Protest or mainstream? How the European anti-immigrant parties have
developed into two separate groups by 1999.1
Wouter van der Brug*
Meindert Fennema *
Accepted for publication in the European Journal for Political Research: to be published in the
spring issue of 2003
1
We like to thank Susan Olzak and Jean Tillie for stimulating discussions and critical remarks on this paper.
*
Wouter van der Brug is assistant professor of methodology at the Department of Communication Science at the
Universiteit van Amsterdam. His research is funded by the Amsterdam School of Communications Research
(ASCoR).
*
Meindert Fennema is professor of political theory at the Universiteit van Amsterdam and program leader at the
Institute for Migration and Ethnic Studies. During the year 2000-2001 he was fellow of the Netherlands Institute
for Advanced Study in the Humanities and Social Sciences (NIAS). He wants to thank the NIAS staff for its
hospitality and in particular Anne Simpson for correcting the English.
Protest or mainstream? How the European anti-immigrant parties have
developed into two separate groups by 1999
Abstract
In their 1994 study on electoral preferences for anti-immigrant parties Van der Brug,
Fennema & Tillie (2000) concluded that most voters consider anti-immigrant parties as
normal parties. In this 1999 replica the main conclusions of the 1994 study are confirmed for
FPÖ, Alleanza Nazionale, Dansk Folkeparti and Vlaams Blok. Preferences for these parties
are largely determined by the same variables that determine preferences for other parties.
Votes for these parties should not be interpreted as protest votes. Things turned out to be
different for the Lega Nord, the Centrumdemocraten, the Fremskridtspartiet, the French Front
national, the Republikaner and the Wallonian Front National. The motivations to vote for any
of the latter parties turn out to be substantially different in 1999 than they were in 1994. In
1999 these parties attract more protest voters, or rather, they lost their ideological and
pragmatic voters. The research note finishes with a discussion of the implications of these
findings for our theoretical understanding of the electoral support for these parties, as well as
for the ongoing political debate about the appropriate strategy to fight them.
Introduction
On the basis of survey data from 1994 Van der Brug, Fennema and Tillie (2000) concluded
that votes for 7 European anti-immigrant parties could be largely explained by the same
causal mechanism that explains votes for other parties. Ideological agreement with these
parties, intolerant attitudes towards immigrants and pragmatic considerations were the
strongest predictors of voting for these parties.1 So, they concluded that votes for these parties
should be interpreted as motivated by the same pragmatic and ideological considerations that
apply to votes for other parties; these votes should not be considered as protest votes. A new
data set collected around the elections for the European Parliament in 1999 enables us to
replicate the analyses of 1994. There are three good reasons for doing this.
1
See for the theoretical argument for labeling this party family anti-immigrant: Fennema (1997). The present
analysis will add empirical evidence that supports this choice.
1
First, since 1994 anti-immigrant parties have in various European countries been very
successful in elections or are doing well in opinion polls (Italy, Belgium, Austria and
Denmark); in other countries they have done very badly (France, Germany and The
Netherlands). It is important to understand the voters’ preference for these parties and it could
well be that the causes of voters’ preferences have changed between 1994 and 1999. This
research note will demonstrate that this is indeed the case. Secondly, the 1999 data set
contains information about three successful parties of the radical right that were not included
in the 1994 study: the Alleanza Nazionale, the Dansk Folkeparti and the Austrian FPÖ.
Thirdly, in response to the recent electoral success of the FPÖ in Austria – propelling it into a
government coalition with the Catholic People’s Party – various prominent observers (such as
Simon Wiesenthal) have argued in the public debate that votes for the FPÖ are protest votes,
that is votes without a policy content. According to these observers many people cast their
votes for the FPÖ not because they agreed with the political program of this party, but to
express discontent with the political clientelism of the SPÖ and ÖVP. The theoretical
framework and research design proposed by Van der Brug, Fennema and Tillie (2000) allow
us to test this proposition.
In the next section we will briefly introduce the theoretical framework. This will be
followed by a section which discusses the data that will be used here and also the research
design. Analyses will then be presented pertaining to the electoral support for 10 parties
generally considered to belong to the radical right: the Fremskridtspartiet, the Dansk
Folkeparti (both Danish), Alleanza Nationale, Lega Nord (both Italian), Vlaams Blok and
Front National (both Belgian), Republikaner (German), FPÖ (Austrian), Centrumdemocraten
(Dutch) and Front national (French).
Theoretical framework
Various political analysts as well as academic researchers have argued that votes for antiimmigrant parties should be interpreted as protest votes (e.g., Van Holsteyn 1990;
Stouthuysen 1993; Betz 1994; Mayer & Moreau 1995; Derks & Deschouwer 1998). Little
conceptual clarity exists, however, in the use of the term protest vote. Derks & Deschouwer
(1998) combine abstainers and voters for a number of parties they consider protest parties,
and label them protest votes. They may have a good theoretical argument to consider
abstaining from the polls as protest behavior since voting in Belgium is obligatory and
abstention thus can be considered as a form of civic disobedience. However, they provide no
2
empirical evidence for calling non-voters protest voters, while leaving the contradiction in
terms unexplained. Furthermore, on the basis of assigned party characteristics they draw
inferences about the motives of voters that support these parties.
A similar kind of reasoning can be traced in the work of Betz (1994, 1998), where he
introduces the concept of populism. Betz speaks of the politics of resentment, which he calls
"the most distinct feature of the new populism of the 1980s and 1990s” (Betz, 1998:4). He
argues that the electoral success of the radical populist right “reflects to a large extent the
psychological strain associated with uncertainties produced by large-scale socio-economic
and socio-structural change.” Hence, the success of the radical right is “the result of a
profound and largely psychological crisis of the ‘popular classes’” (Betz, 1998a: 8) According
to this explanation party preference is primarily explained by social cleavages and by
rebellion against ‘the Powers That Be’. Although in the volume he has edited with Immerfall
Betz suggests that there is ample support for this contention, empirical proof of his theory is
virtually absent (Betz & Immerfall 1998). When discussing “neo-populist parties and
movements” Betz explicitly refers to the strategy of the radical right wing parties and says
very little about the motives of the voters. Instead, we feel that the notion of a protest vote is
only meaningful if it is defined in terms of the motives of voters and not in terms of party
characteristics.
