The impact of identity and economic cues on citizens’ EU support: An experimental study on the effects of party communication in the run-up to the 2009 European Parliament elections

Abstract

In this article we seek to understand whether national parties have an impact on citizens’ EU support by publicly cueing Europe as a risk to or as an opportunity for the economy or identity. In order to answer this question, we have conducted a cross-country survey experiment (covering Austria, the Czech Republic, Germany, the United Kingdom, the Netherlands, Poland, Portugal, Spain and Sweden) relying on real-world stimuli from party campaign communication in the run-up to the 2009 European Parliament elections. By introducing this new methodology to cueing research we show substantial evidence for cueing effects even when thoroughly controlling for nuisance variables drawn from EU research as well as country contexts. We find support for the general cueing hypothesis in experimental groups that were exposed to negative economic messages while in two other groups partisanship works as a relevant moderator of the effects of persuasive messages. These findings are explained by distinguishing between consensual and conflicting issues and show in what circumstances campaign messages might reach beyond the particular partisan base.

Keywords

Campaign effects electoral campaign research European Parliament elections EU attitudes experimental research partisan cueing

Introduction

Citizens’ attitudes towards European integration are at the core of political and research debates alike. However, whether citizens’ attitudes have turned into ‘reluctant acceptance’ (Mittag and Wessels, 2003: 47), have become ambivalent (Steenbergen and De Vries, in press), or even constitute a ‘constraining dissensus’ for elites (Hooghe and Marks, 2005: 426) remains an open question. Also, the question of what drives citizens to support or oppose the European Union (EU) has triggered extensive research, but no scholarly consensus has yet emerged (Boomgaarden et al., 2011; Hooghe and Marks, 2005). Hooghe and Marks (2005: 420ff.) have grouped different approaches of explanations into three ‘families’:¹ economic considerations, community/identity-related considerations, and political cueing. In the economic domain it is argued that EU integration creates new forms of competition that lead to societal winners and losers (Kriesi et al., 2006). Those who profit from EU integration tend to support it, whereas the losers prefer demarcation (for example, Anderson and Reichert, 1996; Boomgaarden et al., 2011; Gabel, 1998). Individual predictors of EU attitudes such as education, occupation or income are summarized in this group along with collective factors such as the type of welfare state, objective economic measures and subjective evaluations of the national and EU economy. Community- or identity-related research refers to the psychology of group membership. From this perspective, EU support is strongly tied to identity constructions (for example, Carey, 2002; McLaren, 2002). The relationship between national identity and EU identity is twofold: national identity and European identity may reinforce each other, thereby creating multiple identities (for example, Boomgaarden et al., 2011; Bruter, 2003; Cidrin and Sides, 2004). Yet national identity – particularly if it is exclusive – may well lead to opposition to EU integration (Carey, 2002; Hooghe and Marks, 2005). Specific predictors from this group are, for example, national pride, perceived cultural threat, anti-immigration attitudes (see Boomgaarden et al., 2011) and territorial attachment.

These first two schools focus on citizens’ internal system of perceptions and predispositions. However, because opinions are a ‘marriage of information and predispositions’ (Zaller, 1992: 6, emphasis added), a third family centres on the impact of external information on EU attitudes.² Cueing theory argues that political actors, that is, political parties, elites and mass media, influence public opinion by tying together, for instance, persuasive messages to political predispositions (see, for example, Campbell et al., 1980; Converse, 2006/1964; Kam, 2005; Mondak, 1993; Zaller, 1992). Cues ‘provide citizens with cognitive shortcuts that help them decide what is in their interest’ (De Vries and Edwards, 2009: 8). Cueing is all the more important regarding the issue of EU integration because Europe is ‘rarely foremost in citizens’ minds’ (Hooghe and Marks, 2005: 420), and therefore opinions are particularly susceptible to construction (Hooghe and Marks, 2008: 13; also see Zaller, 1992). Political ideologies and values (for example, post-materialism or left/right self-placement), party and elite cues (for example, partisan support) and other political cues (for example, system or government support) belong to this category. Recent research has empirically validated the relevance of elite cues for citizens’ attitudes towards Europe (for example, De Vries and Edwards, 2009; Gabel and Scheve, 2007; Hooghe, 2007; Hooghe and Marks, 2005; Steenbergen et al., 2007). The increasing ambivalence of EU attitudes might boost the effectiveness of cues.

Based on national expert surveys or manifesto data, for example, parties’ stances on EU integration, intra-party dissent, elite division, the salience of EU issues, the existence of left- and right-wing Eurosceptics or even macroeconomic system factors (De Vries and Edwards, 2009; Down and Wilson, 2010; Gabel and Scheve, 2007; Garry and Tilley, 2009; Hooghe and Marks, 2005; Ray, 2003; Steenbergen et al., 2007), cues have primarily been conceptualized as country contexts of the party system. In this logic, structural features of the domestic political systems have been used to predict which cues are most likely to be publicly communicated and thus affect citizens’ EU attitudes. Although this proxy approach is valuable, its applicability is limited when we seek to link cues more directly to actual attitudes. In our article we want to shed light on the causal mechanism underlying this elite-cueing process. In order to accomplish this, we add to the existing cueing literature in two significant ways. First, instead of relying on a proxy approach we actually study the process of cue transmission. Second, we will go beyond analysing the level of Euroscepticism in a country and instead focus more closely on the content of cues and the resulting effects on EU attitudes (for first steps, see De Vries and Edwards, 2009).

Keeping in mind that support for and opposition to EU integration might be framed in terms of identity or economy, we ask the following research question: do national parties have an impact on citizens’ EU support by publicly cueing Europe as a risk or an opportunity for the economy or identity? To answer this question we apply a different methodological approach from traditional cueing research: we conduct a cross-country survey experiment relying on real-world stimuli from party campaign communication in the run-up to the 2009 EP elections systematically selected by means of content analysis. In the experiment, participants were randomly exposed to four different types of elite cues. In the following we will identify research lacunas and derive research hypotheses. Then, in a second step, we will present our data, operationalizations and the experimental design, followed by the presentation of results of mainly multiple regression analyses. Finally, we will discuss the implications of our findings.

Cueing and EU attitudes

Who is cueing whom has been the main question of many recent contributions to the EU literature. From a bottom-up perspective, mass publics might influence domestic parties’ positions on the conflictual process of EU integration (Van der Eijk and Franklin, 2004). Consequently, the notion of strategic party competition allows one to understand the emergence of Eurosceptic parties and discourses (for example, Hooghe and Marks, 2008; Kriesi, 2008). Empirically, Carrubba (2001) has established this electoral connection even for the pre-Maastricht period.

Furthermore, research has also accumulated evidence for top-down cueing, that is, elites shaping public opinion on EU integration (for example, De Vries and Edwards, 2009; Feld and Wildgen, 1976; Gabel and Scheve, 2007; Hooghe and Marks, 2005; Ray, 2003). One of the more recent studies on mass–elite linkages in Europe lends support to both bottom-up and top-down processes of cueing: party elites shape public opinion but at the same time they are influenced by these public opinions under specific conditions (Steenbergen et al., 2007).

