I Want a Savior, Not a Victim. The Impact of Media Representations of the EU and Economic Shocks on Citizens’ Supportive Attitudes on the European Integration

I Want a Savior, Not a Victim. The Impact of Media Representations of the EU and Economic Shocks... Abstract The Euro crisis substantially diminished citizens’ support for the European integration. Therefore, we intend to analyze the independent effects of both the depiction of the Euro crisis in national news media and macroeconomic contexts on citizens’ support for a deeper integration. Based on a combination of Eurobarometer data, results from a content analysis, and data on the economic performance of citizens’ home regions, we show that presenting the European Union (EU) as a victim impedes citizens’ supportive attitudes concerning the European integration, whereas depicting the EU as responsible for solving the crisis has a strengthening effect. Additionally, sharp economic developments in citizens’ home regions also have significant effects on their prointegration attitudes, suggesting a differential influence of close and remote economic indicators. From the Maastricht Treaty in 1992 to the Lisbon Treaty in 2009, the development of the European Union (EU) was constantly driven toward ever-deeper integration. Public support by the citizens of the EU member countries, however, provided but a fragile basis for such far-reaching institutional transformations (Eichenberg & Dalton, 2007; Serricchio, Tsakatika, & Quaglia, 2013; Toshkov, 2011). The political and economic disruptions of the Euro crisis have further nourished public skepticism toward the EU both as a political institution and as a community of countries (European Commission, 2012; Kriesi, 2014). In explaining citizens’ attitudes concerning the European integration, research has traditionally focused on determinants such as subjective and objective utilitarianism (Gabel, 1998; Guerra, 2013), perceptions of the national political system, or citizens’ identity (Hooghe & Marks, 2005; McLaren, 2004). Recently, the mass media coverage of EU-related issues, namely, the specific valence of framing, has been shown to be crucial in explaining citizens’ EU-related thoughts (de Vreese, 2004) and attitudes (Desmet, van Spanje, & de Vreese, 2015; de Vreese & Boomgaarden, 2006; Vliegenthart, Schuck, Boomgaarden, & de Vreese, 2008). It is even argued that pronounced negativity in the media debate on EU issues may result in a downward spiral of Euroscepticism (Galpin & Trenz, 2017). Beyond the mere valence of EU-related frames, however, it can be assumed that the specific manner, in which the media characterizes the EU as an actor handling problematic situations such as the Euro crisis, will be considered by recipients when forming an opinion on the future of the European project. For example, if the media coverage presents the EU as responsible for solving the crisis, the sovereignty of the EU is emphasized. Whereas presenting the EU as passive, and its possibilities to react as constraint (i.e., as a victim of the crisis), may question its capacity to manage critical situations. In consequence, such attributions of different roles to the EU can be assumed to influence citizens’ opinions on the further development of the EU. Additionally, as the Euro crisis is mainly of an economic nature, it seems plausible to consider the macroeconomic context next to the attitudinal determinants and the media coverage as a contributor to citizens’ perceptions and evaluation of the further development of the EU. The sharp economic downturn with its serious social consequences may represent a “sudden frustration of expectations” (Easton, 1975, p. 445), which is assumed to erode even the sturdiest supportive attitudes toward a political system. Therefore, pronounced economic developments during the Euro crisis in citizen’s immediate environment (e.g., when many people in the local region lose their jobs in a short period of time) may serve as powerful predictors of citizens’ opinions on the European future in times of economic crises. Therefore, it is the aim of our study to enhance our understanding of the media coverage and economic shocks as potential determinants shaping citizens’ supportive attitudes toward the European future. We seek to investigate the independent effects of both the depiction of the EU in the national news media coverage about the Euro crisis and outstanding economic developments in citizens’ immediate everyday environment (i.e., their home region) during the crisis on public attitudes toward a further integration of the EU. At the same time, we control for the impact of attitudinal and evaluative determinants traditionally studied in the context of public support for the EU (i.e., utilitarianism, perceptions of the national and European political system). In doing so, we merge data from (1) a content analysis of the media coverage on the Euro crisis in 10 European countries, and (2) indicators of economic shocks on a regional level with (3) data from the Eurobarometer Survey (November 2012) in a multilevel analysis. Political Attitudes on Europe’s future in the aftermath of the Euro crisis Since 2007, Europe has been troubled by multilayered financial crises: First, the burst of the U.S. housing bubble and the bankruptcy of the Lehman Brothers investment bank caused severe disturbances on the international financial markets, forcing European governments to impose extensive policies to stabilize capital markets. Accordingly, the public debt of several European countries rose dangerously, resulting in these countries experiencing difficulties in refinancing their national deficit (i.e., European Sovereign Debt Crisis). To avoid the collapse of the common currency, the European community established huge bailout packages to help the countries in trouble (cf., Lane, 2012). Severe social and political consequences accompanied the economic and fiscal turmoil (Kriesi, 2014) giving rise to Eurosceptic attitudes all over Europe (Armingeon & Ceka, 2014; Braun & Tausendpfund, 2014; Clements, Nanou, & Verney, 2014). According to David Easton (1965, 1975), one may conclude that such a decline in supportive attitudes may even endanger the stability of the EU. Based on his theory of political systems, Easton (1965, 1975) claims that democratic political systems—national as well as supranational regimes such as the EU—can only be persistent in the long run, if they rest on the support of their citizens. Citizens’ supportive attitudes can be divided into specific and diffuse support (Easton, 1965, 1975). Whereas specific support is related to the short-term output of a political system and its incumbent political authorities, diffuse support is more basic and stable in nature and stems from a generalized evaluation of political systems’ vital objects. These are a sense of trust in and legitimacy for the political regime and the political authorities in general (independent of current incumbents), as well as a sense of belonging to the political community (Easton, 1965, 1975). Drawing on Easton’s general concept of support for political systems, Kopecký and Mudde (2002) conceptualize Euroscepticism in terms of diffuse and specific support toward the European integration. Whereas the first refers to the “ideas of European integration”, the latter applies to “the European Union as the current embodiment of these ideas” (Kopecký & Mudde, 2002, p. 300). Diffuse support can be deemed crucial for the persistence of a political system, as it can compensate for a short-term lack of specific support (Easton, 1965). Hence, while European policymakers can withstand a short-term lack of specific support, pronounced Eurosceptic attitudes in terms of a lack in diffuse support endanger the fundamental supportive basis for the EU’s further development. Therefore, it is the aim of our study to investigate the determinants explaining citizens’ diffuse support for the European integration in times of economic crises. European integration can be conceived in multiple terms (Karp & Bowler, 2006); hence, a further specification seems to be expedient. Similar to Easton’s general concept of trust in political regimes, which means “to believe that the regime’s