Scholarly article on topic 'Projecting EU Referendums: Fear of Immigration and Support for European Integration'

Projecting EU Referendums: Fear of Immigration and Support for European Integration Academic research paper on "Political Science"

0
0
Share paper
Academic journal
European Union Politics
OECD Field of science
Keywords
{""}

Academic research paper on topic "Projecting EU Referendums: Fear of Immigration and Support for European Integration"

European Union Politics

DOI: 10.1177/1465116505049608 Volume 6 (1): 59-82 Copyright© 2005 SAGE Publications London, Thousand Oaks CA, New Delhi

Projecting EU Referendums

Fear of Immigration and Support for European Integration

^ Claes H. de Vreese

Amsterdam School of Communications Research, The Netherlands

^ Hajo G. Boomgaarden

Amsterdam School of Communications Research, The Netherlands

ABSTRACT

This study tests competing hypotheses about public support for European integration and projects referendum voting behaviour. It emphasizes anti-immigration sentiments as a key variable for understanding reluctance about integration. Drawing on survey data, it is shown that anti-immigration sentiments, economic considerations and the evaluation of domestic governments are the strongest predictors of both attitudinal support for integration and individuals' propensity to vote 'yes' in a referendum on the enlargement of the European Union (EU).

KEY WORDS

■ attitudes towards immigrants

■ public opinion

■ referendums

■ support for European integration

Why do some people embrace the notion of European integration whereas others oppose it? This question has generated a considerable and growing amount of scholarly research and popular debate. The usual suspects for understanding variation in popular support for integration include levels of cognitive mobilization (e.g. Inglehart, 1970), utilitarian and economic considerations (e.g. Gabel and Palmer, 1995), satisfaction with the incumbent government (e.g. Franklin et al., 1995; Ray, 2003), as well as social-demographic characteristics and political-ideological preferences (e.g. Gabel, 1998a). More recently, considerations of national identity such as national pride and territorial identity (Carey, 2002), national attachment and exclusive national identity (Marks and Hooghe, 2003), fear of loss of national identity (Christin and Trechsel, 2002), and perceived cultural threat (McLaren, 2002) have been added to the equation.

We first briefly review each of these explanations. Then we discuss the relevance of anti-immigration sentiments in explaining EU support. McLaren (2002) emphasizes cultural threat as an important negative predictor of EU support. We argue that it is rather people's negative out-group bias that explains opposition to European integration. Drawing on originally collected survey data, we simultaneously assess the power of the different explanations. Our study confirms the importance of economic evaluations and domestic political considerations and it explicates the importance of immigration-related attitudes. In addition to explaining attitudinal support for the EU, we also demonstrate how the same predictors might influence the outcome of future referendums on European integration issues. Our individual-level model of the propensity to vote 'yes' in an EU referendum shows that the predictors driving attitudes towards European integration are also the strongest predictors when modelling vote choice.

Why public opinion about the European Union matters

The European Union is often referred to as an elitist project that does not have widespread public support. Considerable effort has gone into exposing and analysing the democratic deficit of the EU. Inherent in this debate is the notion of legitimacy. As Scharpf (1970) argues, legitimacy builds upon principles of the authorization of power-holders, responsiveness in the exercise of power, and accountability. The democratic deficit is a lack of legitimacy, in that EU decisions are not sufficiently responsive to public preferences and scrutiny (Scharpf, 1997). This gap between elite opinion and public opinion on European integration has also been established in more empirical terms,

showing differences in the areas in which political elites and European publics are willing to confer power to the EU (Hooghe, 2003).

Given the weak system of authorization of power-holders through European Parliamentary elections (e.g. van der Eijk and Franklin, 1996), arguably legitimacy has to come not only from elections but also from the process of governance. This can take a number of forms. Most directly, public preferences can be established through referendums on issues of European integration. European integration is the most voted-on issue in the world (de Vreese and Semetko, 2004a), and the use of referendums to decide on issues of membership, key policies, endorsement of treaties and constitutional documents is increasing. Nevertheless, obviously only a small fraction of EU decisions is legitimized by public support in a referendum.

In general, political systems are viable on condition that they enjoy public support. This is particularly true for the European Union, which, in the absence of, for example, supranational means of law enforcement, depends on public support and acceptance (Caldeira and Gibson, 1995). The ingredients shaping public support for European integration are, however, contested. Below, we review previous strands of research as well as articulate the importance of anti-immigration sentiments in public support for European integration.

Cognitive mobilization

Inglehart (1970) suggested that citizens with high cognitive mobilization are more at ease with a supranational entity. Later, Inglehart (1977) argued that citizens who have a political value system favouring non-material values (such as self-fulfilment and concerns with democracy) above material values (such as financial security) are more favourable towards European integration. The first group, dubbed as post-materialists, is likely to perceive European integration as a vehicle for social reform and tend to consider politics at a more abstract level, which (according to Inglehart) promotes support for the EU.

This idea was demonstrated using bivariate analyses of Eurobarometer data but - as, for example, Gabel (1998a) has suggested - the conclusions are at best tentative given the lack of consistent empirical support in the data and the absence of controls for other, potentially confounding factors such as education. Later analyses have found only limited support for this idea, either by limiting the argument to original member states only (Anderson and Reichert, 1996) or by controlling for a number of the explanations outlined below (Gabel, 1998a).

62 European Union Politics 6(1) Cost/benefit analysis

The second group of studies posits that 'EU citizens from different socioeconomic situations experience different costs and benefits from integrative policy' (Gabel, 1998a: 336). These studies explain support for European integration in terms of income, education, occupational skills and proximity to border regions (e.g. Anderson and Reichert, 1996; Gabel and Palmer, 1995). Pepermans and Verleye (1998) found national economic pride and satisfaction to be a key explanatory variable in support for the euro across the EU countries (for conflicting evidence, see Bosch and Newton, 1995). Anderson (1998) showed that, when contrasting economic and political effects in a simultaneous multivariate analysis, economic variables were in part mediated by political variables. This, he suggests, calls for the inclusion of political variables to understand variation in popular support.