Empirical studies that inquire directly into the motivations behind electoral
choices are rare, but these also lend little support to the protest vote hypothesis. Only in the
case of the Dutch Centrumdemocraten and the Lega Nord is there evidence that their voters
are motivated by feelings of alienation from the political system and thus can be considered as
protest voters (Betz, 1998b; Mudde and Holsteyn, 2000; Van der Brug, Fennema and Tillie,
2000). However, a much more sophisticated and elaborate analysis of the Vlaams Blok
voters’ motives shows that the attitude towards immigrants rather that a general feeling of
political powerlessness is a good predictor of Vlaams Blok voting. Hence Billiet and De Witte
(1995:194) conclude “our results … reject the ‘protest’ hypothesis”.
Some of the evidence in favor of the protest hypothesis stems from open-ended survey questions
asking voters why they voted for a certain party. Quite a large number of voters for anti-immigrant
parties mentioned discontent with the elite or other parties in exit polls (e.g., Rose 2000;
Swyngedouw, 2001). However, even though these voters may themselves mention, "protest" as their
prime motive, this does not mean that these voters are protest voters. In line with Kaplan (1964:32),
we feel that when explaining a social phenomenon we have to make a distinction between "act
meaning" (what an act means to an actor) and "action meaning" (how the researcher interprets this
3
act). Even though act meaning is important, it does not yield explanations that can be tested
empirically. Voters who agree with the political program of an anti-immigrant party may consequently
dislike other parties. When the rhetoric of anti-immigrant parties is directed against "the system",
voters for an anti-immigrant party may mention "protest" as their prime motive. Even though this
answer is meaningful, it could still be the case that these voters are motivated by ideological or policy
based orientations. Our definition of a protest vote is a vote that is of a qualitatively different nature
than policy votes: a vote primarily cast to scare the elite and a vote that is not policy driven. So, by
our definition, protest voting only occurs when political attitudes are of minor importance (see also:
Kitschelt, 1995: 276). It is possible that those who agree with the policies of anti-immigrant
parties would also like to scare the political elite. However, this is not how the concept of a
"protest vote" is normally conceptualised. Most authors use the term "protest votes" to
describe electoral decisions that are of a qualitatively different nature than "policy votes". As
a case in point, Lubbers & Scheepers (2000:69) introduce protest votes as follows "political
attitudes … are expected to be of minor importance" (see also, Mayer & Perrineau, 1992; Van
Holsteyn 1990). Our approach is in line with this conceptualisation.
Much of the empirical evidence for the protest hypothesis stems from analyses of
social characteristics of voters for anti-immigrant parties. For two reasons the results of such
studies have yielded little empirical evidence for the notion that the prime motive behind
votes for anti-immigrant parties is to show discontent with the political elite rather than to
espouse a particular set of political views. First, if a relationship exists between social
characteristics of voters and their electoral choices, this does not allow valid conclusions
about the motives of voters. This may be one of the reasons why the protest vote hypotheses
proposed by Lubbers & Scheepers (2000:70) and Lubbers, Scheepers & Billiet (2000:369) are
the same as the hypotheses these authors propose to test the "economic interests" theory. This
conceptual confusion just underlines the fact that the notion of protest voting is theoretically
undeveloped in electoral research.
Secondly, the relationship between social characteristics and preferences for antiimmigrant parties turns out to be very weak. Mayer (1998: 18) admits that “education and
social position have been rather poor predictors of the FN vote”. Election studies on the
Austrian FPÖ also lend no support to the cleavage model of Betz (Riedlsperger, 1998:35).
The Scandinavian radical right parties do not fit Betz’ profile either. Both Danish and
Norwegian Progress parties “have always recruited across class boundaries” while the
electoral support of the Swedish Ny Democrati was in 1991 “fairly representative of the
4
electorate in general” (Svasand, 1998: 86). Those Italian voters who vote for the Lega Nord
belong, according to Diamanti, to the “class of small entrepreneurs” (cited in Betz, 1998b:
52); it is very doubtful whether these can be called ‘losers of modernity’. Only in the case of
the Dutch Centrumdemocraten, the German Republikaner and the Vlaams Blok does the
electorate seem to conform to the profile that is predicted by Betz. (Lucardie, 1998: 120;
Mudde and Holsteyn, 2000: 157 ff; Winkler and Schuman, 1998: 104; Swyngedouw,
1998:71). Lubbers (2001) has so far done the most extensive study on socio-economic
variables that determine extreme right voting behaviour. He found a relationship between
level of education and voting for extreme right parties. But no such relationship for
unemployment seems to exist.
It is thus far from clear that the voters for radical right wing parties are, as a general rule,
recruited from particular segments of the electorates. When testing the protest vote hypothesis
we will also investigate in how far the votes for anti-immigrant parties can be explained by
the structure of social and economic cleavages in the countries under investigation.
In short, even though the protest hypothesis has had an irresistible attraction for many
students of anti-immigrant parties, little evidence exists in support of the hypothesis. This is
partially due to lack of conceptual clarity. We feel that we have good reason to investigate the
motives of the voters who are attracted by these anti-immigrant parties. Our analysis of the
protest vote thesis builds upon a typology of different motivations behind electoral decisions
proposed by Van der Brug, Fennema & Tillie (2000), who argued that the motives to cast a
protest vote can be analysed fruitfully with the conceptual framework of the rational choice
model (see also: Van der Eijk, Franklin & Marsh 1996). Based on different motivations, four
types of voting decisions are distinguished: pragmatic votes, idealistic votes, protest votes and
clientelist votes. SHOULD CLIENTELISTIC VOTING BE EXPLAINED? The latter voters
always vote for the largest party irrespective of its ideological position. Might is always right.
In this paper we do not investigate the clientelistic voter.
If voters vote for a party because they agree most with this party on the issues they
consider important, we call their vote an idealistic vote. Idealistic voting implies that
ideological proximity and policy considerations are strong determinants of party choice.
Those who want to influence public policies with their vote, may, however, not necessarily
vote for the party with whom they most agree. Some may decide to vote for a lesser preferred,
but larger, party if this party has a better chance of influencing public policies. So, if one's
goal is to affect public policies it is perfectly rational to vote not only on the basis of
5
ideological agreement (or proximity in spatial terms), but also to take each party’s relative
level of power into account. We call this pragmatic voting. Given the way parliamentary
democracies function, party size is a fairly good proxy for parliamentary power. Pragmatic
voting can thus be studied by estimating to what extent voters take into account the relative
size of a party when deciding which party to vote for (see also: Tillie, 1995:118).