Cueing research so far has helped us understand the direction of elite–mass relations in the process of EU integration and to identify conditions under which such influence occurs. Most models of cueing adhere to a very specific mass–elite relationship: it is the parties and their respective supporters who influence each other. Only very few studies postulate that cueing effects might reach beyond the own party base (De Vries and Edwards, 2009; Down and Wilson, 2010). We follow the latter studies since we do not restrict cueing processes per se to a specific segment of the population. Moreover, we do not seek to test bottom-up and top-down cueing against each other. Instead, we concentrate our efforts on top-down cueing because we want to understand the effect of specific cues on citizens’ EU attitudes.

We focus on two blind-spots of cueing theory. First, hardly any researchers have asked how elite cues actually reach the citizens (for critiques, see Carey and Burton, 2004; Maier and Rittberger, 2008). Research on elite cueing takes structural context variables of domestic party systems as proxies for the cues that dominate within any given country. The implicit assumption of such research is that structural features of party systems determine public political communication and that the cues reach out to all citizens or all partisan supporters. Neither of these assumptions is unproblematic. First, exposure of citizens/partisans to cues is a prerequisite for their effectiveness. Second, research has already shown that the content of public political communication cannot simply be determined by the structural features of party systems (Adam and Maier, 2011; Adam et al., in press; Ellinas, 2010; Szczerbiak and Taggart, 2003).

If we are to shed light on the process of cue transmission, we need to specify the channels through which cues reach citizens and the methodological approaches to measuring such processes. Party manifestos are a very popular source when analysing political parties’ persuasive communication. The pros and cons of using manifesto data in order to describe parties’ policy stances have been discussed in a large number of articles (for example, Franzmann and Kaiser, 2006; Gabel and Huber, 2000; for a summary, see Dinas and Gemenis, 2010). However, the crucial reason we did not choose party manifestos as the basis for our study is that voters hardly ever read these strategic papers (see Dinas and Gemenis, 2010). Two alternative channels for the transmission of elite cues have far bigger audiences than party manifestos. First, parties can reach citizens directly, for example with party advertisements (for evidence regarding the effects of this channel, see, for example, Huber and Arceneaux, 2007; Lau and Pomper, 2002) or through their own Internet web pages. Second, elite communication is transferred to the masses by the media (for evidence regarding the effects of this channel, see Druckman, 2001; Iyengar and Kinder, 1987; Zaller, 1992). For information on Europe, citizens still regard mass media as their most important source of information (De Vreese et al., 2007). Mass media, however, do not mirror political cue-giving. Instead, depending on their logic, they decide not only which cues become prominent but also how they are framed.

When studying these two channels of cue transmission, two methodological approaches are applicable: experiments and real-world studies that combine information on cues with survey data. Both of these approaches have their strengths and shortcomings. Real-world studies that connect elite cues with citizens’ surveys have the advantage that one can observe processes over time and examine how the accumulation of conflict between cues affects opinion. In such studies, elite cues are indirectly analysed – by looking either at the structural features of party systems (classical cueing literature, for example, De Vries and Edwards, 2009; Feld and Wildgen, 1976; Gabel and Scheve, 2007; Hooghe and Marks, 2005; Ray, 2003) or at the content of mass media (for example, Dalton and Duval, 1986; De Vreese and Boomgaarden, 2006; Schuck and De Vreese, 2009). Among these studies, earlier ones use aggregated data, but they rarely ask whether cues actually reach citizens and therefore they neglect the process of cue transmission. The latter studies (for example, De Vreese and Boomgaarden, 2006; Schuck and De Vreese, 2009) ask about individual exposure to cues. Experimental research has the advantage of establishing causal links between specific cues and opinion formation by controlling the process of cue transmission very strictly. However, by means of experimental research, studies on the formation of opinions on Europe are still the exception. Only few studies exist that experimentally connect media content with opinion formation (for example, De Vreese, 2004; Maier and Rittberger, 2008; Schuck and De Vreese, 2006, 2009). Experimental studies that link party cueing with citizens’ opinion formation are the core concern of this article. We would name Slothuus and De Vreese (2010) and Kumlin (2011) as relevant here. However, their studies focus more strongly on the concept of party framing than on cueing.³

This article studies the impact of elite cues (not transformed by the media and not approximated by country contexts) on citizens’ attitudes while controlling the process of cue transmission in an experimental setting. Hypothesis 1 summarizes the basic assumption of cueing research. It claims that publicly communicated party cues influence citizens’ support for the EU and that citizens accept the advocated cues. As discussed above, most cueing research restricts the influence of party cues to party supporters. Referring to the work of Zaller (1992), party supporters’ predispositions are assumed to increase the likelihood of them accepting cues advocated by their preferred party (see also Kumlin, 2011; Slothuus and De Vreese, 2010). Hypothesis 2, which we label the partisan cueing hypothesis, claims that parties are (especially) successful in making their partisan supporters follow them, whereas non-partisans are less affected. Whether this restricted-effects paradigm is valid needs to be tested empirically.

The second blind-spot of cueing theory refers to the content of cues and the associated effects. Because we assume that the content of cues matters, we study effects for different types of cues. Here we distinguish between two dimensions of how to identify the relevant cues. First, to date most research on cueing effects primarily takes into account systemic proxies for the level and distribution of EU support and opposition among domestic parties. We follow this line of reasoning when distinguishing between cues on the basis of their EU evaluation: a positive cue presents the EU as an opportunity for the future and describes it with positive characteristics whereas a negative cue does the opposite (Maier and Rittberger, 2008; Schuck and De Vreese, 2006). Research has shown that such evaluative content of cues has an impact on political attitudes (Druckman, 2001; Iyengar, 1991). In the context of EU integration, media impact studies have already confirmed that positive and negative cueing of the EU (in respect to EU enlargement) affects public support and opposition (De Vreese and Boomgaarden, 2006; Schuck and De Vreese, 2006).

Second, we know that such support and opposition might be presented in different ways. Hence, in addition to the evaluative tone, we distinguish between cues that focus on identity or group membership questions and cues that centre on the economy. As empirical studies have shown, such a differentiation enhances our understanding of public opinion formation (De Vries and Edwards, 2009; Maier and Rittberger, 2008). We argue in this paper that citizens’ general EU support depends on whether they perceive their key economic interest or their national identity as being threatened by processes of integration (see Hooghe and Marks, 2005). More particularly, we claim that this cueing dimension is of special importance in the context of the EU because it structures not only citizens’ opinion formation but also parties’ positioning towards Europe.