Domestic politics

The key political variables come from the domestic political realm. The argument is that, given the low level of actual information about the integration processes, citizens are likely to resort to proxies when formulating their view on integration, and these proxies are likely to be based on national political considerations (Anderson, 1998; Franklin et al., 1994). In particular, the importance of government approval and support for incumbent political parties has been considered.

Franklin and colleagues (Franklin et al., 1994; Franklin et al., 1995) even go on to say that domestic political considerations drive not only opinions about integration but also voting behaviour in European elections and national referendums on European issues: 'referenda conducted in the context of national party politics, with the government of the day urging ratification of a treaty they have themselves negotiated, will inevitably be contaminated by popular feelings about the government' (Franklin et al., 1995: 102).1

National identity and threats to national and cultural integrity

A number of recent studies suggest that citizens' feelings of national attachment and their perceptions of threats to the nation-state and to the nation's interests and cultural integrity are potential considerations when expressing support for the EU (Kritzinger, 2003). Marks and Hooghe (2003) differentiate between cultural and economic threats and find these, when controlling for economic evaluations, to have a significant impact on EU

support. For the Swiss case, Christin and Trechsel (2002) show that perceived threat to national interests is a strong predictor of support for joining the EU. This perceived threat in turn is strongly related to the image people hold of neighbouring countries and to their attachment to core Swiss values. Furthermore, their analysis reveals that a strong national identity and perceptions of negative consequences of EU membership for the general economy are directly and negatively related to support for membership.

Whereas Christin and Trechsel (2002) consider the perceived threat to national interests, McLaren (2002) argues that reluctance about integration is a function of the perceived cultural threat. She contends that it is the 'changing nature of the nation and the nation-state that will lead many Europeans to be critical of the EU - since this institution is likely to be seen as contributing to this change' (McLaren, 2002: 554). Her argument is that, given that European citizens have been socialized to accept the power and sovereignty of the nation-state, the idea of advanced European integration, which implies a potentially weakened role for the nation-state and a redistribution of sovereignty, poses a threat to the symbolic value of the nation-state. Threats may come from non-national changes in society, including immigration and globalization.

Both studies frame their threat argument in terms of feelings of national identity and national attachment, but the conceptualization of threat differs considerably. McLaren (2002: 555) contends that 'it is the protection of the in-group (the nation) and the group identity that is at stake'. Although she acknowledges that her variables indeed measure underlying feelings of xenophobia, she presents her findings in the light of in-group protection and national attachment. Christin and Trechsel (2002), by contrast, measure the strength of national identity and find that national identity relates to EU support as well as to the perceived threat to national interests of EU integration. However, since they are not simultaneously controlling for general economic expectations, it remains an open question whether it is indeed the perceived threat to national interests that drives EU membership support.2

In sum, these studies leave us with an inconclusive picture. We know that identity considerations matter, but it remains open whether it is individuals' feelings of attachment to the nation-state and its values and institutions, whether it is their sense of having an exclusive national identity, or whether it is the perceived threat to the nation-state that really matter for supporting European integration.

64 European Union Politics 6(1) Immigration

The present study extends and explicates McLaren's (2002) argument that it is the threat posed by outsiders that fuels public anti-EU integration sentiments rather than merely identity-based considerations. Recent political developments and changes in the political landscape in a number of European countries might have implications for political attitudes. Therefore, it is important to consider certain attitudes beyond support for the incumbent government as a reference point for citizens when expressing their opinion about European integration. Europe has experienced an increase in popularity of anti-immigrant and often anti-EU populist political parties.3 We know that, on the individual level, anti-immigration sentiments are among the core predictors of support for populist parties (Lubbers et al., 2002; Van der Brug et al., 2000). In recent years, anti-immigration sentiments have been moving to the forefront of politically relevant attitudes in a number of European countries. The popularity of populist anti-immigrant parties serves as an indicator of the relevance and significance of immigration-related attitudes for political opinion formation. We are not thereby asserting that right-wing parties cause anti-EU sentiments; however, we believe that an increasing and negative emphasis on immigration-related issues in domestic politics might cause people to consider their stance on this issue when forming an opinion about the EU.

Why might people's attitudes towards immigration be related to support for European integration? Sniderman et al. (2000: 62-8) identify people's readiness to categorize themselves or others into groups as the central and common factor in the formation of hostility towards immigrants. Minimal group experiments (e.g. Tajfel, 1981) and social identity theory (e.g. Tajfel and Turner, 1979; for an overview see Brown, 2000) inform us that, based upon group categorization, people tend to show a favourable bias towards members of their own group and an unfavourable one against members of other groups. Thus, people have a tendency to make in-group versus out-group distinctions that are advantageous for their in-group and unfavourable for the out-group. It has been shown that people who tend to categorize immigrants into an out-group are also more likely to categorize others into out-groups in general and also to show hostility towards these (Sniderman et al., 2000: 87-9).

Following this, we can expect that Europeans holding negative attitudes towards immigrants will show a greater readiness to categorize others in general, which is likely to yield unfavourable evaluations of these out-groups. European integration brings together people from different countries, regions and cultures, and arguably with different religions and ethnicities, who can

be readily categorized on the basis of these features. This fuels negative assessments of these groups, and therefore people holding negative attitudes towards immigrants are more likely to reject the idea of further European integration. It is not in-group favouritism but rather a negative out-group bias, indicated by hostility towards immigrants, that matters in understanding support for the EU.

McLaren (2002) demonstrates the importance of immigration attitudes in understanding support for European integration. Although she frames her argument as a perceived threat to the nation-state, the measures from the Eurobarometer utilized to operationalize this perceived threat are in fact closely related to indicators of anti-immigration sentiment. She thereby forcefully demonstrates that previous research on public support has missed a key variable driving opposition to European integration. We extend McLaren's work by employing measures directly concerning immigration and a greater array of themes, and not solely measuring threat perceptions. National identity and economic concerns are related to anti-immigration sentiments but anti-immigration attitudes are conceptually different from national identity (Sniderman et al., 2004). Anti-immigration sentiments tap people's readiness to show negative out-group bias and therefore to oppose further European integration; they are not a proxy measure for national identity because personality traits and personal values, as well as perceptions of group competition or a general sense of insecurity, can cause people's hostility towards immigrants. It is the combination of these factors into a politically relevant attitude, we believe, that makes people more likely to oppose European integration. Our argument is that there is a great need to specify the particular role that anti-immigration sentiments play in support for European integration, in particular in the context of the increased electoral significance of populist parties campaigning on anti-immigration and anti-EU policies, which indicates the increasing prominence and political relevance of anti-immigration attitudes.