goals (…) can be trusted to provide outcomes equitable for all” (Easton, 1975, p. 448), diffuse support toward the European integration implies a citizen’s approval of the general idea of sharing sovereignty between the member countries of the EU (Kopecký & Mudde, 2002). Sharing of sovereignty can be understood either in terms of broadening the European community (i.e., enlargement with more countries sharing their sovereignty) or in terms of deepening the cooperation between the member countries in different policy fields such as economic and fiscal policy (Karp & Bowler, 2006). Especially during the Euro crisis, issues such as intensifying the solidarity between the existing member countries dominated the debate among political elites and the general public. Therefore, we focus our analysis on the deepening aspect of European integration. To establish citizens’ approval of the idea of deepening the European integration, we rely on one prominent scenario for the further development of the EU: the notion of a political union (Beck & Grande, 2011). Such a development implies an integration process moving away from the ‘Confederacy of European States’, which is indicated by a nations’ almost full independence from other countries, and stepping closer together into a federation of nation states, sharing more sovereignty in vital parts of the polity (Börzel & Risse, 2000). Citizens’ diffuse support for a political system builds on the general meaning they ascribe to it and on the experience they have made with it (Easton, 1975). In the following, we shall argue how the presentation of the EU in the national media coverage and outstanding economic shocks may serve as sources for such experiences with and impressions of the European integration. Formation of Attitudes Toward the European Integration and the Economic Context Diffuse support for a political system may be based not only on socialization processes but also on experiences with its political objects. Citizens develop generalized attitudes toward the political system in terms of a “product of spill-over effects from evaluations of a series of outputs and of performance” produced by the political objects (Easton, 1975, p. 446). As the European integration was originally driven by the idea of enhancing welfare by way of economic cooperation between the EU member countries, the economic performance may well serve as a basis for citizens’ evaluation of the same. Accordingly, studies drawing on the concept of economic voting (Downs, 1957) show that the economic context in terms of the national economic performance influences citizens’ evaluation of the EU. A pronounced growth of the gross domestic product (GDP), low interest rates, low unemployment rates, and intense intra-EU trade has strengthening effects on citizens’ supportive attitudes (Armingeon & Ceka, 2014; Eichenberg & Dalton, 2007; Serricchio et al., 2013). Most of these studies draw on aggregate economic figures that represent the latest national economic status quo (e.g., most recent GDP growth rate) (Eichenberg & Dalton, 2007). However, such a conceptualization of the economic context does not reflect citizens’ scope of economic experience. Citizens cannot directly observe minor changes of the real economy, as they are displayed by slight increases or decreases of key macroeconomic figures on an aggregate level (e.g., unemployment rate, public debt rate, GDP). Yet, when the macroeconomic context manifests itself in citizens’ immediate environment (e.g., fear of job losses may come up in everyday talk, cost-cutting measures in major companies may result in wage cuts among family members or acquaintances), it may well be perceivable for ordinary citizens (Paldam & Nannestad, 2000). Therefore, especially, the economic consequences in citizens’ local region are supposedly influential. Likewise, citizens have no possibility to evaluate (in absolute terms) the status of the economic situation at one specific point in time (e.g., in 2012). However, sharp economic developments, as they occur during an economic crisis (e.g., the youth unemployment rate in Spain reached a high of 55% at the end of 2012 and Greece experienced a 9% downturn of the GDP in 2011; Eurostat, 2015a), may well leave a lasting impression (e.g., when many people lose their jobs in a brief period of time) (Armingeon & Ceka, 2014). Such sharp economic turbulences may also represent what Easton (1975, p. 445) calls “a sudden frustration of expectations” that has the potential to erode even the more stable entity of diffuse support. Therefore, we assume: H1: The more pronounced the macroeconomic consequences of the Euro crisis were at the regional level, the less supportive will a citizen living in this region be toward the European integration. H2: Outstanding economic developments in a citizen’s direct environment will have a stronger impact on this person's support for a deeper European integration as compared with the aggregate current economic development. Yet, even when relying on economic indicators located on a level that is close to a citizen’s life (i.e., the local region), some economic indicators are directly relevant to their everyday experience (e.g., unemployment, prices). On the other hand, measures of the overall growth of the companies’ output in a region (i.e., GDP) represent more remote aspects of a person's economic experience, as they do not have as great an impact on the everyday economic conditions (Paldam & Nannestad, 2000). We take these considerations into account and assume: H3: The impact of macroeconomic indicators close to citizens’ everyday experience will be more pronounced than the impact of macroeconomic indicators remote from citizens’ everyday life. National Media Coverage and Interpretations of the Crisis Via Role Attributions In its most basic sense, diffuse support “refers to evaluations of what an object is or represents—to the general meaning it has for a person” (Easton, 1975, p. 444). Such a general impression of the European integration may not only be impacted by a sudden frustration of citizens’ expectations on the economic output of the EU. Yet, also the mass media, as the central transmission mechanism between the real world and the pictures in a person's head, is supposed to shape such general impressions of the European integration by presenting European political actors or policies during the Euro crisis in a specific way (Picard, 2015b). Journalists usually try to contextualize and interpret issues using coherent storylines that help recipients make sense of these issues (Galpin & Trenz, 2017). Such basic storylines are called ‘frames’ (Entman, 1993; Sniderman & Theriault, 2004). At the core of framing theory stands the idea that journalists are never able to tell every aspect and every conceivable interpretation of an issue. Rather, they emphasize some sets of considerations while excluding or downplaying others. Consequently, only those aspects that are (frequently) referred to in the media discourse have a notable chance of gaining salience in the minds of the recipients (de Vreese, 2010; Sniderman & Theriault, 2004). Research has accumulated extensive evidence for an influence of media framing on citizens’ EU-related attitudes. In particular, the valence of news frames has been shown to guide recipients’ reasoning about EU matters (Desmet et al., 2015; de Vreese & Boomgaarden, 2006; Vliegenthart et al., 2008). However, research has widely ignored the possible impact of specific content dimensions of the media coverage on the recipients’ EU-related attitudes. One possible set of such content dimensions that help recipients to make sense of the complex developments during economic crises is to identify clearly the roles of the actors involved: the identification of primary victims (Who is suffering most from the crisis?) is a possibility for rendering the impact of a crisis more tangible. The attribution of a treatment responsibility (Who is responsible for solving the crisis?) points to actors potentially capable of solving the crisis. Research has shown that a responsibility frame (relying on the attribution of causal and treatment responsibility) and the victim's perspective were key interpretations in the