Method

To test whether national parties have an impact on citizens’ EU support by publicly cueing Europe as a risk or an opportunity for the economy or identity we conducted a survey experiment in nine countries (Austria, the Czech Republic, Germany, Spain, the UK, the Netherlands, Poland, Portugal and Sweden). As a consequence of our considerations regarding the two dimensions of cueing (evaluation and content), respondents were randomly provided with four different types of party cues. One group of participants was exposed to positive economic cues (Group 1), one to negative economic cues (Group 2), one to positive identity cues (Group 3) and the last to negative identity cues (Group 4). In order to analyse the effects of party cueing we compared these experimental groups with a control group. The experiment took place in the week before the 2009 European Parliament (EP) elections and relied on real-world campaign material of domestic parties. The material was systematically selected on the basis of a thorough content analysis and country experts’ judgements (see below).

With our study design we have departed from the standard approach of cueing theory and apply an experimental logic. Experimental research has advantages as well as shortcomings. Our research design, however, addressed three of these drawbacks: the artificiality of experiments, the ‘captive’ populations involved in experiments (Iyengar and Simon, 2000: 163) and the blindness towards country contexts.

Experiments in general are characterized not only by an artificial reception situation but often also by stimuli that do not correspond to the real world. We decided to use real-world cues in our experiments stemming from the 2009 EP election campaigns. This procedure limits the artificiality of our experiments but also reduces our capabilities to fully control the experimental setting. Owing to a possible experimental threat brought about by the lack of homogeneity of the treatment material within experimental groups as well as by the lack of variance of the material between the groups (see Araujo et al., 2007; Brunswik, 1956; Kingstone et al., 2003; Rozin, 2009; Maier et al., 2011a; Maier et al., 2011b), we have selected our stimulus material in a multi-step procedure combining systematic content analysis with country experts’ judgements (detailed information regarding this content analysis can be found in the Web Appendix).

In a content analysis, all televised advertisements and posters of those parties winning more than 3 percent of the votes in the 2009 EP elections were analysed regarding formal and content characteristics (Adam and Maier, 2009). To ensure the reliability of coding, all coders took part in a common training programme and a reliability test was conducted. With an average inter-coder reliability of .71 for the television spots and .82 for the campaign posters across all nine countries, the reliability test according to the ‘Holsti Formula’ (Kolb, 2004; North et al., 1963) delivered satisfactory results.

On the basis of the content analysis, we attempted to build a sample for each of the four cueing conditions described above in as many of the nine countries included in the study as possible. The material should stem from one party per condition and should consist of one to two televised advertisements plus, if possible, a couple of posters (a description of the selected campaign items may be found in the Web Appendix). As a result, these samples are not necessarily representative of a party campaign as a whole but rather represent a specific part of a party’s campaign material that allowed us to maximize the fit with the above-mentioned criteria. A necessary precondition for the selection of campaign material as a stimulus was its focus on European issues and actors (so-called ‘first-order’ campaign materials; see Reif and Schmitt, 1980) since this material typically contains elite cues regarding Europe. Two further prerequisites had to be met: first, we chose only material in which identity or economic questions were discussed and in which one of these two cleavages clearly prevailed; second, our research design required that our stimulus material contained a clear evaluation of the EU. Because the degree of opposition to and support for the EU varies within parties’ campaign material, we created an additive index ranging from −3 as an extreme form of EU scepticism to +3 expressing the strongest support. This index is based upon three variables from our content analysis: (a) the evaluation of the EU as a risk or an opportunity with regard to the dominant cleavage (economy or identity); (b) the overall evaluation of the EU as it is today;⁴ and (c) the vision of how the country should act within the EU, ranging from ‘full involvement’ to the extreme demand to ‘withdraw’ from membership. This metric classification of our stimulus material serves to prevent us from mixing campaigns that feature strong EU opposition (for example, the United Kingdom Independence Party, UKIP) with campaigns in which some critical remarks about Europe are voiced (for example, the Dutch Partijvoor de Vrijheid, PVV). In addition to the content analysis, a native-speaker project partner from the relevant country had to verify the correct assignment of the selected materials to one of the four conditions.

This selection procedure allowed us to identify the following campaigns as fitting our treatment cells:

Positive economic cues. Four campaigns – Austria: Österreichische Volkspartei (ÖVP); the Netherlands: Partij van der Arbeid (PvdA); Portugal: Partido Socialista (PS); UK: Liberal Democrats;

Negative economic cues. Five campaigns – the Netherlands: ChristenUnie (CU), Libertas; Spain: Libertas; Sweden: Folkpartijet; UK: UKIP;

Positive identity cues. Three campaigns – Czech: Strana zelených (SZ); Poland: Platforma Obywatelska (PO), Centrolewica;

Negative identity cues. Four campaigns – Czech: Suverenita, Strana svobodných občanů (SSO); Germany: Deutsche Volksunion (DVU); the Netherlands: PVV.

In total, we were able to assemble suitable campaign material for 16 experimental groups.

Experiments are often conducted with captive populations (Iyengar and Simon, 2000), such as students. To avoid the problems typical of such samples, we relied on a professional online access panel provided by a market research institute (GfK, Nuremberg, Germany) for all nine countries under study. The panel fulfilled the highest standards of online research according to the guidelines of the leading professional associations such as ESOMAR (see www.esomar.org) and German ADM (Arbeitskreis Deutscher Markt- und Sozialforscher e.V.; see www.adm-ev.de), for example, online and offline recruitment of participants. The quota samples drawn from this panel should represent each country’s population with regard to age, education, gender and geographical region.

For each experimental group, a minimum of 50 participants were recruited. In total, 896 participants were assigned to the four different treatment cells. Including one control group for each country, which filled in the questionnaire without receiving the treatment, a total of 1379 participants from nine countries took part in our experiment. The survey consisted of an initial questionnaire comprising questions on non-treatment-sensitive attitudes. Following this, participants were randomly assigned to the treatment groups and the control group. After the treatment, a second questionnaire collected the indicators for the dependent variable (all variables relevant for this article can be found in the Web Appendix).

Finally, experiments are normally conducted in one country setting without taking into account the fact that a given country context may be a variable that needs to be controlled when studying effects (for a critique, see Maier and Rittberger, 2008: 261). To remedy this shortcoming we have included nine European countries – from the East and the West – in our experimental study. This research design makes us more confident that our experimental results are valid outside the individual country.

Data and strategy of analysis

The assessment of the effects of parties’ public campaign communication on citizens’ attitudes towards EU integration first of all requires an adequate measure of these attitudes. For this purpose, an index (see also Hooghe and Marks, 2005: 427) of five original questions was created, each measured on a five-point Likert scale (see the Web Appendix), assessing (a) satisfaction with EU politics; (b) evaluation of the country’s EU membership; (c) evaluation of the advantages and disadvantages of EU integration for the particular country; (d) views on further EU enlargement; and (e) views on the degree of EU integration. The index, which also ranged from 1 (‘very negative attitudes’) to 5 (‘very positive attitudes’), had a Cronbach’s alpha of .86.