Hypotheses

We hypothesize the following relationships between a number of individual characteristics and attitudes and support for European integration:

H1: Low levels of fear of immigration are related to higher levels of support, as argued above.

H2: Positive evaluations of the incumbent government are related to higher support (Franklin et al., 1994).

H3: Positive economic evaluations are related to higher support (Anderson, 1998).

H4: Higher levels of political sophistication are related to higher support (Ingle-

hart, 1970).

In addition - for a full model specification - we control for gender, age, education, occupation, political values and ideology, which we predict to be related to support as follows: men are more supportive (see Gabel, 1998a; Nelsen and Guth, 2000), higher educational groups are more supportive (see Gabel, 1998a), executive and managerial occupations are more supportive (see Gabel, 1998a), post-materialistic values are related to higher support (see Inglehart, 1970), and centre and right ideological preferences are related to higher support (see Gabel, 1998a).

We test these hypotheses in one model. However, we take our argument one step further and link attitudes towards integration to referendum voting behaviour (see also Gabel, 1998b). Previous studies of EU-related referendums have not included measures for all the relevant antecedents of public support for the EU (for an overview, see de Vreese and Semetko, 2004b). Moreover, none of these referendum studies has specifically addressed the role that attitudes towards immigration might play. We therefore also assess the impact of the different predictors of support for European integration in the form of vote choice in a referendum on a European integration topic. We model vote choice using the same measures of support for European integration - also hypothesizing the same direction of influence - given that previous research on EU-related referendums has emphasized these variables too (see, e.g., de Vreese and Semetko, 2004b; Franklin et al., 1995; Hug, 2003; Siune and Svensson, 1993; Svensson, 2002). We thereby employ an attitudinal as well as a behaviour-intention dependent variable to show that the predictors driving attitudes towards the EU also matter when people are asked to support or reject further integration through direct voting behaviour in a referendum.

Methods

We test our model of support for European integration by drawing on newly collected survey data in two EU countries: Denmark and the Netherlands. We chose these two countries for a variety of reasons: first, the populations in the two countries show a similar level of support for the EU (Eurobarometer 58, 2003);4 second, satisfaction with EU democracy is comparable in the two countries (Karp et al., 2003); third, the economic situation in the two countries is similar (OECD, 2002); fourth, and of particular significance for this study, both countries have experienced the electoral success of populist

parties campaigning on strong anti-immigration messages, which indicates that anti-immigration sentiments are an important political attitude among both populations.5

We rely on two identical surveys of representative samples of the Danish and Dutch adult populations. The surveys were carried out in November 2002.6 The sample sizes were 1444 in Denmark and 2396 in the Netherlands. The response rates were 77.9% in Denmark and 70.9% in the Netherlands.7 All previous research on public support for European integration has relied on Eurobarometer data. With our data we can provide a novel link between attitudes towards European integration and voting intention, which is not possible using Eurobarometer data.

Measures

The first dependent variable was an index of support for the EU, measured by five items tapping general attitudes and opinions about (the extent of) European integration (Denmark: M = 3.12, SD = 0.86, alpha = 0.82; the Netherlands: M = 2.99, SD = 0.61, alpha = 0.68). Answers were given on five-point agree-disagree scales:

(1) European integration is being pushed too fast

(2) The EU is a threat to smaller countries such as Denmark/the Netherlands

(3) I would be willing to make a sacrifice to help a less strong country

(4) The membership of Denmark/the Netherlands is a good thing

(5) The EU has more disadvantages than advantages for people like me

The items were recoded when appropriate to form a scale of EU support.8

Our second dependent variable was a measure of voting intention in a referendum on the enlargement of the European Union. The question was: 'If a referendum were held on the issue of the enlargement of the EU, would you vote in favour or against the enlargement?'9 We emphasize that we are not making inferences about the substantive level of support ('yes' votes) in a referendum, but rather are focusing on the underlying explanations.

The independent variables included gender (coded as female), age (in years), education10 and occupational status. For the last variable, we follow Gabel (1998a: 343), who summarizes that 'manual laborers and the unemployed will be less supportive of integration than executives and professionals'.11 Ideological preference was tapped by a 10-point left-right self-placement measure. To test the 'cognitive mobilization' theory, we used an index of political sophistication. This is a combined measure of political

knowledge and political interest. Political values were assessed using a standard measure tapping post-materialism (Inglehart, 1990) by asking respondents to rate the two most important tasks of the government (see the appendix). We used respondents' assessment of the domestic government on a five-point scale from 'very bad' to 'very good' to measure the impact of domestic political considerations. To test the utilitarian perspective, we included occupational status and in addition used a measure of economic evaluation (following Anderson, 1998) in the form of economic expectations for the coming 12 months ranging from 'a lot worse' to 'a lot better'.

Finally, we included a five-item index measuring anti-immigrant sentiments focusing on out-group perceptions (Denmark: M = 3.43, SD = 0.83, alpha = 0.83; the Netherlands: M = 3.40, SD = 0.78, alpha = 0.82).12 The means and standard deviations of the indexes as well as the specific wording of all items can be found in the appendix.13 The specified regression models are ordinary least squares (OLS) models. The voting intention model is a logistic regression model with 'intention to vote Yes' (coded as 1) as the dependent variable.

Results

Our multivariate analyses (Table 1), in which the explanatory value of the different predictors is assessed simultaneously, shows that in both countries the strongest predictor was anti-immigration sentiments, negatively predicting support for European integration (H1). Positive evaluations of the incumbent government (H2) and optimistic economic assessments (H3) were, as hypothesized, positive, significant predictors of support. Political sophistication (H4) positively predicted support for integration and this was significant in Denmark. Gender was a negative predictor of support for European integration (significant in Denmark) and age was a significant negative predictor in the Netherlands. Education was a positive predictor in both countries. We found no significant effects of occupational status. In Denmark, postmaterialist values were negatively related to support for European integration. Finally, left ideological preference was a significant negative predictor in Denmark whereas right ideological preference was a significant negative predictor in the Netherlands.