news discourse about EU policies in general (Semetko & Valkenburg, 2000) and the media coverage about the Euro crisis in particular (Nienstedt, Kepplinger, & Quiring 2015). Additionally, such role attributions in the media coverage influence citizens’ attitudes toward political actors in general (Maestas, Atkeson, Croom, & Bryant, 2008) and the EU in particular (Hobolt & Tilley, 2014). Yet, research has mainly focused on the attribution of a causal responsibility for different political or economic issues (e.g., unemployment). In contrast to causal attributions, which imply a failure of policymakers (Hobolt & Tilley, 2014), presenting the EU as responsible for solving the economic turbulences (treatment responsibility) underscores the EU’s sovereignty to manage problematic situations and may—in consequence—strengthen citizens’ approval of sharing more sovereignty with other members of the EU. Identifying the EU as suffering from the crisis may instead draw a picture of a helpless and overcharged actor and can potentially thwart the idea of a further integration of the EU. Finally, such storylines presenting the EU as suffering from or responsible for solving the economic crisis are not coherently conveyed in different national media systems. Still today, one cannot speak of a European public sphere with pan-European mass media that are used by all or at least the majority of EU-citizens (de Vreese, 2007). Therefore, citizens almost entirely rely on the mediated representations of European issues in their respective country’s national mass media (European Commission, 2004). As the national media discourses hardly take a genuine European perspective either (Kleinen-von Königslow, 2012), the media coverage in different EU countries differs in terms of the stand on European issues (e.g., the Euro crisis) and the pattern of attributing the victim status or the treatment responsibility to the EU (Kepplinger, Koehler, & Post, 2015). Therefore, we focus in our study on the specific impact exerted by the attribution of a victim status and treatment responsibility to the EU in the national news media coverage on citizens’ prointegration attitudes: H4: Attributing the treatment responsibility in the national news media coverage to the EU will have a positive impact on citizens’ support for a deeper European integration. H5: Attributing the victim status in the national news media coverage to the EU will have a negative impact on citizens’ support for a deeper European integration. Besides the impact of the media representations of the EU and citizens’ regional economic context, research suggests several attitudinal factors on the individual level to be powerful in explaining citizens’ supportive attitudes toward the European integration. To avoid biased estimations because of omitted factors, we include predictors representing objective and subjective utilitarianism (Gabel, 1998; Gabel & Whitten, 1997), evaluations of the national political system derived from cue theory (Armingeon & Ceka, 2014) as well as general perceptions of the EU (Serricchio et al., 2013) into our analysis. Although identity-based concepts have been proven influential as well (Hooghe & Marks, 2005; McLaren, 2004), we are not able to study such aspects because of limitations of the used Eurobarometer survey. Method Data and Sampling As it is our aim to analyze the impact of the real economic development during the Euro crisis and the attribution of different roles to the EU in the national news media on citizens’ attitudes toward a deeper integration of the EU, we merged data from three different sources. We draw on (1) data from a comparative quantitative content analysis of media coverage during the Euro crisis in 10 European countries (Belgium, Finland, France, Germany, Greece, Italy, The Netherlands, Poland, Spain, and the United Kingdom) (cf., Supplementary Appendix and Picard, 2015a). The selection of countries was guided by the idea to include nations with different roles during the crisis (countries with vs. without sovereign debt problems, members vs. nonmembers of the Eurozone), as one may expect different positions of the national media coverage on crucial issues regarding the crisis because of these different roles. In each country, the coverage of four major newspapers (tabloid, quality papers, financial press1) was analyzed in the context of 11 major political and economic events (e.g., change of EU contract; Mario Monti’s appointment as Italy’s prime minister) between February 2010 and July 2012 (1 week before and after the events; total number of articles: 13,718). These events were chosen as they were closely linked to actions taken by European institutions in the course of the crisis (Picard, 2015b). The selection of newspapers was based on the leading papers of each category in each country. Although the media sample is not representative for the media landscape in the different countries under analysis, the discourse of the national opinion-forming newspapers with respect to different strata of the public (economic specialist, liberal and conservative quality press, tabloids) was captured (comparable with the procedure of Vliegenthart et al., 2008). The sample period begins when the Greek government admitted difficulties in refinancing their budget deficit on the international financial markets (Clements et al., 2014). It ends in July 2012 with the broadest spread of the crisis throughout Europe. The data on the media coverage were combined with (2) data from the Eurobarometer survey, November 2012 (EB 78.1; ZA5685; European Commission, 2015). The Eurobarometer survey is a biannual face-to-face survey on public opinion in Europe, representative for the national populations aged ≥15 years (European Commission, 2015). In total, the survey covers 10,967 participants in the 10 countries considered in the content analysis. Of those, 3,298 had to be excluded because of missing values on one or more of the variables under investigation. Hence, the final sample consisted of 7,669 participants. As the media coverage indicators were weighted by measures of individual media use, the survey and media content data comprise the individual level of our analysis. We supplemented the individual-level data with (3) macroeconomic indicators on the economic performance of the local regions of the participants, using data from Eurostat. As de Vreese (2010, p. 206) claims that “[f]uture research must address the cumulative effects of repeated exposure to consonant news framing,” we analyze the cumulative effects of the media presentation of the Euro crisis over a period of almost 1.5 years on citizens’ attitudes 4 months after the end of this sample period (see Figure 1 for the time frame of the analysis). Existing research from single-shot experiments suggests that the impact of frames may last for almost 2 weeks—although these effects lose considerable strength (cf., Lecheler & de Vreese, 2011). However, long-standing and profound attitudinal effects are supposed to be based on the cumulatively and repeatedly presented distillate of the mass media discourse (Noelle-Neumann, 1973), which has led researchers to request more realistic designs in which to study potential longstanding effects (Lecheler & de Vreese, 2011). Hence, we build our analysis on a combination of content analysis and survey data on an aggregate level to capture such cumulative influences (instead of frames in single news items). The comparably long-time lag between our content analysis data and the survey data ensures that potential applicability effects from the role attributions in the media coverage about the Euro crisis on citizens’ opinion-building have already been consolidated (Baden & Lecheler, 2012), so that we do not only capture the impact of the most recent media coverage. To assess whether media coverage from specific time periods would allow for a better prediction of the effects on citizens’ attitudes toward the European integration, we calculated alternative models containing media coverage from different time periods (cf., Supplementary Table A1). The comparison of model fits between our main effects model (cf., Table 2) and the alternative models reveals that the alternative models are