The prime motive behind a protest vote is to show discontent with "the" political elite
by voting for a party that is an outcast in the political arena. We thus treat protest voters as
rational voters whose objectives are to demonstrate rejection of all other parties. We have
evidence that the political elites in all West European countries consider anti-immigrant
parties as dangerous political outcasts (see Fennema and Maussen, 2000; Fennema, 2000).2 If,
therefore, someone wants to use his/her vote to scare the elite - i.e., to cast a protest vote they will consider the party that is stigmatised by the political elite an attractive option. Since
the prime motive of a protest voter is to scare the elite, and not to affect public policies,
ideological proximity and issue-stands will be poor predictors of votes for anti-immigrant
parties, if these votes are protest votes.
Moreover, if someone wants to vote for a party as a sign of protest, a larger party will
not necessarily be more attractive than a smaller one. If the protest vote hypothesis is valid,
we therefore expect to find a weak effect of policy positions and party size on preferences for
anti-immigrant parties. If, however, the same factors determine the preference for antiimmigrant parties that determine preference for mainstream parties, such opinions of the
political elite (often expressed in the joint strategy of a cordon sanitaire) apparently do not
carry much weight.
Data and method
In order to assess whether pragmatic and policy considerations exert a strong or a weak effect
on the electoral attractiveness of anti-immigrant parties, we must compare the motivations for
voting for anti-immigrant parties with motivations to vote for other parties. Data from the
European Elections Studies provide an excellent opportunity to make this comparison,
because the data sets contain comparable information about a large number of parties from all
2
In the case of the Alleanza Nationale it may be doubted a priori whether we should expect citizens to cast a
protest vote for them, since this party is no longer stigmatised by the political elite. Moreover, since 1993-1994 it
no longer made a racist appeal to voters, so that the label anti-immigrant party may not be entirely appropriate
(see also: Griffin 1996; Eatwell 1998).
6
sorts of ideological denominations. For this study we will use data from the European
Election Studies 1999, which was conducted immediately following elections to the European
Parliament of 1999 (Van der Eijk, Franklin and Schmitt, 1999). It consists of cross-sectional
surveys using random samples from the electorates of each of the member states of the
European Union.3 In this study we use the surveys from only seven countries with one or
more parties of the radical right. Approximately 1,000 respondents were interviewed in
Denmark, France, the Netherlands and Germany. The Belgian and Austrian samples consist of
500 respondents each, while the Italian sample includes 3708 individuals. Our subsample thus
consists of 8731 respondents. This difference in subsample size affects the reliability of the
Belgian results in particular because, as we will see below, the Belgian respondents are
divided between two regions, Flanders and Wallonia, which will be considered as two
separate political systems.
The dependent variable of this study is the electoral utility of parties for each
respondent, which is measured by a set of questions, one for each party, that ask how likely it
is that the respondent 'would ever vote' for the party concerned. Tillie (1995) tested the
properties of this survey item thoroughly and concludes that valid inferences can be drawn
about the determinants of party choice by analysing the determinants of party preference. The
correlation between party preference and party choice turned out to be 0.90 in the European
elections of 1989. There are three methodological reasons for using the `probability of future
vote’ question as a dependent variable instead of party choice.
The first reason is that the `probability of future vote’ items allow for a research design
that is truly comparative (see below). Were we to use party choice as our dependent variable,
we would have to conduct separate analyses for each of the countries. Now we can analyse
party preference in one single analysis in which all parties from all countries are included.
Secondly, some of the radical-right wing parties that we are interested in attract so few votes
that estimates of the effects of different variables on decisions to vote for any of these parties
3 This study consists of independent cross-sectional surveys fielded in each member country of the EC/EU
immediately after the EP elections. It was based on a stand-alone survey conducted by telephone interviewing
and was funded largely by the Dutch NWO, the Spanish National Science Foundation, the Universities of
Mannheim and Amsterdam, and Trinity College, Hartford. The data will be deposited in 2001.The European
Elections Studies are extensively documented on the European Elections Studies Web Site
(http://www2.trincoll.edu/~mfrankli/EES.html). This site not only describes the 1999 European Elections Study,
but also those from previous years. Moreover, it contains an extensive list of publications emanating from these
studies.
7
are highly unreliable. Since the `probability of future vote’ items are asked of all respondents,
the parameter estimates are more robust. Finally, the `probability of future vote’ items tap
relevant information about the electoral attractiveness of those parties one did not vote for (for
more details: van der Eijk and Franklin, 1996; van der Brug, Fennema and Tillie, 2000). The
EES 1999 asked the `probability of future vote' question for 10 anti-immigrant parties, all
mentioned in the introduction, from seven European countries. Because different party
systems exist in Flanders and Wallonia, we will treat these two Belgian regions as two
separate political systems, thus yielding 8 systems: Austria, Flanders, Wallonia, Denmark,
France, Germany, Italy, and the Netherlands.
To assess whether voters evaluate these 10 parties by the same criteria as other parties,
our study concentrates on the electoral attractiveness of all parties (63 in total) in the eight
political systems included in this study. A valid way to analyse individual and inter-party
level variations in party preferences simultaneously can be realised by arranging the data in
the so-called `stacked' (or ‘pooled’) form that was first proposed by Stimson (1985) and after
that applied frequently in electoral research (e.g., MacDonald, Listhaug and Rabinowitz,
1991; van der Eijk and Franklin, 1996; Westholm, 1997; van der Brug, Fennema and Tillie,
2000). In this stacked data matrix each respondent is represented by as many `cases' as there
are parties for which (s)he was asked to indicate the probability of a future vote (for details
about the number of cases and weighting procedures, see Appendix). This matrix allows us to
apply multiple regression to explain parties' electoral attractiveness. By adding characteristics
of the political systems and the parties as variables in the stacked data matrix, such
characteristics can be included as variables in these regression analyses. In order to assess
whether voting for anti-immigrant parties involves a different kind of decision than voting for
other parties, we will estimate interaction terms for an anti-immigrant party on the one hand,
and a large set of independent variables on the other. Before getting into this, let us discuss
which independent variables are in the equation for predicting parties' electoral attractiveness,
and how these are treated in the stacked matrix.
The first predictor of party preference is the subjectively perceived distance between a
voter and the respective party in the data matrix on a left-right continuum. Idealistic voting
implies that the closer a party is to someone's own position in terms of left-right, the more
attractive this party will be for the person in question. The questionnaire contained a battery
of items in which respondents were asked to indicate their own position as well as that of each
political party on a 10-point scale of which the extremes were labelled left and right. From
these responses the perceived left-right distances were computed. The stronger the effect of
8
perceived left-right distance on electoral attractiveness, the stronger the extent of idealistic
voting.