We further estimated a model of 'yes' voting intention in a referendum on the enlargement of the EU. Using the same predictors as in our previous model, we find - by and large - a confirmation of the predictors for understanding not only opinions but also behavioural intentions. Table 2 shows the results of a logistic regression model estimating the likelihood of voting

Table 1 Regression analysis of support for European integration

Denmark The Netherlands

Std. beta Standard Std. beta Standard

coefficient error coefficient error

Gender (female) -0.08** 0.05 -0.01 0.03

Age 0.02 0.00 -0.06** 0.00

Education 0.06* 0.03 0.06** 0.02

Unemployed/blue collar -0.04 0.06 -0.04 0.03

Executive/managers 0.03 0.06 0.00 0.03

Left ideological preference -0.09** 0.07 0.00 0.03

Right ideological preference 0.04 0.06 -0.06** 0.03

Anti-immigration attitudes -0.39*** 0.06 -0.34*** 0.00

Post-materialist values -0.08** 0.03 -0.00 0.01

Political sophistication 0.18*** 0.02 0.01 0.01

Evaluation incumbent government 0.19*** 0.03 0.12*** 0.01

Economic evaluations 0.07* 0.04 0.21*** 0.02

n 1084 2040

R2 0.24 0.22

Note: OLS regression.

*** p < 0.001; ** p < 0.01; * p < 0.05.

'yes' with positive coefficients indicating a higher probability of voting 'yes'. The results show that anti-immigration sentiments were a strong negative predictor for a 'yes' vote, namely fuelling the propensity to vote 'no'. Government approval and positive economic evaluations were the most important positive predictors for a 'yes' vote. Additionally, political sophistication was a positive predictor in Denmark, and was significant only in Denmark. Individuals with right-wing political leanings differed significantly from centrists (our reference category) in their negative association with voting 'yes' in the Netherlands, and women were less likely to vote 'yes' in Denmark. Finally, age was positively associated with voting 'yes' in Denmark. Although the analysis enables us to distinguish relevant predictors for understanding a 'yes' vote, the model also allows for a comparison of the relative importance of each predictor. This is shown in the 'predicted probabilities' columns of Table 2.

Given the robustness of our findings that anti-immigration sentiments, economic evaluations and evaluation of the domestic government are

Table 2 Logistic regression analysis of 'yes' vote in referendum on EU enlargement °

Denmark The Netherlands =

Beta Standard Predicted Beta Standard Predicted

coefficient error probabilities coefficient error probabilities

Gender (female) -0.73*** 0.17 0.48 0.01 0.10 1.01

Age 0.03*** 0.01 1.03 0.01 0.01 1.00

Education 0.13 0.11 1.14 -0.11 0.09 0.89

Unemployed/blue collar -0.05 0.23 0.96 0.04 0.12 1.05

Executive/managers -0.21 0.21 0.81 -0.03 0.11 0.97

Left ideological preference -0.31 0.27 0.74 0.02 0.14 1.02

Right ideological preference -0.04 0.21 0.96 -0.42*** 0.12 0.65

Anti-immigration attitudes -1.31*** 0.13 0.27 -0.94*** 0.08 0.39

Post-materialist values -0.18 0.11 0.83 -0.05 0.05 0.95

Political sophistication 0.23*** 0.06 1.25 -0.05 0.03 0.96

Evaluation incumbent government 0.36*** 0.10 1.43 0.16** 0.06 1.18

Economic evaluations 0.34 * 0.14 1.40 0.61*** 0.06 1.84

n 846 2041

Correctly classified 76.5 68.2

-2 log likelihood 855.05 2457.22

Nagelkerke's pseudo R2 0.30 0.22

Note: Logistic regression.

*** p < 0.001; ** p < 0.01; * p < 0.05.

consistently the strongest predictors, we model the effects of each of these key predictors on the likelihood of voting 'yes'.

Figures 1 and 2 are illustrations of the estimates provided in Table 2 and demonstrate the likelihood of voting 'yes' by a one-unit increase in the level of government approval, economic evaluations and anti-immigration

........Economic evaluations

-----Government approval

-Fear of immigration

Figure 1 The Netherlands: Likelihood of voting 'yes' when changing the level of (a) economic evaluations, (b) government approval, or (c) fear of immigration.

........ Economic evaluations

-----Government approval

-Fear of immigration

Figure 2 Denmark: Likelihood of voting 'yes' when changing the level of (a) economic evaluations, (b) government approval, or (c) fear of immigration.

sentiments, respectively. The predicted probabilities of voting 'yes' are calculated by holding all variables constant at their mode (dummy variables) or their mean (remaining variables). Respondents expressing scepticism and reluctance towards immigration were more likely to vote 'no' than were voters generally in favour of or less hesitant towards immigration. The slopes in both countries are relatively steep, indicating that any increase in the level of anti-immigration sentiments is associated with a considerable increase in the likelihood of voting 'no'. In the case of government approval, respondents expressing more satisfaction with the government were more likely to vote 'yes'. This slope is steeper in Denmark than in the Netherlands, indicating that an increase in government approval is related to a modest increase in the likelihood of voting 'yes' in the Netherlands and a considerable increase in the likelihood of voting 'yes' in Denmark. In the case of economic evaluations, this pattern is reversed. In both countries, respondents expressing more optimism about the economy were more likely to vote 'yes'. However, this slope is much steeper in the Netherlands than in Denmark. This suggests that an increase in economic evaluations is related to a modest increase in the likelihood of voting 'yes' in Denmark and a considerable increase in the likelihood of voting 'yes' in the Netherlands. These illustrations stress the importance and relevance of all three predictors.

Looking at Figures 1 and 2, we see that in all cases (except economic evaluations in Denmark) the slope crosses the .50 line, which is the decisive point for a 'yes' or a 'no' vote. This means that each of the factors can lead to significant changes in the propensity to vote in favour of or against a proposal in a referendum. It is particularly important to note that only a slight increase in the level of anti-immigration sentiments decreases the likelihood of voting 'yes'.