not significantly better (cf., Nagelkerke’s R2 values). Table 1 Average of the Maximum Increase in Regional Unemployment Rates and the Minimum Increase in Regional GDP (2010-2012) per Country   Belgium  Finland  France  Germany  The Netherlands  Poland  The United Kingdom  Noncrisis countries  Spain  Greece  Italy  Crisis countries  N = 66 regions  n = 3  n = 1  n = 8  n = 16  n = 4  n = 6  n = 12  n = 50  n = 7  n = 4  n = 5  n = 16  Ø of max. increase in unemployment rate  0.77  0.10  0.78  -0.31  1.02  1.42  0.61  0.47  3.39  6.25  2.46  3.81  Ø of min. GDP growth rate  −0.40  −1.74  0.3  0.62  −2.08  1.10  0.17  0.19  −1.78  −8.02  −2.54  −3.58    Belgium  Finland  France  Germany  The Netherlands  Poland  The United Kingdom  Noncrisis countries  Spain  Greece  Italy  Crisis countries  N = 66 regions  n = 3  n = 1  n = 8  n = 16  n = 4  n = 6  n = 12  n = 50  n = 7  n = 4  n = 5  n = 16  Ø of max. increase in unemployment rate  0.77  0.10  0.78  -0.31  1.02  1.42  0.61  0.47  3.39  6.25  2.46  3.81  Ø of min. GDP growth rate  −0.40  −1.74  0.3  0.62  −2.08  1.10  0.17  0.19  −1.78  −8.02  −2.54  −3.58  Note. Differences between countries with versus without sovereign debt issues—in unemployment rate: t(16.54) = −7.31, p < .001; in GDP growth rate: t(17.03) = 5.321, p < .001. GDP = gross domestic product. Table 1 Average of the Maximum Increase in Regional Unemployment Rates and the Minimum Increase in Regional GDP (2010-2012) per Country   Belgium  Finland  France  Germany  The Netherlands  Poland  The United Kingdom  Noncrisis countries  Spain  Greece  Italy  Crisis countries  N = 66 regions  n = 3  n = 1  n = 8  n = 16  n = 4  n = 6  n = 12  n = 50  n = 7  n = 4  n = 5  n = 16  Ø of max. increase in unemployment rate  0.77  0.10  0.78  -0.31  1.02  1.42  0.61  0.47  3.39  6.25  2.46  3.81  Ø of min. GDP growth rate  −0.40  −1.74  0.3  0.62  −2.08  1.10  0.17  0.19  −1.78  −8.02  −2.54  −3.58    Belgium  Finland  France  Germany  The Netherlands  Poland  The United Kingdom  Noncrisis countries  Spain  Greece  Italy  Crisis countries  N = 66 regions  n = 3  n = 1  n = 8  n = 16  n = 4  n = 6  n = 12  n = 50  n = 7  n = 4  n = 5  n = 16  Ø of max. increase in unemployment rate  0.77  0.10  0.78  -0.31  1.02  1.42  0.61  0.47  3.39  6.25  2.46  3.81  Ø of min. GDP growth rate  −0.40  −1.74  0.3  0.62  −2.08  1.10  0.17  0.19  −1.78  −8.02  −2.54  −3.58  Note. Differences between countries with versus without sovereign debt issues—in unemployment rate: t(16.54) = −7.31, p < .001; in GDP growth rate: t(17.03) = 5.321, p < .001. GDP = gross domestic product. Table 2 Multilevel Models of Support for a Deeper European Integration Model 1: Main effects model  Unstandardized coefficients  Standard errors  Intercept  1.499****  (0.043)      AIC  20,814.26      Individual-level controls          Gender (1 = female)  −0.064***  (0.025)      Age  0.0003  (0.001)      Education  0.007**  (0.003)      Utilitarian evaluation: national economy  −0.033**  (0.015)      Utilitarian evaluation: Individual economic situation  0.029*  (0.016)      Satisfaction with national democracy  −0.006  (0.014)      General attitudes toward EU  0.435****  (0.030)      Δ Deviance  806.25****        AIC  20,022      Contextual macrolevel (outstanding economic development)          Max Δ unemployment (2010–2012)  0.062**  (0.031)      Min Δ GDP (2010–2012)  0.060***  (0.021)      Control: Dummy for countries with sovereign debt issues  0.280***  (0.091)      Δ Deviance  14.21***        AIC  20,013.80      Microlevel independent variables: media coverage          Responsibility: EU  0.084***  (0.029)      Victim: EU  −0.210**  (0.093)      Δ Deviance  8.77**        AIC  20,009.03        Nagelkerke’s Pseudo R2  0.107      Model 2: Alternative contextual macro level variables (most recent economic developments)          Unemployment rate in 2012  0.013***  (0.005)      GDP growth in 2012  0.050**  (0.019)      AIC  20,009.14        Nagelkerke’s Pseudo R2  0.107    Model 1: Main effects model  Unstandardized coefficients  Standard errors  Intercept  1.499****  (0.043)      AIC  20,814.26      Individual-level controls          Gender (1 = female)  −0.064***  (0.025)      Age  0.0003  (0.001)      Education  0.007**  (0.003)      Utilitarian evaluation: national economy  −0.033**  (0.015)      Utilitarian evaluation: Individual economic situation  0.029*  (0.016)      Satisfaction with national democracy  −0.006  (0.014)      General attitudes toward EU  0.435****  (0.030)      Δ Deviance  806.25****        AIC  20,022      Contextual macrolevel (outstanding economic development)          Max Δ unemployment (2010–2012)  0.062**  (0.031)      Min Δ GDP (2010–2012)  0.060***  (0.021)      Control: Dummy for countries with sovereign debt issues  0.280***  (0.091)      Δ Deviance  14.21***        AIC  20,013.80      Microlevel independent variables: media coverage          Responsibility: EU  0.084***  (0.029)      Victim: EU  −0.210**  (0.093)      Δ Deviance  8.77**        AIC  20,009.03        Nagelkerke’s Pseudo R2  0.107      Model 2: Alternative contextual macro level variables (most recent economic developments)          Unemployment rate in 2012  0.013***  (0.005)      GDP growth in 2012  0.050**  (0.019)      AIC  20,009.14        Nagelkerke’s Pseudo R2  0.107    Note. ****p < .001; ***p < .01; **p < .05; *p < .1. Test of significant difference in deviance is based on χ2 test. Robust standard errors in parenthesis; nonstandardized coefficients, FML estimation, grand-centered variables; macrolevel N = 66; microlevel N = 7,977. Multilevel models allow to calculate an R2 only approximately. Thus, we compared the different subordinated models by the more reliable measures of differences in deviances (tested by χ2) and the Akaike information criterion (AIC) (Hox, 2010). R2 values are only reported for the final models to compare different—not subordinated—models with each other. For Model 2 only, the relevant figures (most recent economic events) are displayed to keep the presentation of results clear. EU = European Union; GDP = gross domestic product. Table 2 Multilevel Models of Support for a Deeper European Integration Model 1: Main effects model  Unstandardized coefficients  Standard errors  Intercept  1.499****  (0.043)      AIC  20,814.26      Individual-level controls          Gender (1 = female)  −0.064***  (0.025)      Age  0.0003  (0.001)      Education  0.007**  (0.003)      Utilitarian evaluation: national economy  −0.033**  (0.015)      Utilitarian evaluation: Individual economic situation  0.029*  (0.016)      Satisfaction with national democracy  −0.006  (0.014)      General attitudes toward EU  0.435****  (0.030)      Δ Deviance  806.25****        AIC  20,022      Contextual macrolevel (outstanding economic development)          Max Δ unemployment (2010–2012)  0.062**  (0.031)      Min Δ GDP (2010–2012)  0.060***  (0.021)      Control: Dummy for countries with sovereign debt issues  0.280***  (0.091)      Δ Deviance  14.21***        AIC  20,013.80      Microlevel independent variables: media coverage          Responsibility: EU  0.084***  (0.029)      Victim: EU  −0.210**  (0.093)      Δ Deviance  8.77**        AIC  20,009.03        Nagelkerke’s Pseudo R2  0.107      Model 2: Alternative contextual macro level variables (most recent economic developments)          Unemployment rate in 2012  0.013***  (0.005)      GDP growth in 2012  0.050**  (0.019)      AIC  20,009.14        Nagelkerke’s Pseudo R2  0.107    Model 1: Main effects model  Unstandardized coefficients  Standard errors  Intercept  1.499****  (0.043)      AIC  20,814.26      Individual-level controls          Gender (1 = female)  −0.064***  (0.025)      Age  0.0003  (0.001)      Education  0.007**  (0.003)      Utilitarian evaluation: national economy  −0.033**  (0.015)      Utilitarian evaluation: Individual economic situation  0.029*  (0.016)      Satisfaction with national democracy  −0.006  (0.014)      General attitudes toward EU  0.435****  (0.030)      Δ Deviance  806.25****        AIC  20,022      Contextual macrolevel (outstanding economic development)          Max Δ unemployment (2010–2012)  0.062**  (0.031)      Min Δ GDP (2010–2012)  0.060***  (0.021)      Control: Dummy for countries