The likelihood of someone voting for anti-immigrant parties will also increase when
(s)he agrees with its stance on some concrete issues (e.g., Billiet & De Witte, 1995). The
European Elections Study 1999 contains just one position issue for which respondents'
positions and their perceptions of party positions were measured: European integration. This
item yields one more predictor of party preference, i.e., the perceived distance on this scale
between each respondent and the respective party in the data matrix.
In order to estimate the importance of the effects of policy considerations, attitudes
towards three more policy areas were included as predictors of party preference. These policy
areas are, apart from European integration: immigration, the economy and the environment.
For each of these policy areas two survey questions were asked: ‘how important is this issue’
and ‘are you satisfied with the governments’ policies in this field’. To estimate the
contribution of these attitudes to party preference, a different procedure had to be applied,
because no measures of distances between respondents and parties are available. The
procedure involves linear transformations of the original variables. For an elaborate discussion
of this procedure, see Van der Eijk and Franklin (1996, chapter 20). As a consequence of these
transformations the new predictors of party preference are comparable across parties and
countries.4
Additional control variables were taken from the surveys in order to provide us with a
properly specified model. The relevant literature on party choice tells us that we should
control for the following variables: social class and income (e.g., Nieuwbeerta and Ultee,
1999), education (e.g., Swyngedouw, Billiet et Al., 1998), religion, approval of government
and approval of the European Union (e.g., Van der Eijk, Franklin and Van der Brug, 1999).
Class is measured with two separate variables: income and subjective social class. Religion is a
4 The linear transformations of the original variables are carried out in a number of steps. For each of the policy
areas, and for each of the parties in turn, series of regressions were performed with two predictors. The first
predictor is the respondents’ answer to the question how important (s)he considers the issue in question, and the
second variable is his/her satisfaction with the governments’ policies. The predicted values of these regressions
(y-hats) were saved and inserted in the stacked data matrix as new predictors of party preference. These
predicted values are simply linear transformations of the original variable and may therefore be used again as
predictors of party preference. The new variable is comparable over countries and parties. The actual variable
which is added to the stacked matrix is not the y-hat, but the deviation of the y-hats from their mean for each party. It
must be noted that as a result of the transformations the effect of these variables will always be positive.
9
composite variable of religious denomination and church attendance. To estimate the effect of
these variables on party preference, they were transformed in such a way that the effects are
comparable over countries and parties (the procedure is described in note 4).
Finally, we included a variable at the party level, party size, which represents a
strategic consideration that voters may take into account: when two parties are about equally
attractive on all relevant accounts, voters tend to vote for the largest one because it stands a
better chance of achieving its policy goals. We called this type of voting `pragmatic'. 5
In a number of subsequent steps we will assess to what extent voting for antiimmigrant parties is determined by particular considerations that exert less (or no) effect on
decisions about whether to vote for other parties.6
Results
Table 1 presents the results of three regression analyses. In the first one the model is estimated
for all 63 parties, in the second one only the 10 radical right-wing parties are included, and the
third analysis includes the 53 other parties. In the analyses of all 63 parties a (dummy)
variable was included that distinguishes the 10 anti-immigrant parties from the 53 others. The
regression coefficient for this variable tells us whether any differences exist between the
electoral attractiveness of anti-immigrant parties on the one hand and 'mainstream' parties on
the other, after controlling for the effects of the other independent variables. In other words,
the coefficient tells us whether - after we take the effects of social characteristics, policy
preferences, etceteras into account - anti-immigrant parties are considered more or less
5
Van der Brug, Fennema and Tillie (2000) included in their study an interaction of left-right distance on the one
hand and perceptual agreement about left/right positions of parties on the other. Since this variable had no
significant effect (in the data set with 8731 cases), it was excluded from our analysis here.
6
These party specific considerations are detected in the following way. First, we will start with an estimation of
the regression model on the stacked matrix that includes all parties. Also, we will do the same for the subgroup
of 10 radical right-wing parties, and for the 53 other parties. These analyses will allow only for an ad oculum
comparison of differences in the effect parameters. As a final step we will therefore explore whether significant
interaction effects exist between each of the anti-immigrant parties on the one hand and various predictors of
party preference on the other. This will be done for the model that was estimated for the total of 63 parties. Such
interaction effects, were they to exist, would indicate that voters have indeed party specific reasons to vote for
anti-immigrant parties. If we cannot find such interaction effects then we have to conclude that voters treat antiimmigrant parties just like any other party.
10
attractive than other parties. Here the findings are very similar to the findings of 1994,
presented by Van der Brug, Fennema and Tillie (2000). In both election years the dummy
variable that distinguishes anti-immigrant parties from mainstream parties turned out to yield
the only parameter in the equation that did not deviate significantly from zero. This indicates
that in the eyes of most voters anti-immigrant parties do not stand apart from mainstream
parties. Their electoral appeal stems from the same sources as the electoral appeal of other
parties. This finding has important political consequences. Political elites from various
countries persist in using a strategy that aims to stigmatise these parties, which recently has
led to a political boycott from 14 member states of the EU of the Austrian government. Our
results demonstrate that this strategy does not render the group of radical right-wing parties
less attractive to voters. To the extent that a so-called cordon sanitaire is meant to have an
effect on the electoral support for these parties, this strategy is thus not successful (see also:
van der Brug and Fennema, 2000)
<INSERT TABLE 1 ABOUT HERE>
Because different issues are included in the 1999 study than in the 1994 study, the
results are not fully comparable. In 1994 the policy issues questions focused on national
currencies, employment, national borders and immigration, while in 1999 the policy issue
questions focused on European integration, the economy, the environment and on
immigration. However, a few general remarks can be made. Approval of the EU appears to
exert a somewhat stronger effect in 1999 than it did in 1994. Despite these changes and a
slightly decreasing effect of socio-structural characteristics there is a remarkable stability. In
1999 as well as in 1994 the most important consideration for voters when evaluating parties is
which position each of them occupies in terms of left/right. The significance of the left/right
dimension for structuring the behaviour of voters has been observed by many scholars (e.g.,
Fuchs and Klingemann, 1990; Van der Eijk and Franklin, 1996; Van der Brug, 1999a;
Klingemann, Hofferbert and Budge, 1994; Hix, 1999).