Discussion

In our analysis of the forces that drive popular support for European integration, three explanations emerge as particularly powerful: anti-immigration sentiments, economic evaluations, and support for the domestic government. Extending McLaren's (2002) work, we thus explicate public attitudes towards immigration as an important predictor of public support for European integration. This finding supports our assumption that people's readiness to negatively categorize out-groups, measured by attitudes towards immigrants, is influential in support for European integration. Economic considerations have been articulated strongly in the literature on public support for the EU (e.g. Gabel, 1998a; Gabel and Palmer, 1995). We find partial evidence of this

argument in the multivariate analysis. The liberalization of labour markets in the EU is likely to favour individuals with good job skills and high levels of education (Gabel and Palmer, 1995). We find this relationship confirmed in our analysis, but education is a stronger predictor than occupational status, which shows no significant coefficient. Our findings corroborate Gabel (1998a), who found economic evaluations to be significantly linked to the level of support for EU membership. The evaluation of domestic governments as a predictor of support for European integration has been articulated particularly in relation to referendums on European topics (Franklin et al., 1995). The relevance of domestic political considerations for general EU support, however, has been confirmed in most studies of diffuse support outside the electoral situation (e.g. Anderson, 1998; Gabel, 1998a; McLaren, 2002). We also find that citizens draw on their perceptions of the performance of the national government when expressing support for or opposition to European inte-gration.14

In this study, we went beyond explaining attitudes to assess voting intention as well. We modelled voting intention in a referendum on the enlargement of the EU and illustrated how differences in economic evaluations, government support and anti-immigration sentiments can influence the likelihood of voting 'yes' in such a referendum. We emphasize that we do not draw any substantive conclusions about the level of support for enlargement (as expressed in the share of 'yes' votes), but we stress the importance of the underlying dynamics and predictors of the vote. Given the increase in the use of national referendums on European issues, we can speculate about EU-related referendums, including the wave of national referendums on the EU Constitution.

With the continuing social and political importance of immigration issues and in an economic climate characterized by recession or economic stagnation, the conditions for rejecting proposals on further European integration are present for European citizens. Citizens may differentiate between supporting enlargement (the measure used in our study) and endorsing the Constitution, for example. Whereas the former applies to the notion of wider integration, the latter concerns deeper integration. However, our analyses clearly show that the same mechanisms are at play when understanding both diffuse support for European integration and specific support for enlargement.

Sniderman et al.'s (2000) 'right-shock model' predicts that when a society is exposed to an external shock, such as economic recession or increasing levels of immigration, anti-immigration sentiments will flourish. This view is supported by studies explaining the rise of right-wing parties in Western Europe by high rates of unemployment (Jackman and Volpert, 1996), large

numbers of non-Western residents in a country (Lubbers et al., 2002), or large numbers of asylum seekers (Lubbers and Scheepers, 2001).15 Recent studies, however, find only unemployment to be related to support for right-wing parties in interaction with the level of immigration (Golder, 2003a, 2003b) or even demonstrate a negative relationship between unemployment and radical right support (Knigge, 1998). Anyway, it is shown that contextual factors matter, and if Sniderman et al.'s (2000) model travels beyond the case of Italy there is reason to worry about the level of public support for European integration. The almost Europe-wide economic stagnation in the years 2000-3 may generate higher levels of anti-immigration sentiments. Such attitudes are, as demonstrated in this study, of key importance for understanding public opinion about European integration and they are likely to affect public support for European integration negatively.16

With our study, we emphasize the implications of understanding public opinion about European integration. Of course an investigation with more variation at the contextual level (in terms of differences in economic conditions and popularity of anti-immigration sentiments) is desirable. For instance, Sánchez-Cuenca (2000) demonstrated the contingent importance of domestic political considerations as benchmarks in understanding support for the EU. He found that support is higher in countries that suffer from corruption and have less developed welfare states. In such societies, the EU was viewed positively in comparison with the nation-state, whereas it is perceived as a threat in other countries. We demonstrated our model in two comparable societies with similar economic situations and experience of popular anti-immigration parties. We hope this study will spark broader comparative studies that will provide more variation and power to the explanations. We stress the importance of public support as an inherent requirement of the legitimacy of European integration. This support may be indirect and expressed in surveys but, on significant occasions, the opinions of citizens matter in very explicit and direct ways for political decision-making, such as in the case of referendums on European integration issues. Considering the importance of anti-immigration sentiments and of economic evaluations in a referendum, any government calling a referendum and wanting to see its proposal endorsed must be very popular to compensate for the negative impact of economic pessimism and anti-immigration sentiments.

Appendix: Overview of independent variables

Gender: female = 1, male = 0. Age: in years.

Education: recoded into four categories, comparable across the two countries: 1 - primary school; 2 - high school or equivalent (about 13 years of training); 3 - BA or three years' vocational training or equivalent (16 years); and 4 -Master's or postgraduate training (19+ years).

Low occupational status: recoded dummy variable. Occupation was recoded based on Eurobarometer occupational classifications (see e.g. Eurobarometer 60). Unemployed and blue-collar workers were recoded as 1, otherwise 0.

High occupational status: recoded dummy variable. Occupation was recoded based on Eurobarometer occupational classifications (see e.g., Eurobarometer 58). Executive and managerial occupations were recoded as 1, otherwise 0.

Left political ideology: self-placement on left-right scale, where 1 = left and 10 = right; 1-3 coded as 1, otherwise 0.

Right political ideology: self-placement on left-right scale where 1 = left and 10 = right; 7-10 coded as 1, otherwise 0.

Political sophistication: a combined measure of political knowledge and political interest. Five questions tapped political knowledge:

• 'What is the number of Commissioners in the EU Commission?' - open-ended, correct answer (20 Commissioners) coded as 1, otherwise 0

• 'What is the name of the current President of the European Commission?' - open-ended, correct answer (Prodi) coded as 1, otherwise 0

• 'What is the name of the Danish or Dutch Commissioner?' - open-ended, coded as 1 or 0, correct answer (Nielson or Bolkestein)

• 'Which country currently holds the Presidency of the EU?' - correct answer (Denmark) coded as 1, otherwise 0

• 'What is the number of countries seeking membership of the EU?' -correct answers (10, 12 and 13) coded as 1, otherwise 0.