with sovereign debt issues  0.280***  (0.091)      Δ Deviance  14.21***        AIC  20,013.80      Microlevel independent variables: media coverage          Responsibility: EU  0.084***  (0.029)      Victim: EU  −0.210**  (0.093)      Δ Deviance  8.77**        AIC  20,009.03        Nagelkerke’s Pseudo R2  0.107      Model 2: Alternative contextual macro level variables (most recent economic developments)          Unemployment rate in 2012  0.013***  (0.005)      GDP growth in 2012  0.050**  (0.019)      AIC  20,009.14        Nagelkerke’s Pseudo R2  0.107    Note. ****p < .001; ***p < .01; **p < .05; *p < .1. Test of significant difference in deviance is based on χ2 test. Robust standard errors in parenthesis; nonstandardized coefficients, FML estimation, grand-centered variables; macrolevel N = 66; microlevel N = 7,977. Multilevel models allow to calculate an R2 only approximately. Thus, we compared the different subordinated models by the more reliable measures of differences in deviances (tested by χ2) and the Akaike information criterion (AIC) (Hox, 2010). R2 values are only reported for the final models to compare different—not subordinated—models with each other. For Model 2 only, the relevant figures (most recent economic events) are displayed to keep the presentation of results clear. EU = European Union; GDP = gross domestic product. Figure 1 View largeDownload slide Time frame of analysis: sample periods of content analysis, Eurobarometer survey, and macroeconomic indicators Figure 1 View largeDownload slide Time frame of analysis: sample periods of content analysis, Eurobarometer survey, and macroeconomic indicators Dependent Measure Citizens’ diffuse support for the idea of a deeper integration of the EU in terms of a higher amount of shared sovereignty (Kopecký & Mudde, 2002) was measured by one item from the Eurobarometer survey (EB 78.1; ZA5685, QA20a), which represents the idea of strengthening the EU as a political union. “The EU should develop further into a federation of nation states” ranging from 0 = totally disagree to 3 = totally agree (M = 1.47; SD = 0.95). Contextual Macrolevel Independent Measures To investigate the effects of the real economic development during the Euro crisis on citizens’ attitudes, we used data on unemployment rates (close macroeconomic indicator) and the GDP (remote macroeconomic indicator) provided by Eurostat (2015a). As we intend to include macroeconomic indicators into our analysis that represent citizens’ scope of immediate economic experience, the economic indicators were adapted according to two specific aspects: first, we used regional economic indicators for the local regions of our participants that represent the people's immediate environment (instead of national data). Second, we identified the most pertinent development of both the unemployment rate and the GDP during the sample period to represent outstanding changes in the economic situation (highest increase/decrease in percentage as compared with the previous year in 2010, 2011, and 2012) (maximum increase in unemployment rate: M = 1.28, SD = 1.79; minimum increase of the GDP: M = −0.72, SD = 2.36). To test these indicators against the most recent economic developments in our participants’ local region, we also calculated a model containing the most recent economic measures—the unemployment rate and the GDP growth rate of 2012 (unemployment 2012: M = 10.99, SD = 6.88; GDP 2012: M = -0.38, SD = 1.98). Microlevel Independent Measures: Media Coverage Indicators The depiction of the Euro crisis in the national news media is operationalized by two broad categories (both dummy-coded): the attribution of a treatment responsibility (i.e., who is responsible for solving the crisis) and the attribution of a victim status (i.e., who is suffering from the crisis). For both the treatment responsibility and the victim status, we calculated the average percentage of articles (compared with the total number of articles) in every country that identifies the EU as suffering from/responsible for solving the crisis. Reliability for the coding in Germany was Holsti = .69 for victim status, and Holsti = .69 for treatment responsibility. Estimated reliabilities for the overall sample are Holsti = .75 for victim status, and Holsti = .82 for treatment responsibility.2 We calculated respondents’ potential exposure to the attribution of victim status and treatment responsibility to the EU in the national media, weighting the media content variables by each respondent’s individual media diet (the frequency of using the written press; EB 78.1, ZA5685, QE3: “Could you tell me to what extent you read the written press?” 0 = never; 5 = everyday; M = 3.36; SD = 1.81). The procedure is comparable with that of Desmet et al. (2015). Controls Following established concepts in political science, we included different political values as well as demographic and evaluative determinants as controls on the microlevel into our analysis. First, we incorporated respondents’ education (i.e., age at which you finished school, EB 78.1, ZA5685, D8; M = 19.2 SD = 4.64) into our analysis to control for citizens’ general sociodemographic position (objective utilitarian theory). Furthermore, citizens’ subjective utilitarian evaluations of the national economic performance and the personal economic situation were included into our analysis (i.e., “How would you judge the current situation in each of the following?” 0 = very bad; 3 = very good; national economy: M = 1.05; SD = 0.80; personal economic situation: M = 1.69; SD = 0.77; EB 78.1, ZA5685, QA3a). Citizens’ attitudes toward their national democratic system were established using the item “On the whole, are you very satisfied, fairly satisfied, not very satisfied, or not at all satisfied with the way democracy works in your country?” ranging from 0 = not at all satisfied to 3 = very satisfied (M = 1.48; SD = 0.85; EB 78.1, ZA5685, QA19a). Citizens’ general attitudes on the EU were operationalized using four items, combined into a composite index (“The interests of our country are well taken into account in the EU”; “My voice counts in the EU”; both EB 78.1, ZA5685, QA20a; “Would you say that you are very optimistic, fairly optimistic, fairly pessimistic or very pessimistic about the future of the EU?,” EB 78.1, ZA5685, QA23; “You feel you are a citizen of the EU”; M = 1.44; SD = 0.67; EB 78.1, ZA5685, QD2; α = .773). Additionally, we included participant’s age (M = 51.52; SD = 16.81; EB 78.1, ZA5685, D11) and gender (dummy coded with 1 = female, M = .5; EB 78.1, ZA5685, D10). Finally, we controlled on the macrolevel for the influence that may stem from the fact that our analysis covers regions in different countries, which occupied different roles during the Euro crisis: Some countries were notably concerned with sovereign debt problems, with social and economic consequences having been especially prevalent in countries such as Greece, Spain, or Italy (cf., Lane, 2012). Therefore, we included a dummy-coded variable representing regions in the group of countries, which were affected by the crisis (=1 for regions from Greece, Italy, or Spain, M = 0.24, SD = 0.43). Analysis The impact of the real economic indicators, of citizens’ potential exposure to the different characteristics of the media coverage, as well as the impact of the control variables on citizens’ support for a deeper European integration was analyzed by calculating hierarchical linear models using the software HLM (version 7). Random and fixed effects were estimated by full-maximum-likelihood (FML) procedure to be able to compare the model fits (Hox, 2010). The unit of analysis at the contextual macrolevel was a participant’s home region (NUTS4 1 classification of the EU, Eurostat, 2015b; N = 66). The unemployment indicators and the GDP indicators (growth rate in 2012; maximum increase between 2010 and 2012) were established on the regional level. On the individual level (microlevel), the single participants are the unit of interest. Participants were assigned their demographic and attitudinal characteristics as controls, their potential exposure to the attribution of treatment responsibility and victim status in their national media, and the dependent