eurosceptic?

Against this backdrop, the estimates for both the AfD and the CSU are highly plausible. The general left-right measure paints, however, a very broad-brushed picture of the AfD’s political program. Assessing the question whether the AfD is not only on the right but also radical/extremist, populist, and eurosceptic requires a close reading of its manifesto.

In the theory section, “nativism”, i.e. a mixture of nationalism and xenophobia was proposed as a criterion for separating the Radical Right from other right-wing parties. In line with their overall position on the right of the political spectrum, the AfD is certainly unusually prone (by German standards) to display national symbols and to emphasise Germany’s national interest. “Mut zu Deutschland” (roughly translated: dare to stand by Germany) was the title of their manifesto and their main slogan for the EP 2014 campaign. The phrase is still used prominently on the party’s main website, their social media profiles, and in other party material. The slogan alludes to the common right-wing argument that national pride is systematically discouraged in Germany but was deployed in a more specific sense during the campaign: the AfD wants Germany to act more assertively within the European Union.17

The corresponding section, however, is one of the shortest in the manifesto and makes rather modest demands. The AfD blames the member state governments for breaking the treaties (particularly the Stability and Growth Pact), it demands that the EP should launch a public inquiry into the details of bailout measures, and it suggests (without going into details) that Germany should have a greater say within the European institutions. However, the main opponent for the AfD is an unholy alliance between the EU institutions and Germany’s “Altparteien” (old, i.e. established parties – a term the AfD has borrowed from the Green party of the 1980s). But one would be hard-pressed to find any statement that is nationalistic in the usual sense of the term in this or in fact in any other part of the manifesto.

The section on immigration and asylum also strikes a rather conciliatory tone. The AfD subscribes to the principles of free movement and free choice of residence for all EU citizens, although they want to limit benefits (of which they are critical in general) to long-term residents and their offspring. Moreover, the AfD acknowledges the problems brought about by demographic change and supports a point-based immigration regime for non-EU citizens. Finally, the AfD commits itself to a “humane” asylum system, which implies more financial and logistic support for the member states in the South, common standards for accommodation, and labour market access for asylum seekers. Taken together, these positions are not overly restrictive by German standards and do not display any nativist tendencies.

The AfD rejects Turkish EU membership flat-out and mentions “geographical, cultural and historical borders” in this context. But apart from this, and from a single reference to Europe’s “Christian-occidental values”, religion and culture, which are often used as politically acceptable codewords for non-European ethnic groups (Zúquete 2008), are not at all mentioned in the text. Judging by its manifesto, the AfD is therefore not a Radical Right, let alone an Extreme Right party.

Is the AfD populist? If one defines populism as a “thin ideology”, then there is very little in the manifesto that would support such a claim. The AfD is highly critical of “Brussels”, and of the mainstream parties in Germany. They also argue that the ongoing financial crisis was to a large degree caused by irresponsible behaviour of the banks, which should be regulated more tightly. Moreover, they want to improve the democratic legitimacy of the EU in general and demand that future enlargements as well as important decisions on the Euro should be put to a referendum.

But that alone does hardly make them populists. Their manifesto does not contain a single reference to “elites”, the “political class”, or the “eurocrats”. Corruption is mentioned only once, in the innocuous context of the UN’s anti-corruption charter. But even if one opts for a broader, softer definition that primarily treats populism as a style of political communication “that refers to the people” (Jagers and Walgrave 2007, 322) there is nothing in the manifesto that would appear as particularly populist in that sense.

The AfD’s manifesto does not even conform with every day notions of populism that imply appeal to emotions, oversimplification, and a degree of opportunism (Mudde 2004, 542–543). On the contrary: The AfD’s manifesto contains lengthy references to economic theory, is largely written in a rather technical and stilted language and even contains a couple of footnotes that cross-reference political demands to articles in the Treaty on the Functioning of the European Union.

That leaves the issue of euroscepticism. The AfD is clearly not a “hard” eurosceptic party. They are opposed to the currency union in its present form, to current and future bailouts, and more generally to a federal European state. But at the same time, they are committed to the European Union as such and have dropped their erstwhile demand for a return to the Deutschmark from their manifesto. While they want to strengthen the principle of subsidiarity (which was established in the Treaty of Maastricht at the behest of the German Länder), they don’t intend to reduce the EU to a trade bloc. Although they are highly suspicious of secretive intergovernmental co-operation in Justice and Home Affairs, they support the pursuit of a Common Foreign and Security Policy based on lowest common denominator solutions. Taken together, “soft euroscepticism” best describes the political positions articulated in the manifesto.

The Alternative’s internet presence

In Germany, parties are legally obliged to draw up comprehensive manifestos and lodge them with the Federal Returning Officer. These platforms are routinely scrutinised by researchers and the media. The lack of any obviously radical and populist content in the AfD’s manifesto could therefore be misleading. Indeed, a speech delivered by Konrad Adam on June 27, 201318 gives a rather different impression. Adam encourages party members to become “dangerous citziens” (“gefährliche Bürger”) who dare to take on the elites. Politicians of other parties are portrayed as greedy, lazy, and incompetent predators who are after the money of ordinary taxpayers and sell out the national interest to the EU, and the mainstream media help them to cover up.

Other speeches documented on the website, however (four by Lucke, one by Starbatty and one by Henkel) strike a similar, yet clearly more moderate tone. Starbatty and Henkel in particular discuss intricacies of social and economic policy in great detail, while Lucke often focuses on his vision for the further development of the party. None of these speeches could be considered populist or radical.19

To get a more rounded impression of the party’s appeal, it therefore makes sense to analyse the party’s presence on the internet. The party’s main website ( is built with the wordpress platform. It is professionally designed and maintained and currently (July 2014) consists of more than 1,300 unique HTML pages20 plus 62 PDF documents. Many pages contain redundant content because they serve as archives that bring together posts related to a specific author, tag, category, or date of publication. Other pages are simply of an administrative nature (e.g. contact information). The following analysis is therefore restricted to 371 blog-post like pages, which contain comments on media reports, current events, or simply document statements by prominent party leaders.

For the analysis, only text in the main body of the pages was extracted. Stopword removal and stemming were conducted as outlined above. The resulting corpus consists of 45,990 words, which can be reduced to 9,745 stems. Obviously, posts on a website serve a function that is different from that of the manifesto, and this is reflected in the language used. A simple count demonstrates that the AfD is chiefly talking about itself and its leadership. “AfD” (573), “Alternative” (338), and “Deutschland” (“Germany”, 531) are amongst the five most popular stems. Also very prominent is the name of Bernd Lucke (274), who clearly overshadows his fellow leaders Adam (42) and Petry (48). Far more important than Adam and Petry are deputy leaders Gauland (151) and Henkel (90), while (female) deputy leader Patricia Casale is not mentioned at all.21

Taken together, the website leaves no doubt that the AfD is a right-wing party. Only about 60 per cent of the references to “Germany” are due to the use of the full party name. The party simply talks a lot about Germany, everything German (253), Europe (379), the Euro (327), the EU (175), and the Eurozone (62). The tone is slightly harsher than that of the manifesto, with the occasional attack on the ECJ or refugees “who abuse the right to hospitality”. While “Volk” (people) is rare, the more intellectual “Bevölkerung” (population, 42) and particularly “Bürger” (citizen(s), 169) crop up much more often. Attacks on the AfD’s political competitors are also frequent: the CSU is mentioned 41 times, the FDP receives 75 references, and the CDU is mentioned 104 times.

But all in all, there is still little evidence of populism or right-wing radicalism. Immigrants and immigration are mentioned only 23 times (equivalent to a single mention in six per cent of all posts), and not necessarily in a negative context. Bulgaria and Romania are each referenced less than 15 times, Muslims hardly play a role at all, and even Turkey and the Turks are mentioned only 23 times. Remarkably, the AfD shows an unusual degree of sympathy for Russia and distrust for the US, both common in German Radical Right circles. But as far as foreigners are concerned, the main focus is clearly Greece and the Greeks (297 references).

The main website does not, however, include any interactive elements (guest books, comments, fora). Instead, the party relies on social media websites to interact with members, supporters, the media, and the general public.
Facebook is of particular importance for the party. As of July 2014, the official fanpage of the AfD’s federal organisation counts almost 122,000 “likes”.22 This is nearly twice as much as the SPD (just under 75,000) or the CDU (almost 84,000) can muster. On Twitter, the AfD federal organisation’s handle has only about 9,600 “followers”. The analysis will therefore focus on the AfD’s Facebook fanpage.

From a party’s point of view, fanpages are attractive because they provide a focus for political conversation about the party that is actually under the control of the party. Whereas communication on Twitter is largely unmoderated, spontaneous, and ephemeral, Facebook fanpages resemble traditional home pages. Crucially, fanpage administrators can remove posts, restrict who may post, and even ban individual users from the page.

Montage of Soundbite and Party Logo

Facebook has created an application programming interface (API) that makes it easy to programmatically access posts on fanpages as well as their meta data. Data were collected using version 0.4 of the “Rfacebook” package for R. The AfD launched its fanpage on March 3, 2013, and posted for the first time on March 7. As of July 11, 2014, the AfD have updated their status 1,702 times, or roughly 24 times per week, with markedly higher frequencies immediately after the launch of the party, during the federal campaign, during the candidate selection conference in January/February, and finally during the European parliamentary campaign.23 Many of these updates include images that combine text and pictures. Figure 2 is quite a typical specimen that brings together a photo of Gauland, a short quote (“Germany is not the USA’s doormat”), and the party logo and signature blue background. It is well known that such photo updates create quicker and stronger reactions with Facebook users and are privileged by Facebook’s selection algorithm. Moreover, text within images is never truncated by Facebook and can be easily shared both on Facebook and across other channels with minimal effort.