Having specified the experimental treatment and the dependent variable, we now turn to our strategy of analysis. We conduct stepwise ordinary least squares (OLS) regression analyses.⁵ We do so separately for each of our four experimental situations by comparing the effect in the particular treatment group with that in the control group. This differentiation is necessary because we assume that variation in the content of cues (which can be seen in the different experimental treatment cells) might lead to variations in effects. The basic model (Model 1) tests whether the treatments yield significant effects in the expected direction (Hypothesis 1). Since we expect effects to be strongest for partisans (Hypothesis 2), we include the interaction between the treatment and the strength of identification with the party that produced the particular campaign material as an additional explanatory variable in Model 2.⁶

To test for the stability of treatment effects, in Model 3 we control for those individual-level variables that have proved to be relevant for the formation of EU attitudes in numerous studies. This strategy to control the effects of nuisance factors that are not the focus of the experimental design (see Huitema, 1980) has also been applied in similar studies by Maier and Rittberger (2008) and Sigalas (2010). Three types of controls are included. First, the two most prominent factors in EU research are economic and identity considerations (Hooghe and Marks, 2005). We take education as a proxy for economic considerations (Anderson and Reichert, 1996; Carey, 2002; Gabel and Palmer, 1995; McLaren, 2002; Medrano, 2003). The identity variable was created by subtracting individuals’ rating of the extent to which they identified themselves with Europe – values ranged from 1 (‘not at all’) to 5 (‘very strongly’) – from the extent to which they identified themselves with their country (using the same range of values). This measure has values from −4 (‘exclusive European identity’) to +4 (‘exclusive national identity’). Second, we add two sociodemographic factors that have proved to be related to EU opinions (Hooghe and Marks, 2005; Hix, 2005): age and gender. Third, we control for the effect of political cognitive mobilization, which, as the famous Inglehart study in 1970 showed (see also Gabel and Whitten, 1997; Inglehart et al., 1991; Janssen, 1991), should have positive effects on EU attitudes, that is, the variables involvement and information efficacy. The findings of Mondak (1993) suggest that these variables, together with formal education, should be strong controls for cueing effects, even though Karp et al. (2003) found that political awareness has negative effects on satisfaction with the EU. In our study, political involvement was measured as an index derived from four items (all on five-point Likert scales): (a) reception of political news; (b) engagement in interpersonal communication about politics; (c) active seeking of political information; and (d) thinking about political issues. This index had a Cronbach’s alpha of .88 and ranged from 1 (‘very low political involvement’) to 5 (‘very high political involvement’) (see also Otto and Bacherle, 2011). Political information efficacy was measured as an adaptation of the concept of Kaid and Bystrom (2007) using the items (a) feeling competent to participate in EU politics; (b) feeling better informed than most people regarding EU politics; (c) understanding EU issues; and (d) having enough information to advise a friend regarding EU politics (all five-point Likert scales). This index had a Cronbach’s alpha of .89 and values ranging from 1 (‘very low information efficacy’) to 5 (‘very high information efficacy’).⁷ Following the reasoning of Hooghe and Marks (2005), Model 4 finally adds country dummies as control variables for the different national contexts.

Empirical results

In the following, we test our hypotheses about whether significant treatment effects exist in the sense of elite cueing, whether partisanship is a relevant moderating variable, whether the effects are stable when controlling for nuisance variables and whether country contexts play a significant role. Tables 1 to 4 follow the logic of analysis described above. In Model 1 the effect of the particular treatment is tested by bivariate regression (Basic model).⁸ Model 2 adds the interaction between experimental treatment and party identification (PI model). Model 3 then controls for important individual-level predictors of EU attitudes – for example, identity and economic considerations, cognitive mobilization, socio-structural variables and political ideology (Individual predictors model). Finally, Model 4 takes into account country dummies (Context model).⁹

Table 1.

Positive economic treatment (group 1) vs. control group

	Basic model		Party identification model		Individual-level predictors model		Country-level predictors model
	B	SE	B	SE	B	SE	B	SE
Intercept	2.787	.062	2.779	.061	3.115	.295	2.923	.314
Treatment
Treatment (main effect)	−.107*	.041	−.141**	.042	−.063^a	.035	−.045	.036
Treatment x PI			.202***	.055	.148**	.047	.133**	.046
Individual nuisance factors
Exclusive national identity					−.324***	.030	−.300***	.030
Education					.066^a	.039	.079*	.039
Political involvement					.062	.059	.048	.060
Information efficacy (EU)					−.019	.050	.009	.050
Age					−.010**	.003	−.007^c	.004
Gender					−.057	.081	−.023	.081
Controls for country contexts ^y
Country dummy UK							−.211^d	.108
Country dummy POR							.180	.127
Country dummy AU							−.133	.125
Adjusted R²	.01		.04		.35		.36
F	6.78		10.14		29.34		23.05
N	439		439		428		428

Notes: Cell entries are unstandardized B coefficients and standard errors.

Reference country: The Netherlands.

***p < .001, **p < .01, *p < .05, ^ap = .076, ^bp = .093, ^cp = .050, ^dp = .051.

Out of our four regressions (Model 1), two show significant effects of the treatment variables:¹⁰ both the positive and the negative economic treatments yield statistically significant negative effects on participants’ EU attitudes (see Tables 1 and 2). The cultural treatments, however, do not show significant main effects (see Tables 3 and 4). These findings are in line with current research (De Vreese et al., 2012) but only partially agree with the general cueing hypothesis: although the highly significant effect of the negative economic treatments is in line with our expectation, the statistically significant negative effects of the positive economic conditions are contradictory to Hypothesis 1. Obviously, participants are not following the evaluative tone of the clearly positive persuasive message here. It may be speculated that this effect (if it proves to be stable in further analyses) might be linked to the global economic crisis, which was at a peak just before the 2009 EP elections: any exposure to economic campaign materials may have activated negative attitudes connected to this crisis, thereby resulting in a negative EU evaluation. For the cultural treatments (positive and negative), we do not find any significant effects, which would have supported the general cueing hypothesis.

Table 2.

Negative economic treatment (group 2) vs. control group

	Basic model		Party identification model		Individual-level predictors model		Country-level predictors model
	B	SE	B	SE	B	SE	B	SE
Intercept	2.800	.057	2.800	.057	2.859	.309	2.533	.301
Treatment
Treatment (main effect)	−.197***	.043	−.192***	.044	−.121**	.039	−.074^a	.040
Treatment x PI			−.057	.079	−.028	.067	.011	.065
Individual nuisance factors
Exclusive national identity					−.380***	.030	−.333***	.030
Education					.070	.043	.067	.042
Political involvement					.107*	.054	.061	.053
Information efficacy (EU)					−.033	.050	−.018	.048
Age					−.002	.003	.003	.003
Gender					−.084	.079	−.089	.077
Controls for country contexts ^y
Country dummy UK							−.256*	.108
Country dummy ES							.488***	.108
Country dummy SW							.215*	.105
Adjusted R²		.04		.04		.31		.36
F		20.49		10.49		27.35		25.38
N		479		479		470		470

Notes: Cell entries are unstandardized B coefficients and standard errors.