variable (N = 7,669) (see Figure 2 for the data structure). The models were calculated using robust standard errors for all variables centered on their grand mean (except of gender and the dummy for countries with sovereign debt problems; Hox, 2010). Figure 2 View largeDownload slide Structure of data Figure 2 View largeDownload slide Structure of data Results Descriptive Analysis Before analyzing the determinants of citizens’ support for a deeper European integration, we first conducted a descriptive analysis of the percentage of articles in the national media coverage of each country attributing the victim role and the treatment responsibility to the EU (cf., Figure 3). Figure 3 View largeDownload slide Descriptive analysis of role attributions in the national media coverage. Differences between countries with versus without sovereign debt issues—attribution of treatment responsibility: χ2(1, N = 13,718) = 1.86, p = .09; attribution of victim role: χ2(1, N = 13,718) = 26.38, p < .001; N = 13,718 articles Figure 3 View largeDownload slide Descriptive analysis of role attributions in the national media coverage. Differences between countries with versus without sovereign debt issues—attribution of treatment responsibility: χ2(1, N = 13,718) = 1.86, p = .09; attribution of victim role: χ2(1, N = 13,718) = 26.38, p < .001; N = 13,718 articles On average, 10% of the articles in all countries claimed that the EU is suffering from the crisis, whereas 21% attributed the responsibility for solving the crisis to the EU. The coverage differed significantly (for treatment responsibility only by tendency), but only marginally between countries, which were affected by the crisis as compared with those who were not affected. Nevertheless, the media coverage in those countries being affected economically and socially by the Euro crisis showed a relatively homogeneous pattern for both attributing the victim role and the treatment responsibility to the EU. The coverage in countries not concerned with sovereign debt problems was, in contrast, more heterogeneous. Additionally, we calculated the average of the maximum increase in regional unemployment rates and the minimum regional GDP growth rates (2010–2012) per country (cf., Table 1). While media coverage differed only slightly between countries with and without sovereign debt problems, the average regional economic indicators revealed pronounced gaps: on average, the regions in Greece, Italy, and Spain were more affected by increased unemployment and a reduction in the economic power as compared with their counterparts in Belgium, The Netherlands, Finland, Germany, the United Kingdom, Poland, and France. Determinants of Citizens’ Support for a Deeper European Integration In a first instance, the empty model was computed to separate the variance components of the individual versus the regional level. The intraclass correlation coefficient, calculated on the basis of the empty model (not shown in the table; ɣ00 = 1.519 (SE = 0.042); σu0 = 0.108; σe = 0.882), indicated that 11% of the total variance in participants’ support for a deeper European integration was explained by the macro level (i.e., the region—though not necessarily by the macroeconomic indicators). Hence, the region in which participants live contributed substantially to explaining their attitudes toward the European integration. Subsequently, the individual-level predictors as well as the macrolevel indicators were included into the model step by step to test, whether and to which extent the single variables contribute to the overall model fit (cf., Table 2; differences in deviance and AIC values are based on the single step-by-step models). Finally, models with random slopes were estimated to test for any possible interaction effects between the real economic indicators for participants’ local regions and the media coverage indicators. The results of this final step will not be reported because of the lack of significant cross-level interactions. Conclusively, our results do not suggest that the media effects depend on the severity of economic shocks sustained by the local region of the participants. With respect to the control variables, the results were mainly in line with previous research: gender had a significant effect, with men being more supportive of the European integration. Education had a positive impact on the support of the further integration of the EU. Age had no effect, and the impact of the egocentric utilitarian evaluation was only marginally significant, with positive evaluations of one’s own economic situation resulting in more support for a deeper integration of the EU. However, citizens’ satisfaction with the national political system had no significant effect, whereas general attitudes toward the EU had a highly significant and positive impact on their support of the further integration of the EU. The comparison of the AIC between the empty model and the model containing the control variables underscored the strong impact of these political evaluations and attitudes as well as the demographic characteristics on citizens’ prointegration attitudes. The contextual macrolevel as well as the participants’ potential exposure to the media coverage improved the model fit significantly, though not to the extent of the control variables. Impact of the Macroeconomic Indicators In H1, we assumed that pronounced macroeconomic consequences of the crisis in participants’ home region would have a negative impact on their support of a deeper European integration. The effect of the minimum increase of the regional GDP supported this assumption: participants living in a region with a moderate minimum of the regional GDP growth between 2010 and 2012 were more supportive toward the EU compared with those living in a region with a low minimum of the GDP growth rate. Yet, concerning unemployment rates, a reversed pattern revealed that the worse the economic downturn in one’s home region in terms of high unemployment between 2010 and 2012, the more supportive a person in terms of a deeper European integration (H1 partly confirmed). Although not directly connected to H1, this pattern of results could also be observed in case of the dummy for countries with sovereign debt issues: participants living in countries that have experienced social and economic disturbances were on average more supportive toward the European integration than did participants living in countries without sovereign debt issues. Also on the individual level, participants’ evaluation of their national economic performance produced a comparable effect: the worse they subjectively evaluated their national economy, the more they supported a deeper integration. The results regarding H1 already suggested that the impact of the macroeconomic context was differential with respect to close and remote macroeconomic indicators as assumed in H3. Therefore, we turn to H3 before analyzing H2. We proposed that the impact of economic indicators that can be experienced immediately in everyday live should be more pronounced than the influence of remote economic indicators. The results did not support our assumption, as the effect of the maximum increase in unemployment (close indicator) and the effect of the minimum increase in the regional GDP (remote indicator) were almost identical in strength (H3 rejected). Yet, the effects pointed at opposite directions. Finally, in H2, we assumed that sharp economic developments would better explain citizens’ attitudes toward the European integration than most recent economic developments. To test this assumption, we calculated an alternative to our main effects model (Model 2 in Table 1; only the relevant fixed effects are displayed) that contained the most recent instead of the steepest economic developments during our sample period. First, the comparison of the Nagelkerke’s R2 values indicate that Model 2, containing the most recent economic predictors, did not prove to be a better fit to our data as compared with Model 1, containing the most pertinent economic developments. Both economic indicators for the most recent development in 2012 had a significant impact but were—giving the same scaling—not as influential as the maximum increase in