The use of such imagery shows the professionalism of the AfD’s social media team but is an obstacle for text analysis, because the use of various fonts and designs renders reliable optical character recognition virtually impossible. Fortunately, most images are complemented by some text, which usually re-iterates the main points or raises some additional issues. The following analysis is based on 1,223 posts that contain at least some text. Together, these posts make up some 72,000 words.24

Facebook is a popular medium for political communication because links to other content on the internet can be quickly posted (often with a preview of the other site’s content), distributed, and commented upon. Until July, the AfD had posted 1,622 unique URLs which point to 187 separate domains. From these figures, it is clear that the AfD does not simply use Facebook to re-publish the content of its main website. Indeed, there are only 69 links (less than five per cent) to the party’s national website, and 70 links to the websites of 19 local or regional party organisations.

The vast majority (795) of links refer to other content on Facebook. The party also makes extensive use of video clips hosted on (79). Amongst the other sites, (the right-most mainstream broadsheet) and (a centre-right broadsheet) are particularly prominent with 109 and 60 links, respectively. Other important mainstream sources include the business news sites (68) and (29) as well as news magazines (92, centre-right) and (58, centre-left). Finally, there is a host of links to various blogs and other websites.

In summary, the AfD use their Facebook page to direct attention to news articles that support the party’s positions, to “spin” stories on issues that will chime with their supporters, and occasionally to poke fun at their political adversaries. But what exactly are they talking about in the text that accompanies links, videos, and images? Amongst the most frequent words in the posts are once more “AfD” (1,182), “Germany” (701), “Euro” (488) “alternative” (466), “EU” (400), and “Lucke” (367). Again, other party leaders are mentioned far less often. Greece (together with Greeks and “Athens”) feature prominently once more with 202 mentions and are far more important than Turkey/Turks (48), Muslims/Islam (20), or Romania (9) and Bulgaria (8).

Apart from that, it is slightly easier to find populist rhetoric on the Facebook page than on the main website. While “elites” (which by any reasonable definition would include many AfD leaders) are hardly ever mentioned, there are ample references to a conflict between “politicians” and “citizens” as well as many calls for protecting “freedom” and “democracy”. But even the most overtly populist post, the party anthem “We don’t give up” is relatively tame: Germans are a “really super people” (“ein wirklich tolles Volk”) who nonetheless “suffer”, Chancellor Merkel is accused of treating “us” like a bunch of “right-less monkeys” while politicians more generally are guilty of writing incomprehensible and self-serving laws. The only solution to this crisis is to vote for the AfD.

Collectively, the AfD’s posts were “shared” (copied to users’ profiles or other pages) more than 500,000 times. They received more than 1.9 million “likes” (on average, more than 1,100 per post) and over 325,000 comments, which amounts to over 200 comments per post.

Bloggers and mainstream journalists have repeatedly suggested that the AfD buys phantom fans and fake likes on Facebook. Comments, however, are much more difficult to simulate than shares and likes, and even a cursory glance at the AfD’s page shows a remarkable level of real political interaction between users.

Somewhat surprisingly, not just “fans” but any Facebook user may comment and even post new content on the AfD’s page. As of July 11 2014, almost 79,000 user-generated posts are accessible on the page. This is roughly equivalent to a corpus of 3.4 million words. Together, the posts have attracted more than 212,000 comments and just over 51,000 shares.

While members and supporters dominate, some critical voices exist amongst those who post on the AfD’s page. At 7,980, the number of original posters is relatively low, and the distribution of posts across users is heavily skewed to the right (18.7). The median number of posts per user is just one. A minority of five per cent has posted 30 times or more, and a tiny group of 25 users (less than 0.5 per cent) is collectively responsible for a quarter of all posts.25 While some of these 25 show off their sympathy for the AfD in their profile pictures and three hold party offices at the local level, none of them plays any significant role within the party or holds public office.

In terms of content, the 79,000 user-generated posts resemble the material posted by the AfD themselves. Again, “AfD” (26,311), “Germany” (12,816), “Euro” (11,152), and EU (9,295) are amongst the most frequent words, while “Bulgaria”, “Romania”, “Turkey”, and Muslims/Islam are of lesser importance. However, quite a few posts strike a tone that is markedly different from the party’s carefully crafted statements. Resentment and nationalism colour many posts. Complaints about ungrateful immigrants, privileged homosexuals, and greedy politicians are frequent. Links to obscure right-wing sites abound.

The AfD have created a space for their supporters where this kind of talk is tolerated. But only up to a point: Racist slurs and even common expletives are very rare. This does not prove, but suggests, continuous interventions by the party. In various comments, the AfD has made it clear that they are actively monitoring the page, and that they delete racist or otherwise illegal content including links to right-wing extremist websites. There is no way of knowing how many items have been posted and subsequently deleted, but the party is treading a narrow line. On the one hand, the AfD does not want to annoy their most vocal supporters on the internet, on the other, Lucke is very wary of allegations of populism and radicalism.


This article set out to answer the question whether the AfD is a right-wing, populist, and eurosceptic party. A careful quantitative and qualitative analysis of the 2014 EP manifesto shows that the AfD is indeed located at the far-right end of Germany’s political spectrum because of their nationalism, their resistance against state support for sexual diversity and gender mainstreaming, and their market liberalism. They do, however, not qualify as “radical”: There is no evidence of nativism or populism in their manifesto, which sets them apart from most of the other new right parties in Europe. Moreover, their euroscepticisim is of the “soft” variety. This assessment is largely confirmed by an analysis of the AfD’s communication on the web, although statements by their facebook fans hint at more radical currents amongst supporters and the party rank-and-file.

Important nuances not withstanding, their current programmatic appeal most closely resembles that of the CSU. But while the CSU is essentially an “ethno-regional” (Falkenhagen 2013) party that does not stand candidates outside Bavaria, the AfD aims at attracting a much bigger and broader national constituency.

Continued electoral support for the AfD would have profound repercussions for the existing German party system, most obviously by undermining the position of the CDU, which so far have fared much better than other Christian Democratic parties in Western Europe (Bale and Krouwel 2013). In the longer run, it would directly or indirectly affect domestic, immigration, and European integration policies. Thus far, Chancellor Merkel has ruled out coalitions with the AfD. If the FDP’s decline proves permanent and the AfD prevails, maintaining this cordon sanitaire will weaken the position of the CDU by forcing it to exclusively enter coalitions with the parties of the Left. At the same time, the AfD’s success is already fuelling internal backlash from conservatives against Merkel’s socially liberal policies.

Irrespective of the AfD’s perspectives for long-term survival, the fact that the party has done so well in five consecutive elections reflects the scope of partisan dealignment in Germany and the increasingly fluid nature of its party system. But does this remarkable mobilisation guarantee a bright future for the AfD? Not necessarily. In absolute terms, the AfD’s support has essentially stagnated, although the timing of the elections was close to optimal: In 2013, the AfD won 2.06 million votes, in the 2014 election it was 2.07 millions. In the eastern state elections, the AfD even suffered a small net loss of some 18,000 votes compared to their state-level results in the European election. Euroscepticism, the party’s current core issue, is still not very salient in Germany. In a survey immediately before the EP election, only 47 per cent of the AfD’s own voters had a wholly negative view of Germany’s membership in the EU. Only 22 per cent rejected both Juncker and Schulz as president of the commission, and 40 per cent said they supported the AfD to register a protest vote, not because of their policies.

Precisely what these policies are might therefore change over the near future. So far, the “moderates” have dominated the party leadership. Lucke has been able to commit the AfD to civic nationalism, financial prudence, and soft euroscepticisim, with Gauland and Adam catering for a more broadly “liberal-conservative” right-wing audience that may feel left behind by Merkel’s move to the centre. In 2014, Lucke’s creation more closely resembles the British Conservatives than UKIP, let alone the French FN or the Austrian Freedom Party.

Yet Lucke’s control over the party seems to be wavering. His plans to replace the joint leadership structure with a more traditional sole-leader role have been rejected by a party conference and have been met with criticism from his colleagues.26 State and district party chapters are still struggling to keep right-wing extremists out. Meanwhile, their new status as MEPs has given other leading figures such as von Storch and Pretzell a platform. They represent less savoury brands of right-wing politics that could ultimately prove more attractive to voters than Lucke’s polite exercises in economic theory. Just how long the party resists that temptation remains to be seen.

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

Arzheimer, Kai. 2009. “Contextual Factors and the Extreme Right Vote in Western Europe, 1980–2002.” American Journal of Political Science 53 (2): 259–275. doi:10.1111/j.1540-5907.2009.00369.x.

Bakker, Ryan, Seth Jolly, and Jonathan Polk. 2012. “Complexity in the European Party Space: Exploring Dimensionality with Experts.” European Union Politics 13 (2): 219–245. doi:10.1177/1465116512436995.

Bale, Tim, and André Krouwel. 2013. “Down but Not Out. a Comparison of Germany’s CDU/CSU with Christian Democratic Parties in Austria, Belgium, Italy and the Netherlands.” German Politics 22 (1-2): 16–45. doi:10.1080/09644008.2013.794452.