Another stable finding is that party size is the variable with the second strongest effect
on party preference. The positive beta for party size shows that, after controlling for policy
positions and social characteristics, voters consider a larger party more attractive than a
smaller one. Voters who wish to influence policy making take into account the strategic
consideration that a large party has a better chance than a smaller one to realise its policy
11
goals. So, as in 1994, the analysis of party preferences show that electoral choices are
determined by a combination of ideological and pragmatic considerations.
How does this general model compare to the model for the 10 parties of the radical
right? The general conclusion of Table 1 should be that the two considerations that apply
mostly to voting behaviour in general, are just as important for these particular parties. The
beta’s for party size and left/right distance are of a very similar magnitude across the three
models. If votes for these parties were predominantly protest votes, i.e. votes cast to express
discontent with the mainstream parties, neither the size of a party nor its ideological position
would be factors that would increase its electoral attractiveness. The fact that the antiimmigrant parties are considered more attractive once they are larger and once they are
ideologically closer, shows that votes for these parties are generally not protest votes.
In contrast to what one might expect a priori on the basis of the nationalist ideologies
of parties of the radical right, differences on the issue of European integration exert an effect
on preferences for radical right-wing parties that is very similar to the effect it has on
preferences for other parties. The same goes for citizens’ satisfaction with the EU and with
the government.
Along with the similarities among the three models, there are some important
differences as well. Negative attitudes towards migration is a stronger determinant of party
preference for the radical right-wing parties than for other parties. Given the strong emphasis
of many of these parties on the issue of migration, this is not an unexpected finding, and it is
in line with the findings of 1994. It warrants the label of anti-immigrant parties. In addition,
the variables social class and religion are considerably less important as predictors of
preferences for radical right-wing parties than of preferences for other parties. In
contradistinction to some traditional parties that depend for their support on specific social
groups in a society, parties of the radical-right manage to attract their support from across all
social and religious groups. Various scholars have observed that cleavage politics is declining
in most countries and that this decline is largely compensated for by an increase in policy
voting (Franklin, 1992:400). Instead of relying on social positions as a 'cue' to decide which
party to vote for, the autonomous citizens vote largely on the basis of their policy preferences
(e.g., Rose and McAllister, 1986; Dalton, 1996). Thus the modern voters do not cast their
votes in agreement with the social group they belong to but in agreement with their own
ideological and policy preferences. In That sense the voters for anti-immigrant parties tend to
be more modern that the voters that support the traditional cleavage parties. If they are the
‘losers of modernity’ as Betz calls them, they appear to be modern losers of modernity.
12
The comparisons between electoral preferences for various anti-immigrant parties and other
parties have so far been made for all 10 parties of the radical right together, and on an ad
oculum basis. The design of our analyses, with a stacked data matrix in which electoral
preferences are studied for all parties simultaneously, provides the opportunity to
systematically study differences among the radical right-wing parties, and, also between
radical right-wing parties and other parties. If a variable has a different effect for one party
than for all other parties, the regression model should contain an interaction term between the
respective party on the one hand and this variable on the other.
To assess whether interaction terms should be added to the model, the following
procedure was employed. First, we estimated the model presented in Table 1 for all 63 parties
and saved its residuals (the part of the variance left unexplained by the general model). Then
we estimated the same model again for the set of radical right-wing parties, as well as for each
of them separately, with these residuals as the new dependent variable. If the independent
variables had a different effect on the electoral attractiveness of the radical right-wing parties
than on other parties, they should be correlated with the residuals of the overall regressions. If
significant effects are observed in the party specific regressions (with the residuals of the
overall analysis as the dependent variable), this would suggest that interaction terms should be
specified. This method thus allows us to detect whether left/right, European unification, and
party size exert effects on the electoral attractiveness of the radical right-wing parties that
differ from the 63 parties as a whole.7 Table 2 presents the parameter estimates of these
regressions.
<INSERT TABLE 2 ABOUT HERE>
The analyses for the 10 anti-immigrant parties together, presented in the first row of Table 2,
yield significant positive effects for left/right distance and for party size. The regressions were
estimated on the residuals of the overall regression presented in Table 1. For an interpretation
of the meaning of the parameters one should therefore not only look at Table 2, but also at the
7
The method does not allow us to estimate interaction effects for the other variables in the model. The reason is that
their effects were originally estimated with a procedure that involves a linear transformation of the original variables.
This procedure provides a valid way to estimate the strength of each of the independent variables, but at the same
time rules out the possibility to estimate interaction effects. As the topic of this paper focuses primarily on the effect
of party size and left-right distance (two variables that were not transformed) we do not consider this to be a problem
here.
13
original effects presented in the first column of Table 1. The overall effect from left-right
distance on electoral attractiveness is a negative one (Table 1): the larger the ideological
distance the less attractive a party is. The positive effect of left/right distance in Table 2 shows
that the (originally negative) effect of left-right distance is weaker for radical right-wing
parties than for the other parties. The positive effect of party size in Table 2 shows that radical
right-wing parties gain more by becoming larger than other parties. It should also be
emphasised however, that these effects are rather small. Only 2% of the variance that is left
unexplained by the general model can be explained by adding these specific effects for the
radical-right wing parties. In other words, the general conclusion is still that the motivations
underlying a choice for a radical-right wing party are generally the same as the motivations
for choosing other parties.
Our findings for 1999 have so far largely confirmed the findings of 1994. However,
when we inspect the differences among the various parties of the radical right, we must
conclude that things are quite different in 1999 than they were back in 1994. In 1994 there
was only one single party – the Dutch Centrumdemocraten – for which we found evidence of
protest voting. This continues to be the case in 1999. In 1999 there are more parties for which
this is the case, and there are much larger differences among these parties. In Table 2 we
arranged the parties according to party specific effect of left/right distance. We see that not
only the Centrumdemocraten, but also the Wallonian Front National, de German
Republikaner, the Lega Nord, the French Front national and the Danish Fremskridtparties all
attract a certain amount of protest votes. However, for the FPÖ a significant negative effect is
found. This shows that ideological proximity is a more important determinant of preferences
for this party than it is for preferences for other parties. The voters value the FPÖ in the first
place for its extreme rightwing policy options and not, as Simon Wiesenthal suggested out of
dismay for the Proporzdemokratie. Looking at differences among the effects of left/right and
differences among the proportions in explained variance, we may conclude that votes for the
four most successful parties of the radical right, FPÖ, Alleanza Nationale, Vlaams Blok and
Dansk Folkeparti, are at least as heavily determined by ideology as votes for other parties.