Reference country: The Netherlands.

***p < .001, **p < .01, *p < .05, ^ap = .065.

Table 3.

Positive cultural treatment (group 3) vs. control group

	Basic model		Party identification model		Individual-level predictors model		Country-level predictors model
	B	SE	B	SE	B	SE	B	SE
Intercept	3.280	.068	3.246	.068	2.314	.376	2.826	.377
Treatment
Treatment (main effect)	−.061	.051	−.073	.050	−.080	.050	−.074	.048
Treatment x PI			.127**	.041	.078^a	.041	.058	.040
Individual nuisance factors
Exclusive national identity					−.114*	.047	−.121**	.045
Education					.127*	.056	.036	.057
Political involvement					.090	.072	.101	.069
Information efficacy (EU)					.106^b	.055	.064	.054
Age					−.005	.004	−.003	.004
Gender					.024	.091	.059	.088
Controls for country contexts ^y
Country dummy CZ							−.459***	.097
Adjusted R²	.00		.03		.11		.17
F	1.43		5.59		4.84		7.13
N	267		267		263		263

Notes: Cell entries are unstandardized B coefficients and standard errors.

Reference country: Poland.

***p < .001, **p < .01, *p < .05, ^ap = .075, ^bp = .056.

Table 4.

Negative cultural treatment (group 4) vs. control group

	Basic model		Party identification model		Individual-level predictors model		Country-level predictors model
	B	SE	B	SE	B	SE	B	SE
Intercept	2.738	.062	2.747	.062	2.821	.324	2.749	.331
Treatment
Treatment (main effect)	.016	.047	.029	.047	−.028	.043	−.044	.045
Treatment x PI			−.102*	.048	−.087*	.043	−.081^a	.044
Individual nuisance factors
Exclusive national identity					−.256***	.037	−.251***	.038
Education					.106*	.046	.105*	.047
Political involvement					.082	.057	.064	.060
Information efficacy (EU)					.047	.049	.050	.049
Age					−.010**	.003	−.009**	.003
Gender					−.112	.080	−.108	.080
Controls for country contexts y
Country dummy CZ							.118	.107
Country dummy DE							.083	.114
Adjusted R²	.00		.01		.20		.20
F	.12		2.30		13.61		10.99
N	414		414		404		404

Notes: Cell entries are unstandardized B coefficients and standard errors.

Reference country: The Netherlands.

***p < .001, **p < .01, *p < .05, ^ap = .065.

Model 2 tests whether elite cueing works best for the party supporters – the so-called partisan hypothesis. The results from this interaction between treatment and attachment to the communicating party (PI) support the hypothesis in three out of four cases: participants exposed to a campaign from a party they strongly identify with follow their cue-givers more closely than non-partisans do. Thus, being exposed to positive EU evaluations (economic and cultural) by the preferred party resulted in more positive EU attitudes (see Tables 1 and 3). In contrast, if parties’ campaign messages suggested a negative cultural evaluation of the EU, participants with strong party attachment showed significantly less favourable EU attitudes than citizens without or with only weak party identification (see Table 4). It was only in the case of negative economic messages that the interaction with party attachment did not have any significant effect (see Table 2). However, for this group the main effect of the treatment remains stable.

In Model 3 we test the stability of our treatment effects by controlling for individual-level variables. The results clearly show that cueing matters. Even when controlling for the main variables explaining EU support at an individual level, in three out of the four treatment conditions the effects remain stable (see Tables 1, 2 and 4) – the only exception being the effect of positive cultural campaign messages (see Table 3). However, there is evidence for the general hypothesis (negative economic; see Table 2) as well as the partisan cueing hypotheses (positive economic and negative cultural treatment; see Tables 1 and 4). Among the control variables, identity shows the strongest effects by far: people who describe their own identity as exclusively national (in contrast to seeing oneself as a European citizen) have the strongest negative attitudes toward the EU in all treatment conditions. The second-strongest effect stems from the variable education, which we have included as a proxy for economic considerations: in three out of four treatments citizens with a higher level of education have the best chances of benefiting from EU integration and therefore support it, whereas people with low formal education tend to see themselves as economic losers from integration and therefore oppose it. Of the other variables included as controls, only age has the expected slightly negative effect on EU attitudes, which means that younger people are more open to EU integration than older people. Whereas age has a significant impact in half of the estimated models, political involvement has a systematic effect on EU attitudes only in the negative economic treatment condition: citizens with high interest show more support for the European Union. The impact of information efficacy and gender is always insignificant. This indicates that there are no important differences in EU support in respect of political competence or between males and females.

For the national context in which campaigns take place, we also find significant effects (Model 4). First and most important, we can show that our treatment effects remain stable even when controlling for such country factors. Second, we see differences between countries: when compared with the Netherlands, EU attitudes are significantly less favourable in the UK (positive and negative economic treatment groups; see Tables 1 and 2), whereas participants were more positive in both Spain and Sweden (negative economic treatment group; see Table 2). In comparison with Poland, Czech participants voiced significantly more positive EU attitudes (see Table 3). If we take a look at the other control variables, the picture is very much the same as in Model 3. Identity always has a strong and significant impact on EU attitudes. In a few treatment conditions, education and age show significant effects, too. Political involvement, information efficacy and gender remain not systematically related to EU support.

Summarizing the results thus far, we find partial support for the general cueing hypothesis (negative economic messages) and, in two other groups, partisanship proves to be a relevant moderator for the effects of persuasive messages (positive economic and negative cultural treatment; see Figures 1 and 2). These effects are stable even when thoroughly controlling for nuisance variables drawn from EU research as well as country contexts.

Figure 1.

Effect of positive partisan cueing of economic issues on support for EU integration.

Figure 2.

Effect of negative partisan cueing of cultural issues on support for EU integration.

Conclusions

Our analyses provide empirical evidence that cueing matters under certain conditions. Even the short experimental exposure to party campaign materials (television spots and posters) in the run-up to the 2009 EP elections led to effects on citizens’ EU attitudes. However, our research emphasizes that elite cueing does not work as a simple stimulus–response model. There is only one treatment condition where we find a direct effect: negative economic messages come closest to a general cueing effect on participants in our study. It seems that clearly negative economic campaign messages in the 2009 EP election campaign enforced negative attitudes towards Europe – independently from party identification but also independently from individual nuisance factors and country contexts. In two other conditions – very clearly in the positive economic treatment group and less clearly in the negative cultural treatment group – cueing effects were moderated by partisanship. Participants followed the persuasive message of the campaign material only if they felt strongly attached to the party sponsoring it. It was only in the positive cultural treatment group that no effects of elite cueing could be found.