Bale, Tim, Stijn van Kessel, and Paul Taggart. 2011. “Thrown Around with Abandon? Popular Understandings of Populism as Conveyed by the Print Media. a UK Case Study.” Acta Politica 46 (2): 111–131. doi:10.1057/ap.2011.3.

Benoit, Kenneth, and Michael Laver. 2006. Party Policy in Modern Democracies. London: Routledge.

Bobbio, Norberto. 1997. Left and Right. the Significance of a Political Distinction. Chicago: University of Chicago Press.

Brug, Wouter van der, and Meindert Fennema. 2007. “Causes of Voting for the Radical Right.” International Journal of Public Opinion Research 19: 474–487. 10.1093/ijpor/edm031.

Brug, Wouter van der, and Anthony Mughan. 2007. “Charisma, Leader Effects and Support for Right-Wing Populist Parties.” Party Politics 13 (1): 29–51. doi:10.1177/1354068806071260.

Eichenberg, Richard C., and Russell J. Dalton. 2007. “Post-Maastricht Blues: the Transformation of Citizen Support for European Integration, 1973-2004.” Acta Politica 42 (2-3): 128–152.

Falkenhagen, Frédéric. 2013. “The CSU as an Ethno-Regional Party.” German Politics 22 (4): 396–420. doi:10.1080/09644008.2013.833190.

Grimmer, Justin, and Brandon M. Stewart. 2013. “Text as Data. the Promise and Pitfalls of Automatic Content Analysis Methods for Political Texts.” Political Analysis 21 (3): 267–297. doi:10.1093/pan/mps028.

Häusler, Alexander, and Rainer Roeser. 2014. Rechtspopulismus in Europa Und Die Rechtspopulistische Lücke in Deutschland. Erfurt: MOBIT e.V.

Huber, John, and Ronald Inglehart. 1995. “Expert Interpretations of Party Space and Party Locations in 42 Societies.” Party Politics 1: 73–111.

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

Jagers, Jan, and Stefaan Walgrave. 2007. “Populism as Political Communication Style: an Empirical Study of Political Parties Discourse in Belgium.” European Journal of Political Research 46: 319–345.

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

Knutsen, Oddbjørn. 1997. “The Partisan and the Value-Based Component of Left-Right Self-Placement: a Comparative Study.” International Political Science Review 18 (2): 191–225. doi:10.1177/019251297018002005.

Kropp, Sabine. 2010. “German Parliamentary Party Groups in Europeanised Policymaking: Awakening from the Sleep? Institutions and Heuristics as MPs’ Resources.” German Politics 19 (2): 123–147. doi:10.1080/09644001003774198.

Laver, Michael, Kenneth Benoit, and John Garry. 2003. “Extracting Policy Positions from Political Texts Using Words as Data.” American Political Science Review 97 (02): 311–331. doi:10.1017.S0003055403000698.

Lees, Charles. 2008. “The Limits of Party-Based Euroscepticism in Germany.” In Opposing Europe? the Comparative Party Politics of Euroscepticism, edited by Aleks Szcerbiak and Paul Taggart, 1:16–37. Oxford: Oxford University Press.

Lo, James, Sven-Oliver Proksch, and Thomas Gschwend. 2014. “A Common Left-Right Scale for Voters and Parties in Europe.” Political Analysis 22 (2): 205–223. doi:10.1093/pan/mpt028.

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

Marks, Gary, Liesbet Hooghe, Moira Nelson, and Erica Edwards. 2006. “Party Competition and European Integration in the East and West Different Structure, Same Causality.” Comparative Political Studies 39 (2): 155–175. doi:10.1177/0010414005281932.

Mudde, Cas. 2004. “The Populist Zeitgeist.” Government and Opposition 39 (3): 541–563.

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

———. 2012. “The Comparative Study of Party-Based Euroscepticism. the Sussex Versus the North Carolina School.” East European Politics 28 (2): 193–202.

Raschke, Joachim, and Ralf Tils. 2013. Politische Strategie. Eine Grundlegung. 2nd ed. Wiesbaden: VS Verlag für Sozialwissenschaften.

Ray, Leonard. 1999. “Measuring Party Orientations Towards European Integration: Results from an Expert Survey.” European Journal of Political Research 36 (2): 283–306. doi:10.1111/1475-6765.00471.

Slapin, Jonathan B., and Sven-Oliver Proksch. 2008. “A Scaling Model for Estimating Time-Series Party Positions from Texts.” American Journal of Political Science 52 (3): 705–722.

Stanley, Ben. 2008. “The Thin Ideology of Populism.” Journal of Political Ideologies 13 (1): 95–110.

Szcerbiak, Aleks, and Paul Taggart, ed. 2008. Comparative and Theoretical Perspectives. Opposing Europe? the Comparative Party Politics of Euroscepticism. Vol. 1. Oxford: Oxford University Press.

Vasilopoulou, Sofia. 2013. “Continuity and Change in the Study of Euroscepticism. Plus Ça Change?” Journal of Common Market Studies 51 (1): 153–168. doi:10.1111/j.1468-5965.2012.02306.x.

Wimmel, Andreas, and Erica E. Edwards. 2011. “The Return of ‘Social Europe’: Ideas and Positions of German Parties Towards the Future of European Integration.” German Politics 20 (2): 293–314. doi:10.1080/09644001003774206.

Zúquete, José Pedro. 2008. “The European Extreme-Right and Islam: New Directions?” Journal of Political Ideologies 13 (3): 321–344. doi:10.1080/13569310802377019.

1E.g. Stuttgarter Zeitung 18/09/2013, page 4; TAZ 02/07/2014, page 5.

2Most parties will of course try to alter this environment to their advantage.

3In a broader sense, the party’s organisational structures and deployment of campaign funds could also go under that rubric, although these obviously depend on recruitment and supply of external resources.

4The ‘Law and Order Party’ (PRO) of the early 2000s initially took a similar approach, but was not interested in euroscepticism and remained confined to the city-state of Hamburg.

5The original manifesto is archived at .

6Merkur Online, (03/07/2014).

7RP Online 01/12/2013, (03/07/2014), Spiegel Online 28/12/2013, (03/07/2014), Zeit Online 14/06/2014, (03/07/2014).

8Süddeutsche Online 23/03/2014, (03/07/2014).

9Deutsche Welle Online 29/03/2014, (03/07/2014).

10Lucke himself, Hans-Olaf Henkel, the former president of the Umbrella Organisation of German Industry (BDI), who favours a minimal state and has likened the EU to the former Soviet Union(Handelsblatt 03/10/2011, (03/07/2014)), Bernd Kölmel, a public servant with the Baden-Württemberg State Court of Auditors, Beatrix von Storch, an insolvency lawyer and fringe Christian-conservative homphobe activist, Joachim Starbatty, a retired professor of economics who has repeatedly (though unsuccessfully) sued the government over the Euro, Ulrike Trebesius, a civil engineer, and the aforementioned Marcus Pretzell, a lawyer and property developer.

11The Graphs show the average of the boot-strapped point estimates. For the parameters that determine word “loadings” on the ideological dimension, these tend to differ somewhat from the maximum likelihood estimates, but for the parameters, the two sets are virtually identical.

12For the sake of readability, words are used here instead of the actual stems.

13However, in the vast majority of the 500 bootstrap samples, the CSU is estimated to be slightly more right-wing than the AfD.

14E.g. Spiegel Online 22/03/2014, (07/07/2014).

15Zeit Online 18/03/2013, (07/07/2014).

16With the tacit blessing of the leadership, a group of backbenchers have voted against the bailout legistlation and subsequently asked the Federal Constitutional Court to nullify these bills. The party’s core political project in the 2013-2017 parliament is a special road charge for cars registered abroad that would probably violate of EU law, and the central plank of their 2014 campaign was the slogan “kick welfare cheats out”, which referred to alleged “benefit tourists” from the eastern EU member states.

17This interpretation was emphasised by the design of the campaign posters, which surrounded the “EU” in “Deutschland” with the 12 European stars (shown in Figure 2).

18 .

19More generally, Lucke and his party are the object of very intense scrutiny by their political adversaries and the mainstream media, yet there are very few verifiable public statements by Lucke or other members of the national leadership that could qualify as right-wing populist (see the borderline examples in Häusler and Roeser 2014, 37).

20This number excludes automatically generated overview pages for authors, tags, categories, and year of publication.

21Another relatively prominent male is Starbatty (20), now a MEP. On the other hand, there are only five references to female MEP von Storch , and the second female MEP Trebesius is mentioned just once. The image of the party leadership that the website projects is reflected in the coverage by the German media: For the period from February 1 2013 to July 11 2014, LexisNexis lists 2096 news items mentioning both the AfD and Lucke and 622 items that refer to Henkel and the AfD. Adam, however is mentioned only 232 times, Gauland 231 times and Petry 185 times. von Storch is referenced 169 times, Starbatty 150 times, Trebesius 20 times, Pretzell four times, while there is just one article that mentions Casale.

22All 16 state level chapters as well as various regional and local chapters have set up their own fanpages, but those have much smaller fanbases, ranging from several thousands to less than a hundred.

23This number could be inflated, as the Facebook API returns about 264 posts which do neither contain messages nor links, and which created no reactions. Presumably, these are either drafts or were retracted.