unemployment and the minimum increase in GDP in Model 1. Media Coverage and Citizens’ Attitudes Toward a Deeper European Integration Including the media coverage indicators into the model improved the model fit significantly. This implies that it was not solely the control variables and the macroeconomic context, but also the media coverage of the crisis in the national news media that shaped citizens’ attitudes toward the European integration. The attribution of responsibility for solving the crisis to the EU in the national news media coverage about the Euro crisis had a significant impact on citizens’ support for a deeper European integration. The direction of influence suggests that if the national media identified the EU as being responsible for solving the crisis, citizens’ support for a deeper European integration would be enhanced (supporting H4). Identifying the EU as suffering from the crisis in the national news media was also significant with respect to citizens’ prointegration attitudes. The impact of presenting the EU as a victim of the economic turbulences on citizens’ attitudes was negative (H5 confirmed). As the attribution of a treatment responsibility as well as a victim status were denominated in the same unit of analysis (percentage of articles weighted by individual media use), one can conclude that the identification of a victim status in the national news media exerted a stronger (negative) impact on citizen’s attitudes toward the European integration. Discussion and Conclusion The aim of our study was to investigate the impact of the depiction of the Euro crisis in the national news media and the macroeconomic context on citizens’ approval of a deeper European integration, while controlling for the influence of attitudinal factors. Our analysis suggests that both the media coverage indicators as well as the macroeconomic context were important in explaining citizens’ supportive attitudes toward the European integration. With respect to the impact of the macroeconomic indicators, the assumptions derived from the economic voting approach (Downs, 1957) were only supported in terms of a low minimum in the GDP growth rate. However, the fact that one lives in a region, which is located in Italy, Greece, or Spain (countries affected by sovereign debt crisis), or which was affected by pertinent increases in unemployment, enhanced the support for the concept of a deeper European integration. The same is true for a poor subjective evaluation of one’s national economic performance. These effects stood in contrast to what the economic voting approach and previous studies on supportive attitudes toward the European integration suggest. Although the effects of the unemployment rate as well as the GDP growth rate were equal in strength, the opposed directions may indicate that the impact of close and remote economic indicators has to be differentiated. From a more general perspective, these differential effects, stemming from the diverse economic indicators (subjective evaluation, role of one’s home region as part of a crisis country, close vs. remote indicator of economic performance), may suggest that citizens process and evaluate the economic conditions surrounding them by applying different schemas and heuristics (Paldam & Nannestad, 2000). This, in turn, results in different conclusions regarding whether to support the European integration. Finally, our analysis provides evidence that the specific conceptualization of macroeconomic indicators is important when studying the impact of the economic context on citizens’ attitudes on the European integration in times of economic crises. Indeed, the most recent value for the unemployment rate and the GDP growth rate had a significant effect on citizens’ attitudes toward the European integration. However, the impact of these indicators was lower as compared with the most pertinent economic developments in one’s home region. Therefore, one could conclude that economic indicators should represent citizens’ scope of economic experience in terms of sharp economic developments instead of using most recent economic developments that cannot be as easily observed or experienced in everyday life. Additionally, further research should expand its focus toward different economic phases (boom, other crises) and more economic indicators (interest rates, inflation) to identify a general pattern of influence exerted by economic developments on citizens’ supportive attitudes. With respect to the influence of the potential exposure to media attributions of different roles—the attribution of responsibility and the primary victim status to the EU—results indicate that the media impact on citizens’ attitudes is not only connected to the valence of framing but also to specific content dimensions that are presented in the media coverage about the crisis. Hence, rather than focusing on the attribution of causal responsibility only, future research should consider more diverse categories of role attributions to (collective) political actors to systematically identify the impact of different content dimensions on citizens’ political attitudes. Additionally, it may be worthwhile to differentiate such media attributions with respect to different issues. As the economic situation in general and unemployment and public debt in particular were at the top of citizens’ minds during the Euro crisis (European Commission, 2012), these specific media presentations of the EU (e.g., as capable of acting or as weak and suffering) may fall on fertile soil. However, such effects may change or differ when other issues, such as security or crime, become more salient in citizens’ perceptions. When interpreting our results, some limitations have to be considered: we assumed that the media coverage about the Euro crisis exerted a cumulative influence on citizens’ attitudes toward the EU. The underlying assumption suggested that citizens were able to perceive and process the basic messages of news media coverage over an extended period of time and applied those perceptions when evaluating and judging the EU. As our design did not permit the analysis of such dynamics over time, we were not able to ensure that the media influence was in fact based on the cumulative distillate of how the crisis was represented in the media. A time series approach or a long-term experimental design would better suit such purposes and be more appropriate to test the assumed causality of influences (Schuck, Vliegenthart, & de Vreese, 2016). Furthermore, citizens’ potential exposure to the media frame elements was calculated using the average amount of articles representing the specific content dimensions in one country weighted by citizens’ individual use of the written press. First, we cannot report the more rigid and commonly used Krippendorff’s α as an indicator of coding reliability, as the original study reports Holsti measures only (Picard, 2015a). Yet, the Holsti values for the variables as they were used in our analysis show an acceptable level. Additionally, we did not include citizens’ actual exposure to the news media coverage about the Euro crisis, but computed proxies that can be understood in terms of citizens’ potential exposure to the aggregate presentation of the Euro crisis in the national news media. A dynamic approach considering citizens’ actual media use would take into account several aspects: first, media outlets differ in the picture they draw of the crisis. Second, media use may change over time on the individual level (e.g., more intense media use in times of crises or saturation effects because of long issue carriers). This leads to a general conclusion regarding the methodological assets of our study: as we conducted a secondary analysis of already existing data, we had to choose a still appropriate, yet approximate way of investigating the phenomenon of interest. We perceive our study as an attempt of better understanding the possibilities as well as the limitations of connecting survey data with content analytic data beyond the pure amount and tone of raw news items as suggested by Schuck and colleagues (Schuck et al., 2016). Our study implies that the specific content dimensions of the media representation of EU