Political interest was a single item ranging from 1 ('no political interest') to 4 ('high tapped political interest'). The sophistication index ranges from 1 to 9. Denmark: M = 5.02, SD = 1.60; the Netherlands: M = 3.54, SD = 1.64.

Anti-immigrant sentiments: five-item index measuring anti-immigrant sentiment: 1 - immigration is good for the labour market; 2 -immigrants cause

problems in the schools that their children attend; 3 - immigrants enrich Danish or Dutch culture; 4 - members of immigrant groups misuse Danish or Dutch social welfare; and 5 - their religion is a threat to our way of life. The items were recoded when appropriate to form a scale of anti-immigrant sentiments. Denmark: M = 3.43, SD = 0.83, alpha = 0.83; the Netherlands: M = 3.40, SD = 0.78, alpha = 0.82.

Post-materialism: Two questions in which the respondent was asked to rate the most important and the second most important task of the government. The options were: 1 - maintain law and order (materialistic); 2 - give citizens a greater say in important government decisions (post-materialism); 3 - control prices (materialism); 4 - protect freedom of expression measures (post-materialism). The ratings were recoded as: 1 (two materialism choices), 2 (one materialism and one post-materialism choice), 3 (one post-materialism and one materialism choice) or 4 (two post-materialism choices). Denmark: M = 2.12, SD = 0.82; the Netherlands: M = 2.29, SD = 0.98.

Evaluation of domestic government: a scaled measure ranging from 1 to 5, where 1 = very bad, 3 = neither good nor bad, and 5 = very good. Denmark: M = 3.65, SD = 1.08; the Netherlands: M = 3.25, SD = 1.06.

Economic evaluation: a scaled measure ranging from 1 to 5 of economic expectations in the coming 12 months, where 1 = a lot worse, 3 = neither worse nor better, and 5 = a lot better. Denmark: M = 3.08, SD = 0.63; the Netherlands: M = 3.07, SD = 0.82.

Notes -

The study was made possible through a research grant from the Danish Social Science Research Council to Claes de Vreese. He thanks the Center for European Studies at Harvard University and the Dutch Science Foundation (NWO) for providing a stimulating work environment when the article was being written. The data used for this study will become available through the Danish Data Archive (Dansk Data Arkiv). The authors thank Susan Banducci, Mark Franklin, Stefaan Walgrave and the three anonymous reviewers for their helpful comments.

1 Later, this thesis was modified to apply in particular to referendums on issues that are of low salience to the electorate (Franklin, 2002) and in particular in years in which European elections or referendums are held (Ray, 2003).

2 In addition, Carey (2002) tests the impact of three different conceptualizations of national identity: national pride, territorial identification, and perceived cultural threat. He shows that national pride most strongly and negatively affects support for EU integration. However, territorial identification and perceived cultural threat also are significantly related to EU

support. These findings somewhat counter Christin and Trechsel's (2002) results. They argue that 'national identity' is only weakly related to support for Swiss EU membership, whereas 'perceived threat to national interests' is strongly related to support. Marks and Hooghe (2003) also test the impact of different concepts of national identity and perceived threat. They find that feelings of exclusive national identity and perceived cultural threat are negatively related to EU support whereas attachment to the nation and embracing multiculturalism are positive predictors of support.

3 Although these parties are not easily comparable (see e.g. Fennema, 1997) and their supporters might have different motivations for voting for them (see Billiet and de Witte, 1995; Lubbers et al., 2002; Van der Brug and Fennema, 2003, for competing explanations), they all contributed to bringing the issues of immigration and the integration of foreigners to the top of the political and public agenda.

4 Eurobarometer 58 showed that 61% of the population in Denmark and 69% in the Netherlands support their country's membership of the EU and 69% and 64%, respectively, believe that their country has benefited from membership.

5 Immigration-related issues have been high on the public and political agenda for the past few years in Denmark and the Netherlands (see e.g. Andersen et al., 1999, for Denmark and Sniderman et al., 2004, for the Netherlands). Dansk Folkeparti (DF) in Denmark and Lijst Pim Fortuyn (LPF) in the Netherlands have both contributed to government, DF by constituting the parliamentary majority for the Danish Liberal-Conservative government (2001-) and LPF by forming a short-lived three-party coalition with CDA and VVD in 2002.

6 The specific fieldwork days were 21-28 November 2002 in Denmark, and 19-26 November 2002 in the Netherlands.

7 In Denmark, the sample was drawn from the GfK Danmark database. A nationally representative sample of 1807 Danish adult (age 15+) individuals was invited to participate in the study. The questionnaire was a postal self-administered paper and pencil questionnaire. The response rate is not unusual for survey research in Scandinavia, where turnout is also comparatively high (Granberg and Holmberg, 1991). In the Netherlands, the sample was drawn from the ITM International database, which has more than 55,000 respondents. A nationally representative sample of 5321 Dutch adults (age 15+) was invited to participate in the study. The questionnaire was Web administered. Making use of a similar lay-out for the questionnaire in the two countries, potential confounds owing to question and response category layout were taken into account (Dillman, 2000). To assess the quality of our data, we included the standard Eurobarometer 'support for country's EU membership' question in our survey: 60% of our respondents in Denmark and 66% in the Netherlands reported considering the membership of their country in the EU to be a good thing. This compares with 61% and 69%, respectively, in the autumn Eurobarometer 58, which was carried out in October 2002.

8 A factor analysis (PCA with varimax rotation) confirmed the one-dimensional nature, with all item loadings strongly on one factor, with an Eigen value of 2.95, explaining 59% of the variance.

9 This question measures a behavioural intention (which may differ from actual behaviour) and there is a tendency to over-report participation in intention measures (e.g. Belli et al., 1999). However, although vote turnout is often inflated, there is no reason to expect a structural bias in this voting choice measure.

10 Respondents' reported level of completed education was recoded because of differences in the educational systems; see the appendix.

11 We included these two groups as dummy variables in the analysis. A more elaborate model including unemployed, manual workers, managers, housewives/husbands, white-collar workers, and executives separately (see McLaren, 2002) did not yield different results. We include the recoded categories for presentational reasons.