24This number includes URLs and symbols such as the hashtag sign.

25The number could be even smaller, as three of the most prolific posters have very similar surnames: Otto Blank, Andrea Blanc, and Andrea Cnalb.

26FAZ Online 20/10/2014, (07/11/2014), Spiegel Online 29/11/2014, (29/11/2014).

A New Multinomial Accuracy Measure for Polling Bias


The ungated final version is available here (PDF) and here (HTML).

To install our Stata add-on, type ssc install surveybias or click here.

For replication scripts and data, click on “Data”.

A New Multinomial Accuracy Measure for Polling Bias

1 Introduction

Work on pre-election polls forms a vital part of election analysis and forecasting in both academic publications and media coverage. Frederick Mosteller (Mosteller et al.1949) introduced the notion of accuracy measures to assess polls against actual election results. Those indices are designed for two-party/-candidate races, and cannot easily be applied to multi-party/-candidate elections. An equivalent index has not so far been derived for multi-party elections, limiting the ability of researchers to measure overall polling accuracy in such cases.

Starting from the predictive accuracy measure proposed by Martin, Traugott and Kennedy (2005), we propose such an accuracy measure, B, for elections with more than two parties or candidates. First, we derive this index mathematically from an implementation of the multinomial logistic model, including the relevant tests of statistical significance. We then consider how this aggregate measure may be biased, given it is based upon compositional data, and use a simulation to examine the extent of this bias. Finally, we use the B measure as the dependent variable in an explanatory model of different sources of polling bias, as an illustration of how the measure may be applied empirically.

The rest of this paper is rather math-heavy and does not display well as HTML. Go here for the PDF version of the preprint.

The final version of this paper should appear later this year in Political Analysis. We have created a Stata command surveybias that implements the methods described in the paper. It is due to appear on SSC this summer.

Ein Blick von außen. Anmerkungen zu Steinbrink et al. “Netzwerk(analys)e in der deutschen Humangeographie”


1 Einleitung: Netzwerkanalyse in der Geographie

Dass die Herausgeber uns die Möglichkeit geben, den Beitrag von Steinbrink et al. zu kommentieren, ist für uns ebenso schmeichelhaft wie überraschend. Der Anlass für diese Einladung liegt darin, dass wir 2009 eine ähnliche Analyse der Publikationsstrukturen in unserer eigenen Disziplin, der Politikwissenschaft, veröffentlicht haben (Arzheimer und Schoen, 2009). Anders als Steinbrink et al. haben wir uns dabei auf die Zitationsnetzwerke in Zeitschriften beschränkt, haben allerdings zusätzlich zur Politischen Vierteljahresschrift (PVS) als Flaggschiff der deutschen Politikwissenschaft deren Schwesterblatt Österreichische Zeitschrift für Politikwissenschaft (ÖZP) sowie die beiden wichtigsten politikwissenschaftlichen Zeitschriften aus Großbritannien (British Journal of Political Science/BJPS und Political Studies/PS) berücksichtigt. Auf Basis dieses Samples, das wesentlich mehr Autoren umfasst als der Datensatz von Steinbrink et al., kommen wir zu dem Schluss, dass die deutsche Politikwissenschaft deutlich stärker fragmentiert ist als die britische Politikwissenschaft oder auch die deutsche Humangeographie.

Ein Grund für den letztgenannten Befund liegt sicher in unserer Konzentration auf eine einzige deutsche Zeitschrift, die ihrem Selbstverständnis nach ein Forum für die ganze Breite des Fachs darstellt. Hinzu kommen weitere Erklärungen, auf die wir im folgenden eingehen wollen.

Als Sozialwissenschaftler sind uns die Fragestellungen der Humangeographie zwar nicht gänzlich fremd, die disziplinären Strukturen aber weitestgehend unbekannt. Insofern sind unsere Einschätzungen im gleichen Maße unbefangen wie naiv, und wir müssen in vielerlei Hinsicht um Nachsicht bitten.

2 “In den Professorenstand erhoben”

Im Gegensatz zu unserer eigenen und einer Reihe von vergleichbaren Analysen konzentrieren sich Steinbrink et al. auf eine wichtige Untergruppe innerhalb ihrer Disziplin, nämlich auf die hauptamtlichen Professoren an deutschen Universitäten. Diese Auswahl begründen sie einerseits mit forschungspraktischen Notwendigkeiten, andererseits mit der Steuerungsfunktion dieser Gruppe für die Disziplin.

Während der forschungspraktische Nutzen einer solchen Eingrenzung offensichtlich ist – anders ließe sich die aus unserer Sicht sehr wichtige Kombination von Publikations- und Konferenzdaten wohl kaum realisieren – glauben wir, dass die personelle Engführung das Bild der Humangeographie in systematischer Weise verzerrt, und zwar sowohl inhaltlich als auch strukturell.

Inhaltlich sind Professoren nicht notwendigerweise große Innovatoren. Auch wenn das Einstein zugeschriebene Diktum, “wer seinen großen Beitrag zur Wissenschaft nicht bis zum 30. Lebensjahr geleistet hat, wird dies nie mehr schaffen” inzwischen als widerlegt gilt (Jones und Weinberg, 2011), sind es doch häufiger die Doktoranden und Postdocs, die radikal neue Ideen formulieren oder zumindest aus anderen Disziplinen importieren. Einen – wie wir finden – eindrücklichen Beleg für die kreative und innovierende Rolle von Doktoranden und Postdoktoranden liefert die Analyse von Steinbrink et al. selbst, mit der offenbar die Netzwerkanalyse in die deutsche Humangeographie Einzug hält. Die Altvorderen hingegen scheinen doch eher dazu zu neigen, ihr Oeuvre zu konsolidieren. Professorale Beiträge bieten deshalb nicht unbedingt ein repräsentatives Spiegelbild der wissenschaftlichen Produktion einer Disziplin. Schon deshalb würden wir die Aussage von Steinbrink et al., dass die “Hochschullehrerinnen und Hochschullehrer formal die Hauptakteure im Wissensnetz” (S. 7) sind, mit einem Fragezeichen versehen oder geradezu als Lehrbuchbeispiel für die Einsicht verstanden wissen wollen, dass formale Rollen nicht notwendigerweise etwas über die tatsächliche Bedeutung von Akteuren aussagen.

Darüber hinaus hat die Konzentration auf die Professoren aber auch Konsequenzen für die (wahrgenommene) Struktur des Wissensnetzwerkes. Gerade weil Professoren in aller Regel dem Wissenschaftsbetrieb schon länger angehören, in der Organisation von Tagungen und anderen Publikationskanälen eine zentrale Rolle spielen, gemeinsam an größeren Projekten arbeiten und sich untereinander häufig kennen, ja häufig schon gemeinsam bei denselben akademischen Lehrern studiert haben, wäre es geradezu erschreckend, wenn dieser Personenkreis sich nicht wechselseitig zitieren und auch gemeinsam publizieren würde.

Damit stellt sich die Frage, ob der Befund der relativ hohen Integration in den Zitations- und Publikationsnetzwerken partiell ein Artefakt darstellt, das sich aus der Beschränkung auf eine Teilgruppe der Autoren erklärt. Ohne Kenntnis der Publikationspraxis in den sechs untersuchten Zeitschriften – vielleicht erscheinen dort tatsächlich primär Beiträge von Professoren – lässt sich dies nicht entscheiden. Für unsere Vermutung spricht aber die weitaus geringere Dichte der Verknüpfungen im “Netz der Geographentage”, wo mehrheitlich Doktoranden und Postdoktoranden referieren. Aus unserer Sicht wäre es lohnend, das gemeinsame Netzwerk von Professoren und anderen Wissenschaftlern in den Blick zu nehmen und dabei zu untersuchen, ob sich die von den Autoren als solche wahrgenommene Standesschranke auch empirisch nachweisen läßt.

3 “In Deutschland weltbekannt?”

Einer der großen Vorzüge der Studie von Steinbrink et al. liegt darin, dass die Autoren die deutsche bzw. deutschsprachige Zeitschriftenliteratur im Bereich der Humangeographie mit ihrer Analyse von sechs Zeitschriften vermutlich relativ vollständig erfassen. Auf diese Weise können sie ein faszinierendes Bild ihrer Subdisziplin zeichnen.

Allerdings bleibt auf diese Weise eine wichtige und zusehends wichtiger werdende Dimension ausgeblendet, nämlich die Frage nach der internationalen Vernetzung von Wissenschaftlern, Zeitschriften und Disziplinen. Unsere eigene Analyse stützt sich auf die Auswertung von zwei deutschsprachigen und zwei britischen Journals. Dabei zeigt sich, dass letztere einen höheren Grad der internen Verflechtung aufweisen als ihre deutschsprachigen Pendants. Dies dürfte sich zum Teil daraus erklären, dass Publikationen in deutscher Sprache im weltweiten Wissensnetz der Politikwissenschaft eine immer geringere Rolle spielen – auch und gerade für Forscherinnen und Forscher, die in Deutschland publizieren, aber sich dabei vornehmlich auf internationale Ergebnisse stützen.