In order to test for the stability of these cueing effects, we controlled for a broad range of alternative factors that have been identified in order to explain EU attitudes. These control models reveal two significant findings. First, they reconfirm the extraordinary relevance of identity and economic considerations for EU attitude formation: an exclusive national identity and lower education (our proxy for the economic disadvantages of integration) were by far the strongest predictors of EU-sceptical attitudes. We were also able to reconfirm the relevance of country contexts by including dummy variables. Second, but as important, our control models clearly show that cueing effects are stable. In three out of four treatment conditions, cueing effects significantly influence EU attitudes after all controls have been included. These findings strongly support Zaller’s (1992) insight that opinions are composed of information (new cues) and predispositions (individual-level controls), which, as we argue, are shaped by specific country settings (macro-context factors). Consequently, future studies seeking to fully understand the formation of EU attitudes need to take into account all three factors – cueing, individual-level predispositions and country contexts.

Adding to traditional cueing research we have sought to shed light on the process of cue transmission. So far, cueing researchers have preferred structural context variables of domestic party systems as proxies for cues that dominate within a country. They assume that these cues affect the respective partisan supporters. If it is true that structure first translates into cues and then into effects, this chain of reasoning is not problematic. However, by introducing an experimental methodology to cueing research we want to at least partly test the validity of this implicit assumption: do cues that are publicly communicated and to which citizens are actually exposed affect citizens’ or partisan supporters’ EU attitudes in the expected way? In our experiments we have found evidence for both assumptions of traditional cueing research: the significant main effects in the negative economic treatment group support the general cueing hypothesis that citizens follow the persuasive messages of political elites; the partisan hypothesis, on the other hand, is supported in the positive economic and negative cultural treatment groups.

However, the question remains of why general cueing works in the case of the negative economic stimuli whereas in the other cases it is only partisans who follow the messages of their preferred party. One possible explanation is that participants may have been primed (see Druckman, 2004; Iyengar and Kinder, 1987; Krosnick and Brannon, 1993) by the context of the 2009 EP elections. As we have pointed out before, the financial crisis in 2009 was at its peak during the campaigns, and this priming may have led to the uniform main effect that we find in this specific treatment group. This explanation is in line with Slothuus and De Vreese (2010: 643), who find that ‘parties can be quite powerful in shaping public opinion, not only among their own voters, but also more broadly in the electorate on consensus issues in which citizens do not discriminate as strongly as to which party promotes the frame.’ Within such consensus issues there is a clear dominant – consensual – position amongst the electorate that a party might support. In the context of our study, we claim that the financial crisis might have been such a ‘consensus issue’ in which party attachment was not a necessary prerequisite to join the parties’ complaints about the EU economy. Such complaints were already foremost in citizens’ minds.

In contrast, the opposite should hold true in regard to the so-called ‘conflict issues where parties have signalled that party values are at stake’ (Slothuus and De Vreese, 2010: 643). Putting a conflict issue – more precisely, a conflictual position on a specific issue – on the agenda changes the effect of party communication. It is then not only the evaluation standards (issue cueing or priming) but also the evaluation itself (positional cueing) that becomes relevant. However, the evaluation itself is merely taken over by partisans. Our results suggest that a party stands to gain most by placing an issue on the agenda where it just follows a consensual position. In this case party communication reaches beyond its own followers. In cases where a party contradicts a consensual position (such as the Austrian ÖVP campaigning on a pro-European economic message in the midst of the financial crisis) or takes over a position on an inherently conflictual issue (for example, the Dutch PVV with its Eurosceptic cultural campaign), party communication just serves to uphold party values and strengthens the bond with party followers. Hence, the context of a campaign determines which issues/positions allow parties to trigger priming (issue cueing) effects towards a broader audience and which allow positional cueing only for their respective supporters. Further research needs to show why the treatment effect in the positive cultural cell does indeed have no significant effects.

Finally, future experiments on the effects of party cues should take into account not only the context-driven nature of the content of cues but also the type of messenger putting forward the cue. We know that recipients accept or reject persuasive messages based on source credibility (for example, Iyengar and Valentino, 2000). This is empirically supported in our experiments by the interaction effect between treatment and partisanship. Even beyond this, it might turn out that a highly credible source affects the wider electorate, whereas a non-credible source has difficulties convincing its own partisans. This question becomes all the more relevant in the context of EU matters because research – and, indeed, our stimulus selection – shows that Eurosceptic positions are primarily voiced by the more extreme or single-issue parties. For these parties, moderates may reject the persuasive message owing to the fact that, independently of the content, they object to the messenger.

What our experiments already show is that mainstream parties (which dominate the positive economic and cultural treatments) do not automatically convince a wider audience than more extreme parties (which dominate the negative cultural treatment). Yet it might be that mainstream parties adopting increasingly negative attitudes towards Europe have a wider appeal: it is in our negative economic treatment that two out of five parties (the Dutch CU and the Swedish Folkpartijet) voice mainstream Euroscepticism. This leaves us with two interpretations of why we are observing a broader cueing effect here, thus leaving room for further research: this might be a priming effect on a consensual issue or it might be a messenger effect. These considerations also hint at the limitations of real-world experiments. If we want to disentangle the effects of the message versus the messenger (see also Kam, 2005), we need full control over the contexts of the experiments. Hence, future studies must complement experiments based on real-world stimuli with ones based on artificially altered treatment materials.

In sum, our results suggest that parties can affect citizens’ EU attitudes with new information on Europe: elite cues do indeed matter. However, the effects achieved differ. Regarding most questions, successful campaign communication brings partisans in line – a valuable asset of pro-European parties that are increasingly confronted with the voicing of outspoken Euroscepticism in nearly all member states of the Union. However, there are a few instances where we expect that new information might reach beyond the particular partisan base. Such instances occur if a (credible) party succeeds in following a consensual position. Cueing then leads to a change in the evaluation standards of the wider electorate and, therefore, to opinion formation without persuasion. In theory, Eurosceptics as well as Euro-supporters may profit from this mechanism. At the moment – being in the midst of a European financial crisis and observing the upsurge of right-wing identity mobilization on Europe – Eurosceptics have a more favourable issue environment in terms of following (more or less) consensual issues. If the pro-European mainstream seeks to be successful in this battle over public opinion on Europe, it is not sufficient merely to engage in a positional competition on EU integration. What is necessary is an active shifting of the agenda towards consensual issues, which would allow them to benefit.

Footnotes

Acknowledgements

The study presented here was part of the project ‘Between Integration and Demarcation: Strategies and Effects of Party Campaigns in the Context of the 2009 European Parliament Elections’. This project was directed by Michaela Maier (University of Koblenz-Landau) and Silke Adam (University of Bern). Partial funding was granted by the research group (Kollegforschergruppe) ‘The Transformative Power of Europe’ of the Free University Berlin and the Department of Communication Psychology at the University of Koblenz-Landau. It would not have been possible without the help of many colleagues and assistants who helped with the design, the instruments and especially the data-gathering. We would like to thank all these helping hands for their support – especially Matthias Balzer, Manuela Baumli, Rosa Berganza, Boguslava Dobek-Ostrowska, Claes de Vreese, Marie Grusell, Carlos Jalali, Ralph Negrine, Lars Nord, Frank Schneider, Andreas Schuck, Gilg U.H. Seeber, Vaclav Stetka and Karin Stengel.