12 Unlike in the Eurobarometer, these questions were not asked after a question probing the respondent for whether s/he feels that s/he belongs to one of the majority or minority groups in a country. This potentially biases Eurobarometer respondents towards thinking about in-groups and out-groups prior to answering questions about immigration. Our measures resemble McLaren's (2002) measures of cultural threat but specifically tap attitudes towards immigrants and not just threat perceptions.

13 Unfortunately, our data do not allow us to test national identity-based explanations for EU support. We would encourage studies that disentangle the relationship between different identity approaches and related indicators of anti-immigration sentiments.

14 The different predictive power of government evaluation and economic evaluations might be attributed to the political situation in the two countries in autumn 2002. Denmark was governed by a stable Liberal-Conservative coalition government, which was a likely point of evaluation. The Netherlands, however, was governed by the three-party coalition that emerged out of the May 2002 elections in the aftermath of the assassination of Pim Fortuyn. In November 2002, the government had already announced new general elections for January 2003 and was therefore not perceived to be a powerful acting entity. This de-emphasized the importance of government evaluations as a predictor of support and highlighted economic evaluations. Moreover, Danes are used to having proposals put to them by their government on European matters whereas the Dutch are not. These contextual explanations stress the importance of including measures of both economic and government evaluation, because these are related (as demonstrated by Anderson, 1998; Clarke et al., 2000).

15 Since economic concerns are another exogenous variable in our full model of support for European integration, potential problems of multicollinearity arise. The bivariate correlation between economic evaluations and immigration attitudes is .34, which is significant but does not cause concern about multicollinearity. Furthermore, we know that sociodemographics, personality traits and authoritarian values (Sniderman et al., 2000), as well as exposure to mass media content (Vergeer, 2000), can cause hostility towards immigrants. Future research needs to disentangle the relationship between national identity and immigration attitudes as predictors of EU support. This study represents a first step in emphasizing the importance of considering immigration attitudes.

16 Of course, a bad economic situation as such also affects economic evaluations, which in turn have a negative impact on support for the EU.

References

Andersen, Joergen, Ole Borre, Joergen Goul-Andersen and Hans Joegen Nielsen (1999) Vaelgere med omtake. En analyse af folketingsvalget 1998 [Thoughtful Voters: An Analysis of the 1998 General Elections]. Aarhus: Systime.

Anderson, Chris J. (1998) 'When in Doubt Use Proxies: Attitudes to Domestic Politics and Support for the EU', Comparative Political Studies 31(5): 569-601.

Anderson, Chris J. and Michael S. Reichert (1996) 'Economic Benefits and Support for Membership in the European Union: A Cross-National Analysis', Journal of Public Policy 15(3): 231-49.

Belli, Robert F., Michael W. Traugott, Margaret Young and Katherine A. McGonagle (1999) 'Reducing Vote Over-Reporting in Surveys: Social Desirability, Memory Failure and Source Monitoring', Public Opinion Quarterly 63(1): 90-108.

Billiet, Jean and Hans de Witte (1995) 'Attitudinal Dispositions to Vote for a "New" Extreme Right-Wing Party: The Case of VB', European Journal of Political Research 27(2): 181-202.

Bosch, August and Ken Newton (1995) 'Economic Calculus or Familiarity Breeds Content?', in Oscar Niedermeyer and Richard Sinnott (eds) Public Opinion and Internationalized Governance, pp. 73-104. Oxford: Oxford University Press.

Brown, Rupert J. (2000) 'Social Identity Theory: Past Achievements, Current Problems and Future Challenges', European Journal of Social Psychology 30(6): 745-78.

Caldeira, Gregory and James Gibson (1995) 'The Legitimacy of the Court of Justice in the European Union: Models of Institutional Support', American Political Science Review 89(2): 356-76.

Carey, Sean (2002) 'Undivided Loyalties: Is National Identity an Obstacle to European Integration?', European Union Politics 3(4): 387-413.

Christin, Thomas and Alexander H. Trechsel (2002) 'Joining the EU? Explaining Public Opinion in Switzerland', European Union Politics 3(4): 415-43.

Clarke, Harold D., Allan Kornberg and Peter Wearing (2000) A Polity on the Edge: Canada and the Politics of Fragmentation. Peterborough, Ontario: Broadview Press.

De Vreese, Claes H. and Holli A. Semetko (2004a) Political Campaigning in Referendums: Framing the Referendum Issue. London: Routledge.

De Vreese, Claes H. and Holli A. Semetko (2004b) 'News Matters: Influences on the Vote in a Referendum Campaign', European Journal of Political Research 43(5): 699-722.

Dillman, Don A. (2000) Mail and Internet Surveys: The Tailored Design Method. New York: John Wiley.

Eurobarometer 58 (2003) Eurobarometer: Public Opinion in the European Union, Report No. 58. Brussels, Belgium: Directorate-General X.

Fennema, Meindert (1997) 'Some Conceptual Issues and Problems in the

Comparison of Anti-Immigration Parties in Western Europe', Party Politics 3(4): 473-92.

Franklin, Mark (2002) 'Learning from the Danish Case: A Comment on Palle Svensson's Critique of the "Franklin Thesis"', European Journal of Political Research 41(6): 751-7.

Franklin, Mark, Michael Marsh and Christopher Wlezien (1994) 'Attitudes Toward Europe and Referendum Votes: A Response to Siune and Svensson', Electoral Studies 13(2): 117-21.

Franklin, Mark, Cees van der Eijk and Michael Marsh (1995) 'Referendum Outcome and Trust in Government: Public Support for Europe in the Wake of Maastricht', West European Politics 18(3): 101-7.

Gabel, Matthew (1998a) 'Public Support for European Integration: An Empirical Test of Five Theories', Journal of Politics 60(2): 333-54.

Gabel, Matthew (1998b) Interest and Integration: Market Liberalization, Public Opinion, and European Union. Ann Arbor, MI: University of Michigan Press.

Gabel, Matthew and Harvey Palmer (1995) 'Understanding Variation in Public Support for European Integration', European Journal of Political Research 27(1): 3-19.

Golder, Matt (2003a) 'Electoral Institutions, Unemployment and Extreme Right Parties: A Correction', British Journal of Political Science 33(3): 525-34.

Golder, Matt (2003b) 'Explaining Variation in the Success of Extreme Right Parties in Western Europe', Comparative Political Studies 36(4): 432-66.