Wir vermuten, dass sich die Situation in der deutschen Humangeographie ganz ähnlich darstellt. Aus unserer Sicht wäre es deshalb interessant zu wissen, in welchem Umfang die von Steinbrink et al. untersuchten Autoren international publizieren, international zitiert werden und selbst Literatur außerhalb des deutschen Netzwerkes zitieren. In dieser Perspektive könnte sich das Bild einer kleinen, in sich geschlossenen Gemeinschaft relativieren oder aber auch verdichten. Darüber hinaus könnte sich die relative Position einzelner Personen in dem Netzwerk verändern. In einer besonders drastischen, zumindest denkmöglichen, wenn auch nicht sehr wahrscheinlichen Ausprägung könnte dies seinen Niederschlag darin finden, dass national zentrale Personen im internationalen Maßstab am Rande stehen, während nationale Außenseiter in die internationale Community vergleichsweise gut eingebunden sind. In jedem Fall verspricht die internationale Einbettung der Befunde zur deutschen Humangeographie wichtige Zusatzinformationen, die Fehlinterpretationen vermeiden helfen.

4 Netzwerkanalyse – wozu?

Die Netzwerkanalyse als Methode erfreut sich in den letzten Jahren in den verschiedensten Disziplinen wachsender Beliebtheit. Eine Suche auf GoogleScholar ergibt für die Publikationsjahre 2009 und 2010 jeweils über 9100 Treffer. Für das Jahr 2008 sind es nur 6550, für 2011 werden trotz des typischen Nachlaufs von Literaturdatenbanken bereits jetzt über 10.550 Treffer verzeichnet. Die Gründe dafür liegen auf der Hand: günstige, leistungsfähige Software, Zugang zu Netzwerkdatensätzen und das wachsende Bewusstsein für die lange Zeit dem Vergessen anheimgefallene Einsicht, wie häufig Netzwerkphänomene in allen sozialen Bereichen sind – nicht zuletzt in den Wissenschaften, in denen sie beispielsweise zur wellenartigen Ausbreitung neuer Ideen und Methoden beitragen können. Hinzu kommt, dass Menschen ein geradezu naturwüchsiges Interesse an sozialen Netzwerken zu haben scheinen und insbesondere graphische Darstellungen, die soziale Beziehungen illustrieren, intuitiv eingängig sind, suggestiv wirken und daher eine große Anziehungskraft ausüben.

Trotzdem stellt sich die Frage nach dem Erkenntnisgewinn, der durch Netzwerkanalysen tatsächlich zu erzielen ist. Aus Arbeiten wie der von Steinbrink et al. oder unserem eigenen Beitrag, lernen wir zunächst, wer mit wem vernetzt ist, und wer die Stars in einem Wissensnetzwerk sind. Dies befriedigt zwar unsere Neugier (und mag in einigen Fällen unsere lebensweltlich begründeten Einschätzungen bestätigen), tut aber per se noch nichts zur Sache.

Interessanter ist die vergleichende Perspektive, d.h. etwa die Frage, ob das Wissensnetzwerk in der deutschen Humangeographie stärker zentralisiert oder fragmentiert als in den Nachbardisziplinen oder -ländern. Selbst vor einer solchen Vergleichsfolie stellen sich dann aber Folgefragen nach einer möglichen optimalen Struktur eines Wissenschaftsnetzwerkes. Spiegelt die Zentralität einer kleinen Gruppe von Akteuren deren anerkannte und verdiente Spitzenposition wider, oder ist sie vielmehr Ausdruck einer dysfunktionalen Kartellbildung? Sollte die Existenz von Subnetzwerken als „Balkanisierung“ beklagt oder im Sinne einer problemadäquaten Ausdifferenzierung begrüßt werden? Solche und ähnliche Fragen sind letztlich nur subjektiv und vor dem Hintergrund einer intimen Kenntnis der jeweiligen Fachdisziplin zu beantworten.

Der Vergleich sollte sich auch auf die zeitliche Dimension erstrecken. Eine solchermaßen verbreiterte Datenbasis würde es Forscherinnen und Forschern erleichtern, der allzu menschlichen Versuchung zu widerstehen, eine Momentaufnahme als zeitlos gültigen Befund fehlzuinterpretieren. Doch nicht nur das. Der zeitliche Vergleich könnte helfen, eine Reihe reizvoller substantieller Fragen zu klären. Werden die Muster inner- und interdisziplinärer Vernetzung über die Jahrzehnte erfolgreich reproduziert und damit von Forschergeneration zu Forschergeneration weitergegeben, oder lassen sich systematische Verschiebungen erkennen? Unterliegt die interdisziplinäre Vernetzung einem systematischen Wandel, lassen sich dabei bestimmte Individuen als Vorreiter identifizieren? Reagieren die Vertreter einer Disziplin auf strukturelle Veränderungen in der Umwelt, etwa den Bedeutungsgewinn der internationalen Wissenschaftsarena oder aber Änderungen rechtlicher Regeln, mit Anpassungen, in welcher Richtung und in welcher Geschwindigkeit, oder erweisen sich die überkommenen Muster wissenschaftlicher Interaktion als robust gegenüber solchen Veränderungen von Randbedingungen?

Die von Steinbrink et al. unternommene Vergleich von Vortrags- und Publikationsnetzwerken scheint uns hier ein Schritt in die richtige Richtung zu sein, der deutlich über unsere eigene Arbeit hinausgeht. In einer idealen Welt sollten darüber hinaus weitere Affiliationsnetzwerke (gemeinsame Doktorväter und -mütter, gemeinsame Studien- und frühere Arbeitsorte) sowie die Zusammensetzung von Berufungs-, Findungs- und Begutachtungskommissionen einbezogen werden. In einem solchen Wunschszenario würde zudem der Untersuchungszeitraum erheblich ausgedehnt, um individuelle und kollektive, lebenszyklische, generationale und periodenspezifische Dynamiken – etwa die Diffusion der Idee, in der Humangeographie Netzwerkanalysen einzusetzen, und die Kreativität ihrer Urheber im akademischen Lebenszyklus – studieren zu können. Ein solches Unternehmen stößt zwar an forschungspraktische und -ethische Grenzen, verspricht dank dem Vergleich von Netzwerken und der intertemporalen Perspektive aber nochmals ein erhebliches zusätzliches Erkenntnispotential.


Arzheimer, Kai und Harald Schoen (2009). “Isoliert oder gut vernetzt? Eine vergleichende Exploration der Publikationspraxis in der PVS”. In: Politische Vierteljahresschrift 50, S. 604–626.

Jones, Benjamin F. und Bruce A. Weinberg (2011). “Age dynamics in scientific creativity”. In: Proceedings of the National Academy of Sciences. DOI: 10.1073/pnas.1102895108.

Political Opportunity Structures and Right-Wing Extremist Party Success


West European right-wing extremist parties have received a great deal attention in the academic literature due to the success that many of these actors have experienced at the polls. What has received less coverage, however, is the fact that these parties have not enjoyed a consistent level of electoral support in this third wave of right-wing extremist party activity (Beyme, 1988). Instead, their electoral fortunes have risen and fallen over the last two decades. The fact that this question of variation in the electoral support for the parties of the extreme right – both over time and across countries – has attracted relatively little attention in the literature is not overly surprising. For one thing, there continues to be a shortage of comparative studies on the extreme right and in particular on the extreme right’s voters. In addition, as far as the studies that do exist are concerned, it is not surprising that many of these have tended to focus only on why right-wing extremist parties have been successful, rather than on why they have not.


The few works that have addressed the issue of the variation in the electoral support for the parties of the extreme right across Western Europe have tended to offer only partial explanations for this phenomenon. Jackman and Volpert (1996), for example, assess the importance of electoral system, party system and economic factors on the right-wing extremist party vote, but they do not consider the impact of different socio-demographic variables. Likewise, Abedi (2002) concentrates on the effect of party system factors but fails to examine the influence of socio-economic variables and of other institutional characteristics. Knigge (1998), by contrast, explores the effect of some socio-economic factors but does not examine the impact of electoral system or party system factors. Thus, while these studies each add to an overall explanation for the variation in the electoral fortunes of the parties of the extreme right, on their own, they offer an account for the phenomenon that is far from comprehensive.


A more extensive explanation for the uneven electoral success of the parties of the extreme right is to be found in the influential work by Kitschelt (1995) and in the useful study by Lubbers and his colleagues (2002). However, in spite of its comprehensive nature and of the significant contribution that it makes to research on right-wing extremism, the study by Kitschelt also has a number of limitations. In particular, the framework employed does not allow for a precise assessment of the relative influence of the different independent variables on each of the right-wing extremist parties under observation.


The study by Lubbers et al. certainly does not suffer from this limitation. Yet, it is problematic, too, in terms of its methodology, the countries that it covers and its time-span. The decision to combine data from national election studies with data sets from supra-national projects raises potential problems of validity and reliability. In addition, the use of multi-level analysis is open to question.i As for the countries examined, the inclusion of countries where support for the extreme right is extremely low is also not without consequences. Finally, in terms of the time-span covered, Lubbers and his colleagues analyze data from 1994 to 1997 only, and do not cover the early to mid-1980s in which many right-wing extremist parties of the third wave broke through into the electoral arena. Therefore, the variance in explanatory factors such as unemployment, immigration and the positions of other parties is probably severely restricted.