At the other side of the spectrum, we find three electorally unsuccessful parties: the
German Republikaner, the Dutch Centrumdemocraten and the Wallonian Front National,
where ideology matters very little. It could thus well be that votes for these parties are largely
inspired by protest. Then there are three parties that take an intermediate position:
Fremskridtspartiet, the French FN and Lega Nord. Ideology is a somewhat less important
determinant of the electoral attractiveness of these parties than of the electoral attractiveness
14
of other parties. However, if we take into account the proportions of explained variance in
these models, we must conclude that the causal mechanism that generates votes for these
parties is very similar to the mechanism that determines preferences for other parties.
Conclusion and discussion
In their 1994 study on electoral preferences for anti-immigrant parties Van der Brug,
Fennema & Tillie (2000) concluded that most voters consider anti-immigrant parties as
normal parties. In this 1999 replica we find the main conclusions of the 1994 study confirmed
for FPÖ, Alleanza Nazionale, Dansk Folkeparti and Vlaams Blok. Preferences for these
parties are largely determined by the same variables that determine preferences for other
parties, particularly ideological proximity and party size. Other variables play a modest or
minor role. Interestingly enough sociological variables of class and religion play a more
important role for the mainstream parties than for these anti-immigrant parties. Rather than to
attract ‘the losers of modernity’ (Betz, 1994), the latter parties attract support across all social
strata. Given the general decline of cleavage politics, these anti-immigrant parties are (in
terms of the way they attract their electoral support) more modern than various traditional
parties. Inglehart’s ‘silent revolution’ has come full circle (Inglehart, 1977). Also the antiimmigrant parties’ negative attitude toward the European Union does not have a substantive
impact on their attractiveness. Finally, the anti-immigrant parties’ stance on immigration is
more important for its voters than it is for voters of mainstream parties. Thus, we consider our
label for these parties to be well chosen.
In the 1994 study preferences for the anti-immigrant parties FPÖ, the Alleanza
Nazionale and the Dansk Folkeparti were not included. The 1999 results for these three
parties show some interesting results. Contrary to the assumption of many political
commentators, the FPÖ voters are even more motivated by ideological considerations than the
voters for the SPÖ and ÖVP. Thus we may conclude that FPÖ voters have not been ‘led
astray’ by their frustration about political clientelism of the SPÖ and ÖVP. They have voted
for the FPÖ predominantly because of its radical rightwing ideology, not in spite of it.
For the Italian and Danish voters the preference for the Alleanza Nazionale and the Dansk
Folkeparti is determined by ideological considerations to the same extent as the preference for
other parties. The electorate, then, considers these two parties as ‘normal’ parties.
The conclusion of the 1994 study - that voters evaluate anti-immigrant parties largely
by the same standards by which they evaluate mainstream parties - was not substantiated for
15
the Fremskridtspartiet, de French Front national, the Lega Nord, the Republikaner, the
Wallonian Front National and the Centrumdemocraten. These parties distinguish themselves
from the mainstream parties because the ideological distance that the voters see between
themselves and the parties’ program carries less weight than it does for the mainstream
parties. Protest voting may thus play a role here, a fact that was established already for the
Centrumdemocraten (and to a much lesser extent for the Lega Nord and the French Front
national) in the 1994 study. However, for the Wallonian Front National, the Republikaner and
the Fremskridtpartiet this is a new finding, which indicates that preference for these parties
has between 1994 and 1999 changed in the direction of protest voting. How should we
interpret such shift? The first explanation may be that these parties are small and in decline.
The Wallonian Front National has, contrary to the Vlaams Blok, always been a marginal
party. The Republikaner party has suffered from internal struggles since the early nineties and
its electoral support dropped as a result. They are now a marginal party. The Fremskridtpartiet
fell into electoral insignificance after 1994. We may conclude that the smaller parties attract
more protest voters, and that parties in decline lose their ideological and pragmatic voter. The
conclusion that protest voting is more common in the case of small parties is supported by a –
slight – decrease of ideological voting for the French Front national in 1999 as compared to
1994. In 1998 the Front national had split into two rival parties, one led by Bruno Mégret
(Mouvement National) and the other under the leadership of Jean-Marie Le Pen, which
proved disastrous for both of them.
Our finding that the electoral support for the Republikaner is determined by other
considerations in 1999 than in 1994, leads us to critically reflect upon the findings of Lubbers
and Scheepers (2000). Because they used party choice rather than party preference as the
dependent variable they were forced to add the Republikaner voters from 1991, 1992 and
1996 into one compiled dataset in order to obtain a sufficient number of voters for the
Republikaner. By this procedure they obtained 155 Republikaner voters (Lubbers and
Scheepers, 2000: 70), who voted at different times for the Republikaner. Yet the distinction
between the three moments in time is not taken into account in their analyses. The results of
our study, as well as the obvious electoral volatility of anti-immigrant parties, underline how
important it is to take into account this dynamical component in the data. Large differences
exist among the electoral fate of anti-immigrant parties in the 1990s. As a case in point we
may contrast the rise of Vlaams Blok and Dansk Folkeparti with the rise and decay of FPÖ
and with the decay of the French Front nationale. In our view the most important and
16
challenging research question that lies ahead of students of anti-immigrant parties, is to
determine the factors that explain these differences in success.
The Lega Nord may be a case apart. Since this anti-immigrant party explicitly aims at
secession of Padania, its electoral support may partly fall beyond the left-right dimension that
assumes –by definition - a unified political space. Hence the Lega Nord is the only relatively
successful anti-immigrant party that is not considered by many voters as a normal party. The
other large parties under consideration, FPÖ, Vlaams Blok, Alleanza Nazionale and Dansk
Folkeparti, seem to have conquered a political space at the very right of the political spectrum
that makes them normal parties in the eyes of the voters. However, the fate of the French
Front national should warn against reaching too hasty conclusions.
Given the importance of the issue of immigration as well as ideological proximity for
the electoral support of the main anti-immigrant parties, three aspects seem to be important
for the political opportunity structure that determines the success of these parties. First, as
long as the immigration policy is a salient issue, and as long as anti-immigrant parties are
uncontested owners8 of this issue, they can be successful. Secondly, the success of antiimmigrant parties depends largely on their ability to present themselves to voters with a clear
ideological image. Vlaams Blok, for example, emphasises the fact that it is a "right-wing
nationalist" party. This clear ideological profile makes its stands on other issues, such as
crime, more predictable, which in turn allows it to successfully mobilise support around these
other issues. Finally, after an anti-immigrant party has successfully presented itself at the very
right of a political spectrum, its electoral success depends largely on the actions of competing
parties from the right. Voters from Bavaria with intolerant attitudes towards migrants have no
need to vote for the Republikaner, since the CSU presents an alternative.