Notes

References

Adam S and Maier M (2009) Codebook of Content Coding: Strategies of Party Campaigns in the Context of the EP Elections. URL (accessed 18 July 2012): http://www.uni-koblenz-landau.de/landau/fb8/ikms/ikm/forschung/forschungdoc/codebook.pdf.

Adam

Maier

(2011) National parties as politicizers of EU integration? Party campaign communication in the run-up to the 2009 European Parliament Election. European Union Politics 12(3): 431–454.

Adam

Maier

De Vreese

Schuck

ART

Stetka

Jalali

(in press) Campaigning against Europe? The role of Euroskeptic fringe and mainstream parties in the 2009 European Parliament election. Journal of Political Marketing.

Anderson

Reichert

(1996) Economic benefits and support for membership in the EU: A cross-national analysis. Journal of Public Policy 15(3): 231–249.

Araujo

Davids

Passos

(2007) Ecological validity, representative design, and correspondence between experimental task constraints and behavioral setting: Comment on Rogers, Kadar, and Costall (2005). Ecological Psychology 19(1): 69–78.

Boomgaarden

Schuck

ART

Elenbaas

De Vreese

(2011) Mapping EU attitudes: Conceptual and empirical dimensions of euroscepticism and EU support. European Union Politics 12(2): 241–266.

Brunswik

(1956) Perception and the Representative Design of Psychological Experiments, Berkeley: University of California Press.

Bruter

(2003) Winning hearts and minds for Europe: The impact of news and symbols on civic and cultural European identity. Comparative Political Studies 36(10): 1148–1179.

Campbell

Converse

Miller

Stokes

(1980) The American Voter: Unabridged Edition, Chicago: University of Chicago Press.

10.

Carey

(2002) Undivided loyalties: Is national identity an obstacle to European integration? European Union Politics 3(4): 387–413.

11.

Carey

Burton

(2004) Research note: The influence of the press in shaping public opinion towards the European Union in Britain. Political Studies 52(3): 623–640.

12.

Carrubba

(2001) The electorate connection in European Union politics. Journal of Politics 63(1): 141–158.

13.

Cidrin

Sides

(2004) Can there be a Europe without Europeans? Problems of identity in a multinational community. In: Herrmann

Brewer

Risse

(eds) Transnational Identities: Becoming European in the EU, Lanham, MD: Rowman & Littlefield, 161–185.

14.

Converse

(2006) The nature of belief systems in mass publics. Critical Review 18(1–3): 1–74 (reprinted from Apter DE (ed.) Ideology and Discontent. New York: Free Press, 207–261).

15.

Dalton

Duval

(1986) The political environment and foreign policy opinions: British attitudes towards European integration 1972–1979. British Journal of Political Science 16(1): 113–134.

16.

De Vreese

(2004) The effects of frames in political television news on issue interpretation and frame salience. Journalism & Mass Communication Quarterly 81(1): 36–52.

17.

De Vreese

Boomgaarden

(2006) Media effects on public opinion about the enlargement of the European Union. Journal of Common Market Studies 44(2): 419–436.

18.

De Vreese

Lauf

Peter

(2007) The media and European Parliament elections: Second-rate coverage of a second-order event? In: Van der Brug

Van der Eijk

(eds) European Elections & Domestic Politics: Lessons from the Past and Scenarios for the Future, Notre Dame: University of Notre Dame Press, 116–130.

19.

De Vreese

Van der Brug

Hobolt

(2012) Turkey in the EU? How cultural and economic frames affect support for Turkish accession. Comparative European Politics 10: 218–235.

20.

De Vries

Edwards

(2009) Taking Europe to its extremes: Extremist parties and public euroscepticism. Party Politics 15(1): 5–28.

21.

Dinas

Gemenis

(2010) Measuring parties’ ideological positions with manifesto data: A critical evaluation of the competing methods. Party Politics 16(4): 427–450.

22.

Down

Wilson

(2010) Opinion polarization and inter-party competition on Europe. European Union Politics 11(1): 61–87.

23.

Druckman

(2001) On the limits of framing effects: Who can frame? Journal of Politics 63(4): 1041–1066.

24.

Druckman

(2004) Priming the vote: Campaign effects in a U.S. Senate election. Political Psychology 25(4): 577–594.

25.

Druckman

Nelson

(2003) Framing and deliberation: How citizens’ conversations limit elite influence. American Journal of Political Science 47(4): 729–745.

26.

Ellinas

(2010) The Media and the Far Right in Western Europe, Cambridge: Cambridge University Press.

27.

Feld

Wildgen

(1976) Domestic Political Realities and European Unification: A Study of Mass Publics and Elites in the European Community Countries, Boulder, CO: Westview Press.

28.

Franzmann

Kaiser

(2006) Locating political parties in policy space: A reanalysis of party manifesto data. Party Politics 12(2): 163–188.

29.

Gabel

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

30.

Gabel

Huber

(2000) Putting parties in their place: Inferring party left–right ideological positions from party manifesto data. American Journal of Political Science 44(1): 94–103.

31.

Gabel

Palmer

(1995) Understanding variation in public support for European integration. European Journal of Political Research 27(1): 3–19.

32.

Gabel

Scheve

(2007) Mixed messages: Party dissent and mass opinion on European integration. European Union Politics 8(1): 37–59.

33.

Gabel

Whitten

(1997) Economic conditions, economic perceptions, and public support for European integration. Political Behavior 19(1): 81–96.

34.

Garry

Tilley

(2009) The macroeconomic factors conditioning the impact of identity on attitudes towards the EU. European Union Politics 10(3): 361–379.

35.

Hix

(2005) The Political System of the European Union, New York: Palgrave Macmillan.

36.

Hooghe

(2007) What drives euroscepticism? Party–public cueing, ideology and strategic opportunity. European Union Politics 8(1): 5–12.

37.

Hooghe

Marks

(2005) Calculation, community and cues: Public opinion on European integration. European Union Politics 6(4): 419–443.

38.

Hooghe

Marks

(2008) A postfunctionalist theory of European integration: From permissive consensus to constraining dissensus. British Journal of Political Science 39(1): 1–23.

39.

Huber

Arceneaux

(2007) Identifying the persuasive effects of presidential advertising. American Journal of Political Science 51(4): 957–977.

40.

Huitema

(1980) The Analysis of Covariance and Alternatives, New York: Wiley.

41.

Inglehart

(1970) Cognitive mobilization and European identity. Comparative Politics 3(1): 45–70.

42.

Inglehart

Rabier

J-R

Reif

(1991) The evolution of public attitudes toward European integration: 1970–86. In: Reif

Inglehart

(eds) Eurobarometer: The Dynamics of European Public Opinion, London: Macmillan, 111–131.

43.

Iyengar

(1991) Is Anyone Responsible? How Television Frames Political Issues, Chicago: University of Chicago Press.

44.

Iyengar

Kinder

(1987) News That Matters, Chicago: University of Chicago Press.