In light of the limitations of the existing studies, this paper seeks to put forward an explanation for the variation in the right-wing extremist party vote across Western Europe that incorporates a wider range of factors than have been previously considered and that covers a longer time period. More specifically, through the construction of an individual-level model, the paper first examines the impact of socio-demographic variables on the right-wing extremist party vote. Then, by augmenting the model with system-level information, the paper investigates the influence of a whole host of structural factors (which together make up the political opportunity structure) that may potentially affect the extreme right’s performance at the polls. This two-stage approach enables us to assess the extent to which system-level features (relating to the political opportunity structure) account for variation in the extreme right’s success over time and across countries after individual-level features have been controlled for. Moreover, it also allows us to establish whether individual-level characteristics still have an effect on right-wing extremist voting when the political opportunity structure is held constant. The paper concludes with an assessment of which variables have the most power in explaining the uneven electoral success of right-wing extremist parties across Western Europe.



Theoretical Framework


i) Socio-demographic Factors


It has been well documented that certain socio-demographic groups have shown themselves more likely to vote for the parties of the extreme right than others. In the first instance, a significant gender gap in the support for the extreme right has been reported, with male voters exhibiting a greater propensity to vote for right-wing extremist parties than their female counterparts (see for example Betz, 1994; Lubbers et al., 2002).


Similarly, the existing studies have shown that an age effect exists, with both younger and older voters being more likely to support the extreme right than other age groups. A number of theories help explain this U-shaped phenomenon. It has been well documented, for example, that the decline in the effects of social structure has not affected all generations equally, and that younger voters as well as pensioners are more likely to lack social ties. Greater social integration is likely to be reflected not only in higher levels of electoral participation but also in a tendency to refrain from voting for a party of the extreme right. A further explanation for the greater propensity of both young and older voters to support the extreme right rests in these people’s interests and their access to welfare. Since young and old voters depend disproportionately on welfare, these two age groups are more likely to view immigrants as competitors than are people of other age groups.


As regards formal education, it is often hypothesized that people with lower levels of education will exhibit a greater propensity to vote for parties of the extreme right than people with higher levels of education. In the first instance, there is an economic or an interest-based argument to support this presumption: voters with lower levels of education tend to be less skilled, and hence are likely to fall victim to market forces (Falter, 1994: 69). They tend to support parties of the extreme right because these parties pledge to defend the economic interests of these voters by limiting the rights of immigrants and asylum-seekers, who are perceived as direct competitors both in the workplace and in accessing social services and housing. Another argument is value-based. It rests on the premise that, through education, people are intensively exposed to liberal values, and hence the longer a person spends in education, the more likely they are to embrace such values (Warwick, 1998; Weakliem, 2002). A similar argument holds that cognitive style effects explain the link between a person’s propensity to vote for a party of the extreme right and their level of education (Weil, 1985).


Finally, as regards class, a number of national studies have shown shopkeepers, artisans and small-business people to be particularly well represented among the electorates of right-wing extremist parties in several countries. An over-representation of working-class voters among those who support the parties of the extreme right – in some cases right from the start, in other instances growing over the years – is also well-documented by many studies at the national level. Finally, it has also been argued that people in non-manual jobs who enjoy a small degree of autonomy in their work may also develop authoritarian preferences, quite similar to those ascribed to working-class voters (Kitschelt, 1994: 16-17).


To sum up then, based on the evidence that has emerged in much of the existing literature, we expect there to be a greater propensity to vote for parties of the extreme right among men, among voters who are either young or old, among those with lower levels of formal education, and among the working class, the self-employed and those in routine non-manual forms of employment as compared to all other socio-demographic categories of elector



ii) Political Opportunity Structures


To assess the influence of structural or environmental factors on the right-wing extremist vote we draw on the concept of political opportunity structures, which was originally developed in the context of research on social movements to denote the degree of ‘openness’ or ‘accessibility’ of a given political system for would-be political entrepreneurs. In a very influential study Kitschelt describes political opportunity structures as ‘specific configurations of resources, institutional arrangements and historical precedents for social mobilization, which facilitate the development of protest movements in some instances and constrain them in others’ (1986: 58). As their name implies, political opportunity structures therefore emphasize the exogenous conditions for party success and, in so doing, contrast to actor-centred theories of success (Tarrow, 1998: 18).


The concept of political opportunity structures is a broad one and different authors have included different items in their definition of the term. In spite of the differences, however, the majority of studies agree that fixed or permanent institutional features combine with more short-term, volatile or conjectural factors to produce an overall particular opportunity structure (e.g. Kriesi et al., 1995). We therefore propose to adopt a three-pronged approach with which to examine the influence of political opportunity structures on the right-wing extremist party vote: a first set of variables captures the impact of long-term institutional features on the parties of the extreme right; a second set examines medium-term factors which relate to the party system; and a third set of variables examines short-term contextual or conjectural variables.


a) Long-term Institutional Variables


Two institutional variables we regard as being of potential importance to how well parties of the extreme right perform at the polls are (i) the electoral system, and (ii) the degree of decentralization/federalism. As far as electoral systems are concerned, it has long been established that the more proportional the electoral system, the greater the incentives for political entrepreneurs to enter the electoral race and for voters to decide to support a new or a small political party. By contrast, the less proportional the electoral system, the more leaders of new or small parties will be dissuaded from fielding candidates and the more discouraged voters will be from voting for such parties since they stand little change of gaining representation (Duverger, 1951; Blais and Carty, 1991). In view of this relationship, we anticipate that unless they have already reached a certain size and have a chance of continuing to attract a sizable section of the electorate, right-wing extremist parties are likely to suffer from disproportional electoral systems.


The effect of decentralization or federalism is less clear-cut. On the one hand, it can be argued that a high degree of decentralization (including regional parliaments) may foster the development of right-wing extremist parties because voters are more willing to support new and/or radical parties in ‘second order’ elections (Reif and Schmitt, 1980). However, rather than allowing extremist parties to gain a toehold in the electoral arena, it may instead be the case that second order elections serve as a kind of security valve for the political system by providing citizens with an opportunity to express their political frustration with the mainstream parties without overly disturbing the political process on the national level. Therefore, two contrasting – yet equally convincing – hypotheses as to the effect of territorial decentralization exist.


b) Medium-term Party System Variables


Party system variables are less constant than institutional factors. For reasons of parsimony, we restrict ourselves to examining the impact of three such variables: (i) the ideological position of other competitors in the party system, (ii) the degree of convergence between the mainstream parties, and (iii) the coalition format in the respective party systems.ii


We expect the position of the major party of the mainstream right in each of the respective party systems to have an impact on the success of the party of the extreme right, yet it is difficult to predict the exact nature of this impact. On the one hand it can be argued that the more right wing the party of the mainstream right, the less political space will be available to the party of the extreme right. On the other hand, it can be argued that a more right wing party of the mainstream right might legitimize the issues around which the extreme right mobilizes. Thus, two competing hypotheses emerge as to the influence of the ideological position of the mainstream right on the electoral success of the extreme right.


Next, we examine the degree of convergence between the parties of the mainstream right and the parties of the mainstream left in each of the party systems under observation.iii Here too, two contrasting hypotheses present themselves. On the one hand we can argue that right-wing extremist political parties will benefit where the mainstream right and the mainstream left converge (Kitschelt, 1995: 17). In such instances the parties of the extreme right can credibly argue that if voters wish to see a real alternative to both the government and the mainstream opposition, then they should put their support behind the right-wing extremist party. When the mainstream parties are ideologically distinct from each other, it is more difficult for the parties of the extreme right to adopt this strategy. On the other hand, the extreme right might perform well at the polls when the mainstream parties are ideologically quite distinct. First, this distinctiveness may signal the lack of elite consensus (Zaller, 1992), which might further extreme right party success. Second, the mainstream parties may have diverged ideologically in an attempt to curb the advance of the parties of the extreme right in upcoming elections. Either way, ideological divergence between the mainstream parties may be associated with extreme right party success. Once again, therefore, two conflicting hypotheses exist as to the effect of ideological convergence of the mainstream parties on the right-wing extremist party vote.


We then move to consider the coalition format of the party systems under investigation. We suspect that the extreme right will benefit from grand coalitions because (i) voters will feel that there is a lack of other political alternatives during a grand coalition and (ii) supporters of the mainstream right may become alienated if they do not see their preferred policies being enacted and do not enjoy the consolation of seeing their party play the role of a principled opposition (Kitschelt, 1995: 17). Therefore, we anticipate that the right-wing extremist party vote will be higher in (or shortly after) periods of grand coalition government than it will be in periods of alternating government.


c) Short-term Contextual Variables


In addition to being affected by long-term institutional variables and medium-term party system variables, it is also reasonable to expect the right-wing extremist vote to be influenced by a number of short-term contextual factors. More specifically, given the considerable emphasis parties of the extreme right place on the issue of immigration from non-EU countries and on the supposed competition between immigrants and the indigenous population, we anticipate that levels of immigration and unemployment (both straightforward levels and also change in these levels) will exert an effect on how well the parties of the extreme right perform at the polls. We expect the right-wing extremist vote to be positively correlated to both the level of immigration and the level of unemployment.



Data and Methodology


The data in our analysis come from national election studies. The pooling and harmonizing was carried out under the auspices of the Extreme Right Electorates and Party Success (EREPS) Research Group.iv The major advantage of using national election studies is that they reflect voter behaviour at election time. This contrasts to supranational surveys, which may be carried out at a time close to the beginning of the electoral cycle in one country, but near the end of the cycle in another.