We would like to end our research note with a comment on the mainstream parties’ strategy
towards anti-immigrant parties that has been labelled cordon sanitaire. Such a strategy of
non-co-operation with these parties and the attempt to exclude them from the political arena –
sometimes also by legal means – is quite often based on the assumption that many voters for
such parties do not ‘really’ support the program of those parties. We have shown that this is
not true of the most successful ones. Also, the assumption that these parties somehow
represent evil forces in society is not an opinion that is found in the electorate’s perception of
8
For details about the theory of issue ownership, see Budge & Farlie (1983) or Budge, Robertson & Hearl
(1987).
17
these parties. To treat them as normal parties would be more in line with the voters’
conception of them. The coalition of the FPÖ and the ÖVP in Austria also has shown that the
Weimar Republic is not comparable to the contemporary setting of these anti-immigrant
parties. On the contrary, rather than gaining momentum, the FPÖ is rapidly losing popular
support since it has participated in a coalition government. Some anti-racist researchers have
pointed out that a cordon sanitaire is likely to split the anti-immigrant parties in rival factions
that disagree about the ways to confront the cordon sanitaire (Donselaar 1993). We would like
to point out that the opposite could also be argued: the split in the French Front national
occurred after the FN had broken the cordon sanitaire in the region Rhone-Alpes in 1998.
18
References
Betz, H-G. (1994). Radical Right-wing Populism in Western Europe. Basingstoke:
Macmillan.
Betz, H-G. (1998a). Introduction, In: Hans-Georg Betz & Stephan Immerfall (eds.), The new
politics of the right: neo-populist parties and movements in established democracies.
Basingstoke: Macmillan: 1-10.
Betz, H-G. (1998b). Against Rome: The Lega Nord. In: Hans-Georg Betz & Stephan
Immerfall (eds.), The new politics of the right: neo-populist parties and movements in
established democracies. Basingstoke: Macmillan: 45-58.
Betz, H-G. & S. Immerfall (1998). The new politics of the right: neo-populist parties and
movements in established democracies. Basingstoke: Macmillan.
Billiet, J. & H. de Witte (1995). Attitudinal dispositions to vote for a "new" extreme rightwing party: the case of "Vlaams Blok". European Journal of Political Research 27 (2): 181202.
Brug, W. van der (1999). Voters’ perceptions and party dynamics. Party Politics 5 (2): 147169.
Brug, W. van der, M. Fennema & J. Tillie (2000). Anti-immigrant parties in Europe:
Ideological or protest vote? European Journal of Political Research 37:77-102.
Brug, W van der & M. Fennema (2000). Is een cordon sanitaire wel gezond? Streven
September 2000: 692-696.
Budge, I. & D. J. Farlie (1983). Explaining and Predicting Elections. Issue Effects and Party
Strategies in Twenty-three Democracies. London: George Allen and Unwin.
19
Budge, I., D. Robertson, & D. Hearl (1987). Ideology, Strategy and Party Change: Spatial
Analyses of Post-War Election Programs in 19 Democracies. Cambridge: Cambridge
University Press.
Dalton, R.J. (1996). Citizen Politics in Western Democracies; Public Opinion and Political
Parties in Advanced Industrial Democracies. Chatham House, Chatham.
Derks, A. & K. Deschouwer (1998). Vrijzninigen, ongelovigen en protest. In: Marc
Swyngedouw, Jaak Billiet, Ann Carton & Roeland Beerten (eds.), Kiezen is verliezen.
Onderzoek naar de politieke opvattingen van Vlamingen. Acco, Leuven/Amersfoort:85112.22.
Donselaar, J. van (1993). Post-war fascism in the Netherlands, Crime, Law and Social
Change 19: 87-100.
Eatwell, Roger (1998). The Dynamics of Right-wing Electoral Breakthrough. Patterns of
Prejudice 32: 1-31.
Eijk, C. van der & M. Franklin (1996). Choosing Europe? The European Electorate and
National Politics in the Face of Union. Ann Arbor: The University of Michigan Press.
Eijk, C. van der, M. Franklin, & W. van der Brug (1999). Policy Preferences and Party Choice.
In: H. Schmitt & J. Thomassen (eds.), Political Representation and Legitimacy in the European
Union. Oxford: Oxford University Press.
Eijk, C. van der, M. Franklin & H. Schmitt (1999). European Elections Study 1999: Design,
Implementation and Results. Colgne: Zentralarchiv für Emprische Sozialforschung.
Fennema, M. (1997). Some conceptual issues and problems in the comparison of antiimmigrant parties in Western Europe. Party Politics 3: 473-492.
Fennema, M. & M. Maussen (2000). Dealing with Extremists in Public Discussion: Front
national and “Republican Front” in France. The Journal of Political Philosophy 8:379-400.
20
Fennema, M. (2000). Legal repression of extreme-right parties and racial discrimination. In: R.
Koopmans & P. Statham (eds), Challenging immigration and ethnic relations politics.
Oxford: Oxford Uiversity Press: 119-144.
Franklin, M. (1992). The Decline of Cleavage Politics. In: M. Franklin, T. Mackie, & H.
Valen (eds.), Electoral Change. Responses to Evolving Social and Attitudinal Structures in
Western Countries. Cambridge: Cambridge University Press.
Fuchs, D. & H.D. Klingemann (1990). The left-right scheme: theoretical framework. In: M.K.
Jennings & J. van Deth (eds.), Continuities in Political Action: A Longitudinal Study of
Political Orientations in Three Western Democracies. Berlin: De Gruyter.
Griffin, R., 1996. The `post-fascism' of the Alleanza Nazionale: a case study in ideological
morphology. Journal of Political Ideologies 1: 123-145.
Hix, S. (1999). Dimensions and Alignments in European Union Politics: Cognitive
Constraints and Partisan Responses. European Journal of Political Research 35:69-106.
Holsteyn, J. van, 1990. En wij dan? De kiezers van de Centrumdemocraten. Socialisme en
Democratie 47: 158-161.
Inglehart, R. (1977). The Silent Revolution. Changing Values and Political Style among
Western Politics. Princeton: Princeton University Press.
Klingemann, H.D., R.I. Hofferbert, & I. Budge (1994). Parties, Policies, and Democracy.
Boulder, Col: Westview Press.