Granberg, Donald and Sören Holmberg (1991) 'Self-reported Turnout and Voter Validation', American Journal of Political Science 35(2): 448-59.

Hooghe, Lisbeth (2003) 'Europe Divided? Elite vs. Public Opinion on European Integration', European Union Politics 4(3): 281-304.

Hug, Simon (2003) Voices of Europe: Citizens, Referendums and European Integration. Oxford: Rowman & Littlefield.

Inglehart, Ronald (1970) 'Cognitive Mobilization and European Identity', Comparative Politics 3(1): 45-70.

Inglehart, Ronald (1977) The Silent Revolution. Princeton, NJ: Princeton University Press.

Inglehart, Ronald (1990) Culture Shift in Advanced Industrial Societies. Princeton, NJ: Princeton University Press.

Jackman, Robert W. and Karin Volpert (1996) 'Conditions Favouring Parties of the Extreme Right in Western Europe', British Journal of Political Science 26(4): 501-21.

Karp, Jeffrey A., Susan A. Banducci and Shaun Bowler (2003) 'To Know It Is to Love It: Satisfaction with Democracy in the European Union', Comparative Political Studies 36(3): 271-92.

Knigge, Pia (1998) 'The Ecological Correlates of Right-Wing Extremism in Western Europe', European Journal of Political Research 34(2): 249-79.

Kritzinger, Silvia (2003) 'The Influence of the Nation-State on Individual Support for the European Union', European Union Politics 4(2): 219-41.

Lubbers, Marcel and Peer Scheepers (2001) 'Explaining the Trend in Extreme Right-Wing Voting: Germany 1989-1998', European Sociological Review 17(4): 431-49.

Lubbers, Marcel, Mérove Gijsberts and Peer Scheepers (2002) 'Extreme Right-Wing Voting in Western Europe', European Journal of Political Research 41(3): 345-78.

McLaren, Lauren M. (2002) 'Public Support for the European Union: Cost/Benefit Analysis or Perceived Cultural Threat?', Journal of Politics 64(2): 551-66.

Marks, Gary and Lisbeth Hooghe (2003) 'National Identity and Support for European Integration', Discussion Paper SP IV 2003-202. Berlin: Wissenschaftszentrum Berlin für Sozialforschung (WZB).

Nelsen, Brent F. and James L. Guth (2000) 'Exploring the Gender Gap: Women, Men, and Public Attitudes toward European Integration', European Union Politics 1(3): 267-91.

OECD (2002) 'Report on National GDP', URL (consulted March 2004): http://www.oecd.org/dataoecd/30/40/29867116.xls.

Pepermans, Roland and Gino Verleye (1998) 'A Unified Europe? How Euro-Attitudes Relate to Psychological Differences between Countries', Journal of Economic Psychology 19(6): 681-99.

Ray, Leonard (2003) 'Reconsidering the Link between Incumbent Support and Pro-EU Opinion', European Union Politics 4(3): 259-79.

Sánchez-Cuenca, Ignacio (2000) 'The Political Basis for Support for European Integration', European Union Politics 1(2): 147-72.

Scharpf, Fritz W. (1970) Demokratietheorie zwischen Utopie und Anpassung [Theory of Democracy between Utopia and Adaptation]. Konstanz: Konstanz Universitätsverlag.

Scharpf, Fritz W. (1997) 'Economic Integration, Democracy and the Welfare State', Journal of European Public Policy 4(1): 18-36.

Siune, Karen and Palle Svensson (1993) 'The Danes and the Maastricht Treaty: The Danish EC Referendum of June 1992', Electoral Studies 12(2): 99-111.

Sniderman, Paul M., Luuk Hagendoorn and Markus Prior (2004) 'Predispositional Factors and Situational Triggers: Exclusionary Reactions to Immigrant Minorities', American Political Science Review 98(1): 35-50.

Sniderman, Paul M., Pierangelo Peri, Rui J. P. de Figueiredo, Jr, and Thomas Piazza (2000) The Outsider: Prejudice and Politics in Italy. Princeton, NJ: Princeton University Press.

Svensson, Palle (2002) 'Five Danish Referendums on the European Community and European Union: A Critical Assessment of the Franklin Thesis', European Journal of Political Research 41(6): 733-50.

Tajfel, Henri (1981) Human Groups and Social Categories: Studies in Social Psychology. Cambridge: Cambridge University Press.

Tajfel, Henri and John C. Turner (1979) 'An Integrative Theory of Intergroup Conflict', in William G. Austin and Stephen Worchel (eds) Psychology of Intergroup Relations, pp. 7-24. Chicago: Nelson Hall.

Van der Brug, Wouter and Meindert Fennema (2003) 'Protest or Mainstream? How the European Anti-Immigration Parties Developed into Two Separate Groups by 1999', European Journal of Political Research 42(1): 55-76.

Van der Brug, Wouter, Meindert Fennema and Jean Tillie (2000) 'Anti-Immigrant Parties in Europe: Ideological or Protest Vote', European Journal of Political Research 37(1): 77-102.

Van der Eijk, Cees and Mark Franklin (1996) Choosing Europe? The European Electorate and National Politics in the Face of Union. Ann Arbor: University of Michigan Press.

Vergeer, Maurice (2000) 'Een gekleurde blik op de wereld. Een studie naar de relatie tussen blootstelling aan media en opvattingen over ethnische minderheden' [A Coloured View of the World. A Study of the Relation between Media Exposure and Attitudes toward Ethnic Minorities], unpublished doctoral dissertation, Katholieke Universiteit Nijmegen, The Netherlands.

About the authors

Claes H. de Vreese is an associate professor at the Amsterdam School of Communications Research, University of Amsterdam, ASCoR, Kloveniersburgwal 48, 1012 CX Amsterdam, The Netherlands. Fax: +31 20 525 3681 E-mail: C.H.deVreese@uva.nl

Hajo G. Boomgaarden is a PhD candidate at the Amsterdam School of Communications Research, University of Amsterdam, ASCoR, Kloveniersburgwal 48, 1012 CX Amsterdam, The Netherlands. Fax: +31 20 525 3681 E-mail: H.Boomgaarden@uva.nl