45.

Iyengar

Simon

(2000) New perspectives and evidence on political communication and campaign effects. Annual Review of Psychology 51: 149–169.

46.

Iyengar

Valentino

(2000) Who says what? Source credibility as a mediator of campaign advertising. In: Lupia

McCubbins

Popkin

(eds) Elements of Reason: Cognition, Choice and the Bounds of Rationality, New York: Cambridge University Press, 108–129.

47.

Janssen

(1991) Postmaterialists, cognitive mobilization, and support for European integration. British Journal of Political Science 21(4): 443–468.

48.

Kaid

Bystrom

(2007) The Electronic Election: Perspectives on the 1996 Campaign Communication, Mahwah, NJ: Erlbaum.

49.

Kam

(2005) Who toes the party line? Cues, values, and individual differences. Political Behavior 27(2): 163–182.

50.

Karp

Banducci

Bowler

(2003) To know is to love it? Satisfaction with democracy in the European Union. Comparative Political Studies 36(3): 271–292.

51.

Kingstone

Smilek

Ristic

Friesen

Eastwood

(2003) Attention researchers! It is time to look at the real world. Current Directions in Psychological Science 12(5): 176–184.

52.

Kolb

(2004) Verlässlichkeit von Inhaltsanalysedaten: Reliabilitätstest, Errechnen und Interpretieren von Reliabilitätskoeffizienten für mehr als zwei Codierer. Medien und Kommunikationswissenschaft 52(3): 335–354.

53.

Kriesi

(2008) Rejoinder to Liesbet Hooghe and Gary Marks, ‘A Postfunctional Theory of European Integration: From Permissive Consensus to Constraining Dissensus’. British Journal of Political Science 39(1): 221–224.

54.

Kriesi

Grande

Lachat

Dolezal

Bornschier

Frey

(2006) Globalization and the transformation of the national political space: Six European countries compared. European Journal of Political Research 45(6): 921–956.

55.

Krosnick

Brannon

(1993) The impact of war on the ingredients of presidential evaluations. Multidimensional effects of political involvement. American Political Science Review 87(4): 963–975.

56.

Kumlin

(2011) Claiming blame and giving credit? Unintended effects of how government and opposition frame the Europeanization of welfare. European Union Politics 12(4): 575–595.

57.

Lau

Pomper

(2002) Effectiveness of negative campaigning in U.S. Senate elections. American Journal of Political Science 46(1): 47–66.

58.

Loveless M and Rohrschneider R (2008) Public perceptions of the EU as a system of governance. Living Reviews in European Governance 3(1). Available at: http://www.livingreviews.org/lreg-2008–1 (accessed 24 April 2012).

59.

McLaren

(2002) Public support for European integration: Cost/benefit analysis or perceived cultural threat. Journal of Politics 64(2): 551–566.

60.

Maier

Rittberger

(2008) Shifting Europe’s boundaries: Mass media, public opinion and the enlargement of the EU. European Union Politics 9(2): 243–267.

61.

Maier

Adam

De Vreese

Schuck

Berganza

Stetka

(2011a) Between integration and demarcation: Effects of Europeanized and national campaigns on voters in the 2009 European Parliament election. In: Maier

Strömbäck

Kaid

(eds) Political Communication in European Parliamentary Elections, Farnham/Burlington, VA: Ashgate, 233–256.

62.

Maier

Stengel

Ruhrmann

Marschall

Zillich

Göbbel

(2011b) Synchronisierung von Erhebungsinstrumenten zur Erfassung des dynamischen Nachrichtenprozesses am Beispiel der Krisenkommunikation. In: Jandura

Quandt

Vogelgesang

(eds) Methoden der Journalismusforschung, Wiesbaden: VS Verlag für Sozialwissenschaften, 237–255.

63.

Medrano

(2003) Framing Europe: Attitudes to European Integration in Germany, Spain, and the United Kingdom, Princeton, NJ: Princeton University Press.

64.

Mittag

Wessels

(2003) The ‘one’ and the ‘fifteen’? The member states between procedural adaptation and structural revolution. In: Wessels

Maurer

Mittag

(eds) Fifteen into One? The European Union and Its Member States, Manchester: Manchester University Press, 413–457.

65.

Mondak

(1993) Source cues and policy approval: The cognitive dynamics of public support for the Reagan agenda. American Journal of Political Science 37(1): 186–212.

66.

North

Holsti

Zaninovich

Zinnes

(1963) Content Analysis: A Handbook with Applications for the Study of International Crisis, Evanston, IL: Northwestern University Press.

67.

Otto

Bacherle

(2011) Politisches Interesse Kurzskala (PIKS) – Entwicklung und Validierung. Politische Psychologie 1(1): 19–35.

68.

Ray

(2003) When parties matter: The conditional influence of party positions on voter opinion about European integration. Journal of Politics 65(4): 978–994.

69.

Reif

Schmitt

(1980) Nine second-order national elections. A conceptual framework for the analysis of European election results. European Journal of Political Research 8(1): 3–44.

70.

Rozin

(2009) What kind of empirical research should we publish, fund, and reward? A different perspective. Perspectives on Psychological Science 4(4): 435–439.

71.

Schuck

ART

De Vreese

(2006) Between risk and opportunity. News framing and its effects on public support for EU enlargement. European Journal of Communication 21(1): 5–32.

72.

Schuck

ART

De Vreese

(2009) Reversed mobilization in referendum campaigns: How positive news framing can mobilize the skeptics. International Journal of Press/Politics 14(1): 40–66.

73.

Sigalas

(2010) Cross-border mobility and European identity: The effectiveness of intergroup contact during the ERASMUS year abroad. European Union Politics 11(4): 241–265.

74.

Slothuus

De Vreese

(2010) Political parties, motivated reasoning, and issue framing effects. Journal of Politics 72(3): 630–645.

75.

Steenbergen MR and De Vries CE (in press) Variable opinions: The predictability of support for unification in European mass publics. Journal of Political Marketing. URL (accessed 18 July 2012): http://www.catherinedevries.eu/Variable%20Opinions_Steenbergen_deVries_final.pdf.

76.

Steenbergen

Edwards

De Vries

(2007) Who’s cueing whom? Mass–elite linkages and the future of European integration. European Union Politics 8(13): 13–35.

77.

Szczerbiak A and Taggart P (2003) Theorising Party-Based Euroscepticism: Problems of Definition, Measurement and Causality. SEI Working Paper No. 69, Sussex European Institute, University of Sussex, August. URL (accessed 18 July 2012): http://www.sussex.ac.uk/sei/documents/sei-working-paper-no-69.pdf.

78.

Van der Eijk

Franklin

(2004) Potential for contestation on European matters at national elections in Europe. In: Marks

Steenbergen

(eds) European Integration and Political Conflict, Cambridge: Cambridge University Press, 32–50.

79.

Zaller

(1992) The Nature and Origin of Public Opinion, Cambridge: Cambridge University Press.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.09 MB