These national election studies provided us with information on the individual vote choices and the socio-demographic characteristics of West European electors. In contrast to some of existing studies of right-wing extremist electorates (e.g. van der Brug et al., 2000; Swyngedouw, 2001; Lubbers et al., 2002; van der Brug and Fennema, 2003), we do not include variables that capture the different attitudes of voters because there are very substantial problems in finding comparative indicators of attitudes in national election studies, both over time and across countries. Although there is clearly some trade-off to be had in deciding not to include attitudinal variables, we believe that the advantages of using national elections studies (rather than supranational surveys) outweigh any disadvantages that result from excluding attitudinal variables. Furthermore, in contrast to attitudinal data, socio-demographic data are relatively easily compared and are measured with much less error.


The countries included in our analysis are: Austria, Belgium,v Denmark, France, Germany Italy and This means that the parties included in our analysis are: the Freiheitliche Partei Österreichs (FPÖ), the Vlaams Blok (VB); the Fremskridtspartiet (FRPd) and the Dansk Folkeparti (DF); the Front National (FN); the Deutsche Volksunion (DVU), the Nationaldemokratische Partei Deutschlands (NPD) and the Republikaner (REP); the Movimento Sociale Italiano / Alleanza Nazionale (MSI / AN) until 1995;vii and the Fremskrittspartiet (FRPn).


In contrast to the study by Lubbers et al., we have excluded countries where support for the extreme right is extremely low. While we recognize that including countries in which there is no effective extreme right is certainly necessary in a macro-level explanation of the extreme right’s success (and failure to do so would result in selection bias), we believe that incorporating such countries in an analysis of individual voting decisions is problematic for three reasons: (i) voting for the reasonably established extreme right parties in countries like Belgium, France or even Germany is not comparable to voting for a tiny (and often fanatical) political sect, (ii) in countries like Portugal, Spain, Great Britain, and Ireland, extreme right voters are extremely rare, with their numbers in social surveys even lower than the electoral results suggest,viii and (iii) in countries where the extreme right is very weak, prospective extreme right voters are often prevented from supporting an extreme right party because candidates of these parties are only fielded in certain constituencies. which is not reflected in surveys, as such voters are coded either as non-voters or as supporters of another party. Therefore, the inclusion of survey data from countries where support for the extreme right is extremely low or non-existent therefore dilutes and distorts any analysis of individual voting decisions.


While the parties included in our analysis differ from each other in terms of their precise ideological profile, we nonetheless believe that they belong to the same party family, and that they can thus be treated as constituent members of a larger, single group. There has been much debate in the literature over the exact definition of right-wing extremism, and hence over which parties belong to the extreme right party family, but a consensus has nonetheless emerged within this body of work that a separate extreme right party family does indeed exist. While it is perhaps more heterogeneous than other party families, its constituent parts are distinct from the parties of the mainstream right, and they also share a number of ideological features (in particular some combination of racism, xenophobia, nationalism, and a desire for a strong state and law and order), which allow them to be grouped together at the far right end of the left-right political spectrum (see Ignazi, 1992; 2003; Hainsworth, 1992, 2000; Betz, 1994; Mudde, 1996, 2000 among others for further details of this debate). Further evidence of the fact that the parties included in our study belong to a common extreme right party family can be found in the series of expert judgments studies that have been carried out since the beginning of the 1980s (Castles and Mair, 1984; Laver and Hunt, 1992; Huber and Inglehart, 1995; Lubbers, 2000).


Our timeframe spans the years 1984-2001. Our start date is informed by the broad consensus in the literature on right-wing extremism that the 1980s saw the beginning of a third wave of right-wing extremist activity in Western Europe (Beyme, 1988). The majority of scholars of right-wing extremism also agree that the Scandinavian Progress Parties only became part of the right-wing extremist party family in the mid-1980s when refugee and immigration policies became their primary concerns (Kitschelt, 1995: 121; Goul Andersen and Bjørklund, 2000: 203-204; Hainsworth, 2000). We therefore began with the Danish election survey of 1984, and collected all available data for polities where the extreme right was a relevant player in national parliamentary elections.


The socio-demographic variables included in our model are the standard ones: gender, age (up to 24 years, 25-34 years, 35-44 years, 45-54 years, 55-64 years, 65 years and older), formal education (no education/primary education, mid-school, secondary education, university degree), and social class (measured by a simplified Goldthorpe classification: professionals / managers, routine non-manual, self-employed, manual).


For our examination of the influence of political opportunity structures on the right-wing extremist party vote, we augmented the socio-economic data derived from the national election studies with information on the political systems and the party systems of the countries under investigation. To assess the impact of institutional variables we made use of data (derived from Carter, 2002) that measured the disproportionality of the electoral systems according to the Gallagher index (Gallagher, 1991), and we adopted Lijphart’s index of federalism to reflect the degree of territorial decentralization (Lijphart, 1999). This ranges from 1 to 5, with 1 indicating a unitary and centralized state and 5 referring to a federal and decentralized state.


To explore the influence of the position of other political competitors, and to assess the impact of mainstream party convergence and divergence we drew on the data of the Comparative Manifesto Project (CMP) (Budge et al., 2001). From the CMP data we constructed a measure based on the parties’ policies on the issues of multiculturalism, internationalism, the ‘national way of life’, and law and order. While reflecting many of the components that make up the overarching left-right dimension, these policy items are particularly important to the parties of the extreme right as it is primarily along these dimensions that they compete with their mainstream rivals.ix Like all measures that are based on CMP data, it reflects the balance between ‘left’ and ‘right’ statements of a party. Negative figures indicate a leaning to the left, with an empirical minimum of -12.4 for the Norwegian Socialists in the 1990s, while positive numbers indicate a leaning to the right of this dimension. Here, the empirical maximum for a party that is considered a part of the moderate right is 20.4, achieved by the Danish KF during the 1990s. However, the major right parties usually register much lower scores like e.g. 3.6 for the Austrian ÖVP in 1994 or 3.3 for the French RPR in 1997.


To examine the effect of a grand coalition in the period directly before a general election, we drew on data from EJPR data Yearbooks and included an appropriate dummy variable in our model.


Finally, to evaluate the effect of conjectural factors on the decision to vote for the extreme right, we drew on unemployment data at the aggregate levelx, and on data reflecting the number of asylum seekers in the countries under observation.xi We included a measure of the yearly number of asylum-seekers per thousand inhabitants,xii and a measure of the yearly percentage of unemployed people in the total workforce. We also included change rates for both variables in our model because, according to the classical ‘J-curve’ reasoning (see Davies, 1974; Coenders and Scheepers, 1998), people might respond to changes rather than to the actual level of both measures.


In terms of methodology, we estimate a logit model with contextual variables. Our model thus allows us to estimate the probability of a voter voting for a party of the extreme right conditional on (i) his/her individual socio-demographic attributes, and (ii) the particular political opportunity structures present in his/her country at the time of the election. Since there is no strong theoretical argument as to why socio-demographic or system-level explanations for an extreme right vote should vary systematically over countries and across time,xiii we assume that the true regression coefficients are constant across countries and across time after controlling for both individual and contextual variables. Therefore we refrain from inserting dummies and interactions to capture cross-country differences in intercepts and slopes.





Looking at Table 1, we can see that our findings are in line with much of the previous research in the field.xiv The results show that being male substantially raises the odds of voting for the extreme right. Put differently, depending on the respondent’s other attributes, being male increases the probability of an individual being an extreme right voter by more than 50 percent. This coefficient suggests that there is a substantial gender-gap in the support for the extreme right voting in Western Europe even when we control for other socio-demographic variables such as age, education, and social class.


Turning to the influence of age, Table 1 illustrates the U-shaped effect of this variable that we expected to see. Wald tests show that there are no significant (p = 0.45) differences in the respective levels of extreme right support among those voters who are between the ages of 35 and 64, while the level of support for parties of the extreme right among both younger and older voters is higher.xv The propensity to vote for a party of the extreme right among voters who are aged between 25 and 34 is identical (p = 0.91) to that of the reference group (voters who are 65 or older), while voters who are younger than 25 years old are much more likely to vote for the extreme right than any other voters, including the reference group.



Table 1: Socio-demographic model



Independent Variables








Age: -24





Age: 25-34





Age: 35-44





Age: 45-54





Age: 55-64





No/Primary Education










Secondary School








Routine Non-Manual









50 276

Adj. Pseudo-R2 (Mc-Fadden)



-515 293



Robust standard errors adjusted for clustering on country are shown in parentheses (see note Error: Reference source not found).

* significant at 5%; ** significant at 1%



As regards levels of formal education, we predicted that people with lower levels of education would exhibit a greater propensity to vote for parties of the extreme right than people with higher levels of education. When we examine our model, however, things are not as clear-cut. While the low level of support that extreme right parties receive from university-educated voters (the reference group) is in line with the predications advanced above, the coefficient for the group of voters with no education or with primary education is smaller than expected and is not significantly different from zero. We find, instead, that it is people with ‘mid-school’ diplomas who appear to form the core social base of the extreme right. Depending on his or her other characteristics, having a mid-school education more than doubles the probability of an individual voting for the extreme right. The effect of being educated to secondary level is somewhat weaker, but the difference between the two coefficients is not significant (p=0.19).