Lucardie, P. (1998). The Netherlands: The Extreme Right Center Parties. In: Hans-Georg
Betz & Stephan Immerfall (eds.), The new politics of the right: neo-populist parties and
movements in established democracies. Basingstoke: Macmillan: 111-124.
Lucardie, P. (2000). Prophets, purifiers and prolocutors: towards a theory of emergence of
new parties. Party Politics 6, 2: 175-186.
21
Lubbers, M. (2001)
Lubbers, M., P. Scheepers & J. Billiet (2000). Individual and Contextual Characteristics of the
Vlaams Blok vote. Acta Politica, 35: 363-398.
Lubbers, M. & P. Scheepers (2000). Individual and Contextual Characteristics of the German
extreme right vote in the 1990s. European Journal of Political Research 38 (1): 63-94.
Macdonald, S.E., O. Listhaug, & G. Rabinowitz (1991). Issues and Party Support in
Multiparty Systems. American Political Science Review 85:1107-31.
Mayer, Nona (1998). The French National Front. In: Hans-Georg Betz & Stephan Immerfall
(eds.), The new politics of the right: neo-populist parties and movements in established
democracies. Basingstoke: Macmillan: 11-26.
Mayer, N. & P. Perrineau (1992). Why do they vote for Le Pen? European Journal of
Political Research 22 (1): 123-141.
Mayer, N. & P. Moreau (1995). Electoral support for the German Republikaner and the
French National Front 1989-1994. Paper presented at the Workshop on Racist Parties in
Europe of the ECPR Joint Sessions of Workshops, Bordeaux 27 April - 2 May 1995.
Mudde, C. & J. van Holsteyn (2000). The Netherlands: explaining the limited success of the
extreme right. In: P. Hainsworth (ed.), The Politics of the Extreme Right. From the Margins to
the Mainstream. Pinter, London and New York: Pinter: 144-171.
Nieuwbeerta, P. & W. Ultee (1999). Class Voting in Western Industrialized Countries, 19451990: Systematizing and Testing Explanations. European Journal of Political Research
35:123-160.
Riedlsperger, M. (1998). The Freedom Party of Austria: From Protest to Radical Right
Populism. In: Hans-Georg Betz & Stephan Immerfall (eds.), The new politics of the right:
neo-populist parties and movements in established democracies. Basingstoke: Macmillan:2744.
22
Rose, R. & I. McAllister (1986). Voters Begin to Choose. Beverly Hills: Sage.
Stimson, J.A. (1985). Regression in Space and Time: A Statistical Essay. American Journal of
Political Science 29: 914-47.
Stouthuysen, P. (1993). Extreem-Rechts in na-oorlogs Europa. Brussel: VUB Press.
Svasand, L. (1998). Scandinavian Ringt-Wing Radicalism. In: Hans-Georg Betz & Stephan
Immerfall (eds.), The new politics of the right: neo-populist parties and movements in
established democracies. Basingstoke: Macmillan: 77-94.
Swyngedouw, M. (1998). The Extreme Right in Belgium: Of a Non-existent Front National
and an Omnipresent Vlaams Blok. In: Hans-Georg Betz & Stephan Immerfall (eds.), The new
politics of the right: neo-populist parties and movements in established democracies.
Basingstoke: Macmillan: 59-76.
Swyngedouw, M., J. Billiet, A. Carton & R. Beerten (1998). De (on)redelijke kiezer.
Onderzoek naar de politieke opvattingen van Vlamingen. Verkiezingen van 21 mei 1995.
Leuven: ACCO.
Tillie, J. (1995). Party Utility and Voting Behaviour. Amsterdam: Het Spinhuis.
Westholm, A. (1997). Distance Versus Direction: The Illusory Defeat of the Proximity
Theory of Electoral Choice. American Political Science Review 91:865-83.
Winkler, J. & S. Schumann (1998). Radical Right-Wing Parties in Contemporary Germany.
In: Hans-Georg Betz & Stephan Immerfall (eds.), The new politics of the right: neo-populist
parties and movements in established democracies. Basingstoke: Macmillan: 95-110.
23
TABLES
Table 1: Regression analyses to explain party preferences (standardised coefficients)
All Parties 10 radical right-wing
53 other parties
parties
Party size
.24
.27
.23
Left/right distance
European integration
distance
-.29
-.08
-.28
-.08
-.29
-.08
EU-approval
Government approval
.08
.09
.09
.10
.09
.09
Importance and satisfaction
about policies:
European integration
Immigration
The economy
The environment
.03
.06
.05
.07
.02ns
.13
.09
.04
.03
.05
.05
.07
Social class and income
Religion
Education
.09
.12
.04
.03
.07
.06
.010
.13
.04
-.01ns
-
-
Adjusted R2
.31
Source: European Elections Study 1999
.36
.29
Dummy extreme right-wing
party
24
Table 2: Results of regressions to explain residuals of the general model
Ideological
Distance Party size
distance (LeftEuropean
right)
Unification
10 radical right-wing
.07
.03ns
.14
parties
FPÖ
-.24
.03ns
Dansk Folkeparti
-.05ns
-.01ns
Alleanza Nationale
-.01ns
.00ns
Vlaams Blok
.08ns
.00ns
Fremskridtspartiet
.14
-.02ns
FN (French)
.17
.09ns
Lega Nord
.22
-.05ns
Republikaner
.24
.18
FN (Wallonian)
.26
.13ns
Centrumdemocraten
.41
.09
Source: European Elections Study 1999
Adjusted R-square
.02
.05
-.00
-.00
-.01
.02
.04
.04
.10
.08
.19
APPENDIX
The stacked matrix, combining party preferences for the 63 parties from eight political
systems has a total of 76.354 observation units. To estimate the parameters of the regression
models, units of analyses are weighted. The weight variables are computed in two steps. As a
result of the weight factor applied in the first step (using the variable POLWEIGH), respondents
in each system are weighted in such a way that their party choice in the European Elections 1999
reflect exactly the actual election results. This weight factor is applied in the analyses that pertain
to each party separately (see Table 2). In the second step the variable POLWEIGH is multiplied
by a (different) constant for each system, so that the eight systems in the stacked matrix contain
1.091 units of analysis each (yielding a total of 8731, which equals the original sample size).
This weight will be used for the analyses in which parties from the 8 different political systems
are analysed simultaneously. Each time groups of parties are selected, the variable POLWEIGH
is multiplied by yet different constants for each system, so that in all regressions presented in
Table 1 the eight systems in the stacked matrix contain 1.091 units of analysis each (yielding a
total of 8731, which equals the original sample size).
25