As concerns the effect of class, our findings are generally in line with our expectations. The results show that professionals and unclassified voters (the reference group) exhibit the lowest propensity to support extreme right-wing parties while the odds of an extreme right vote are somewhat higher if the respondent has a routine non-manual job, if he or she is self-employed, or if he or she is a manual worker.xvi


In a bid to summarize our socio-demographic findings we calculated the expected probability of an extreme right vote across varying levels of the independent variables (see Table 2). For the sake of brevity, we restricted class to unclassified voters (the reference group) in the upper section of the table, and to workers (the group with the highest propensity to vote for a party of the extreme right) in the lower section of the table. Above all, Table 2 shows the significant variation in the support for the extreme right that exists across the different socio-demographic groups. If, for example, we compare the predicted probability of a vote for the extreme right being cast by a female voter, aged 24 or less, with a university education and whose class is ‘unclassifiable’ with the predicted probability of an extreme right vote being cast by a male voter from the same age group, with a mid-school education and a manual job, we can see the full extent of this variation. Indeed, the figures in Table 2 illustrate that the predicated probability of the female voter just described voting for a party of the extreme right is roughly 5 percent (as shown in bold in the upper section of the table), whereas the predicted probability of the male voter just described voting for the extreme right is roughly 21 percent (as shown in bold in the lower section of the table). This example clearly illustrates that gender and education in particular have a sizeable impact on the probability of a person voting for a party of the extreme right, while age and class are somewhat weaker predictors.



Table 2: Predicted probabilities (in percent) of an extreme right vote, depending on gender, age, education, and social class.


class: unclassified


































































class: manual




































































Typical 95%-confidence intervals based on robust standard errors adjusted for clustering on country: female, less than 25 years old, university educated, class ‘unclassified’: 2.9 – 8.2;

male, less than 25 years old, mid-school education, manual worker: 13.2 – 32.6.



So far, therefore, our discussion has illustrated that a voter’s socio-demographic attributes go a long way in helping to explain his or her propensity to vote for a party of the extreme right at election time. In addition to this, our results have by and large also been in line with those of many of the existing studies on right-wing extremism. In particular, our comparative study of 24 elections in 7 countries confirms that parties of the extreme right are strongest among the more marginalized sections of society, and that (when we control for other socio-demographic variables) their support is predominantly male.


This agreement with the existing studies notwithstanding, our results point to another important finding: the low adjusted (McFadden) pseudo R2 in our model (a mere 0.03) indicates that the variation in the electoral success of right-wing extremist parties both over time and across space cannot simply be explained by the different composition of the respective electorates. Instead, the variation in the electoral fortunes of the parties of the extreme right must be explained by factors other than socio-demographic ones.


To confirm this we added a series of dummies for the 24 elections under study in our model (not shown) so as to create a model that captured all variation in the extreme right vote that could potentially be due to system-level factors. The resulting R2 of 0.09 was substantially higher than the R2 of the model in Table 1, thereby indicating that the extreme right’s electoral success varies considerably over time and across space even if we control for the composition of the electorate. In light of this, we now augment our socio-demographic model shown in Table 1 with variables that relate to the political opportunity structure as discussed above.


Table 3 shows the results of the full model. Looking at the table, the first observation to make is that the coefficients for the socio-demographic variables have not greatly changed since we have augmented the model with the political opportunity structure variables.xvii Second, we see that some of the additional variables have statistically significant and sizeable effects on an individual’s propensity to vote for a party of the extreme right. Finally, we see a significant improvement in the model-fit: the pseudo R2 more than doubles and, more importantly, the BIC is reduced by 1106, meaning that the full model is clearly superior to the socio-demographic one.xviii Given the nature of our explanatory variables, it is also worth noting that multicollinearity is not an issue in our model.xix


Starting with the two long-term institutional variables, we can see that the coefficient for the disproportionality of the electoral system is in fact positive, rather than negative as was anticipated.xx That is, the odds of voting for the extreme right actually increase with the disproportionality of the electoral system. At first we considered that this unexpected result might be caused by the inclusion of the French case, where the unique double-ballot system (whose disproportionality scores are extremely high) obviously did not prevent the ascent of the extreme right.xxi We therefore temporarily excluded France from the analysis but found that the coefficient for the disproportionality score hardly changed.


The absence of a negative relationship between the disproportionality of the electoral system and the right-wing extremist vote has been reported elsewhere (Carter, 2002), and two potential explanations for it have been put forward: (i) right-wing extremist party voters may simply not be aware of the consequences of electoral systems or (ii) their awareness may be overshadowed by other, more pressing concerns so that the psychological effects of electoral systems have only a weak impact on them. This latter hypothesis has clearly yet to be investigated.


As concerns the degree of decentralization and federalism, the coefficient is negative. However, since the coefficient fails the significance test we must accept that our data simply do not provide conclusive evidence as to which of the two hypotheses advanced above holds true in practice.



Table 3: Complete model



Independent Variables








Age: -24





Age: 25-34





Age: 35-44





Age: 45-54





Age: 55-64





No/Primary Education





Mid-School Education





Secondary School Education










Routine Non-Manual















Index of Decentralisation


Ideo. position of major party of mainstream right


Distance between major parties of mainstream left/right


Grand Coalition


Asylum Seekers per 1000 inhabitants


Asylum Seekers: Change


Unemployment Rate (%)


Unemployment Rate: Change






50 276

Pseudo-R2 (Mc-Fadden)



-516 399



Robust standard errors adjusted for clustering on country are shown in parentheses (see note Error: Reference source not found).

* significant at 5%; ** significant at 1%



Turning to the medium-term party system variables we can see that the position of the major party of mainstream right has a positive and borderline-significant (p = 0.05) effect on the right-wing extremist party vote. A move to the right by the major party of the mainstream right raises the odds of an extreme right vote. This suggests that the second hypothesis advanced above (that a mainstream right party may legitimize the policies of the extreme right by adopting some of their positions) has some validity.


The findings also show a positive effect of the distance between the mainstream parties on the right-wing extremist party vote, which is in line with the second hypothesis put forward above. However, since the coefficient does not pass the conventional threshold of significance (p = 0.08), though they are suggestive, our data do not provide conclusive evidence as to which of the competing hypotheses is borne out in practice.


The final medium-term party system variable that we included in our model was one that referred to the coalition format of the party systems under investigation. Our findings in Table 3 show that the existence of a grand coalition government before the election in question does indeed have a substantial effect. As we anticipated, the presence of such a governing coalition raises the odds of voting for the extreme right. Depending on the level of the other variables, the probability of an extreme right vote is roughly doubled.


As concerns the variables that related to short-term contextual factors, table 3 shows that the effect on the extreme right vote of the number of asylum-seekers is in line with the expectations (it is positive), while the coefficient for the change in the number of asylum seekers is negative. However, both these variables miss the usual threshold for statistical significance by a considerable margin. Therefore, we must assume that their true effect is zero.


The effect of unemployment (as a macro variable) on extreme right voting is markedly negative – that is, the odds of voting for the extreme right fall as the rate of unemployment increases. While this clearly does not allow us to draw any conclusions about the extreme right’s appeal to unemployed people (since this would be an instance of ecological fallacy),xxii we can surmise that extreme right parties perform better at the polls in societies where unemployment is low.


Although similar results have been reported in other studies (e.g. Knigge, 1998; Coenders and Scheepers 1998, Lubbers et al., 2002), a substantial explanation for this finding is not readily given. One plausible (yet untested) reason for this negative relationship is that people may turn to the more established and experienced mainstream parties in times of economic uncertainty rather than to the parties of the extreme right that lack such experience (Knigge, 1998: 269-270). The coefficient for the change in the unemployment is positive but is not statistically significant, thus again implying that the true impact of this variable on the likelihood of a vote for the extreme right is zero.


In the same way that we summarized the findings of our socio-demographic model in Table 2, Tables 4a and 4b summarize the findings of our complete model and show the combined impact of the four strongest system-level predictors on two segments of the population. Table 4a depicts the expected probability of an extreme right vote of a group that is least likely to support parties the extreme right (female voters, aged 45-54, with university education, and from the ‘unclassified’ class category); and Table 4b shows estimates for a small, marginal segment of the general population among which the extreme right is usually quite successful (male manual worker, aged 24 or younger, with no or primary education only).


Tables 4a and 4b show the expected probability of an extreme right vote from these two types of voters in situations where:

  1. there is a grand coalition in place in the preceding period of government and when there is not,
  2. the disproportionality of the electoral system is 1 (low) and where it is 5 (high),
  3. the ideological position of the major party of the mainstream right is –5, –1, 1 and 3 (with -5 indicating a rather left-wing position and 3 indicating a more right-wing position), and
  4. the unemployment rate is 2 percent, 4 percent, 6 percent, 8 percent, 10 percent and 12 percent.


First, we note that the socio-demographic variables have a considerable and consistent impact even if we control for system-level variables. If we compare equivalent cells from Table 4a and Table 4b, it is obvious that independent of the socio-political context, the probability of an extreme right vote is about five to six times higher for the young male, primary-educated worker than for the mid-aged, unclassified, university educated female voter.

Table 4a: Predicted probabilities (in percent) of an extreme right vote, depending on various system-level variables. Female voters aged 45-54, with university education, and from the ‘unclassified’ class category.



Female, class unclassified, university education, aged 45-54

Grand Coalition: No

Disproportionality: 1

Disproportionality: 5

Ideo Pos of MRUnempl Rate































































Grand Coalition: Yes

Disproportionality: 1

Disproportionality: 5