issues as well as citizens’ immediate social environment of everyday life should not be neglected when studying the diverse antecedents of citizens’ EU-related attitudes. In times of enduring political, social, and economic turbulences, such a holistic and more diversified research approach seems to be expedient. Supplementary Data Supplementary Data are available at IJPOR online. Christina Koehler is a research assistant at the University of Mainz, Germany. Her research is mainly concerned with questions from the political communication field addressing the structure of political and economic news as well as its impact on citizens’ as well as political elites’ attitudes and behavior. Mathias Weber is a research associate at the University of Mainz, Germany. His research mainly focuses on media socialization during adolescence (social contexts of media use; uses and effects of potentially harmful media content) and economic aspects of social communication. Research has been published, for example, in the International Journal of Press/Politics, the Journal of Children and Media, and the Journal of Consumer Behavior. Oliver Quiring is professor for communication studies at the University of Mainz, Germany. His research interests include social, economic and political communication as well as integrative qualitative-quantitative methodology. Especially the effects of mediated political and economic reality on individuals and the functioning of modern societies are at the heart of his research focus. Research has been published e.g. in the International Journal of Press/Politics, Journal of Communication and the Communication Yearbook. Footnotes 1In France, Greece, Italy, and Spain, no national tabloids exist. In these cases, three national quality papers representing different editorial slants (liberal, centrist, and conservative) were included in the analysis. 2The reliability measures for the original variables established in the course of the comparative content analysis are as follows: Holsti = .60 for treatment responsibility; Holsti = .53 for victim status. However, we did not use the categories, as they were originally established, but dichotomized the variables as to represent only the differentiation between “EU as responsible/victim” and “others are responsible/victim.” A recalculation of the reliability was only possible for the German data. Reliability for attribution of victim status was improved by .21. If we use this change in reliability as a rough proxy for the overall reliabilities of this category for all countries, one could expect an improvement of Holsti to about .75. Applying the same procedure to the category attribution of treatment responsibility, a Holsti value of .82 can be expected. These new approximate inter-coder reliability values indicate an acceptable quality for the data used in this analysis. 3A principal component analysis (PCA) with oblique rotation of these four items reveals a single component solution (Kaiser-Mayer-Olkin index = .77; explained variance 59.41%; factor loadings are all above .75), which justifies the calculation of an average index. Also calculating PCAs for single countries reveals single-component solutions in all countries under analysis with factor loadings being roughly the same for all countries. Correlation between index “general attitudes towards EU” and DV “support for deeper European integration”: r = .31***. 4NUTS (Nomenclature of Territorial Units for Statistics). References Armingeon K., Ceka B. ( 2014). The loss of trust in the European Union during the great recession since 2007: The role of heuristics from the national political system. European Union Politics , 15, 82– 107. doi: 10.1177/1465116513495595 Google Scholar CrossRef Search ADS   Baden C., Lecheler S. ( 2012). Fleeting, fading, or far-reaching? A knowledge-based model of the persistence of framing effects. Communication Theory , 22, 359– 382. doi: 10.1111/j.1468-2885.2012.01413.x Google Scholar CrossRef Search ADS   Beck U., Grande E. ( 2011). Empire Europe: Statehood and political authority in the process of regional integration. In Neyer J., Wiener A. (Eds.), Political theory of the European Union  (pp. 21– 46). Oxford: Oxford University Press. Börzel T. A., Risse T. ( 2000). Who is afraid of a European Federation? How to constitutionalise a multi-level governance system. In Joerges C., Mény Y., Weiler J. H. H., (Eds.), What kind of constitution for what kind of polity  (pp. 46– 58). Florence: European University Institute. Braun D., Tausendpfund M. ( 2014). The impact of the Euro Crisis on citizens' support for the European Union. Journal of European Integration , 36, 231– 345. doi:10.1080/07036337.2014.885751 Google Scholar CrossRef Search ADS   Clements B., Nanou K., Verney S. ( 2014). ‘We no longer love you, but we don’t want to leave you': The Eurozone Crisis and popular Euroscepticism in Greece. Journal of European Integration , 36, 247– 265. doi: 10.1080/07036337.2014.885753 Google Scholar CrossRef Search ADS   de Vreese C. ( 2004). The effects of frames in political television news on issue interpretation and frame salience. Journalism and Mass Communication Quarterly , 81, 36– 52. doi: 10.1177/107769900408100104 Google Scholar CrossRef Search ADS   de Vreese C. ( 2007). The EU as a public sphere. Living Reviews in European Governance , 2. Retrieved from http://www.livingreviews.org/lreg-2007-3 de Vreese C. ( 2010). Framing the economy: Effects of journalistic news frames. In D'Angelo P., Kuypers J. A. (Eds.), Doing news framing analysis. Empirical and theoretical perspectives  (pp. 187– 214). New York, NY: Routledge. de Vreese C., Boomgaarden H. ( 2006). Media effects on public opinion about the enlargement of the European Union. Journal of Common Market Studies , 44, 419– 436. doi: 10.1111/j.1468-5965.2006.00629.x Google Scholar CrossRef Search ADS   Desmet P., van Spanje J., de Vreese C. ( 2015). Discussing the democratic deficit: effects of media and interpersonal communication on satisfaction with democracy in the European Union. International Journal of Communication , 9, 3177– 3198. Downs A. ( 1957). An economic theory of democracy . New York, NY: Harper. Easton D. ( 1965). A system analysis of political life . New York, NY: John Wiley & Sons. Easton D. ( 1975). A re-assessment of the concept of political support. British Journal of Political Science , 5, 435– 457. http://dx.doi.org/10.1017/S0007123400008309 Google Scholar CrossRef Search ADS   Eichenberg R. C., Dalton R. J. ( 2007). Post-maastricht blues: The transformation of citizen support for European integration, 1973–2004. Acta Politica , 42, 128– 152. doi: 10.1057/palgrave.ap.5500182 Google Scholar CrossRef Search ADS   Entman, R. M. ( 1993). Framing: Toward clarification of a fractured paradigm. Journal of Communication , 43, 51– 58. CrossRef Search ADS   European Commission ( 2004). Eurobarometer Spring 2004. Public Opinion on Europe. Retrieved from ec.europa.eu/public_opinion/archives/eb/eb61/eb61_en.pdf European Commission ( 2012). Public Opinion in the EU. Standard Eurobarometer 78, autumn 2012. Retrieved from http://ec.europa.eu/COMMFrontOffice/PublicOpinion/index.cfm/Survey/getSurveyDetail/yearFrom/1973/yearTo/2012/search/Eurobarometer/surveyKy/1069 European Commission ( 2015). Eurobarometer 78.1 (2012). TNS opinion, Brussels [producer]. GESIS Data Archive, Cologne. ZA5685 Data file Version 2.0.0. https://dbk.gesis.org/dbksearch/sdesc2.asp?no=5685&db=e Eurostat ( 2015a). Database. Retrieved from http://ec.europa.eu/eurostat/data/database Eurostat ( 2015b). NUTS, Retrieved from http://ec.europa.eu/eurostat/ramon/nomenclatures/index.cfm?TargetUrl=LST_NOM_DTL&StrNom=NUTS_2013&StrLanguageCode=EN&IntPcKey=&StrLayoutCode=HIERARCHIC Gabel M. ( 1998). Public support for European integration. An empirical test of five theories. Journal of Politics , 60, 333– 354. doi: 10.2307/2647912 Google Scholar CrossRef Search ADS   Gabel M., Whitten G. D. ( 1997). Economic conditions, economic perceptions, and public support for European integration. Political Behavior , 19, 81– 96. doi: 10.1023/A:1024801923824 Google Scholar CrossRef Search ADS   Galpin C., Trenz H. J. ( 2017). The spiral of Euroscepticism: Media negativity, framing and opposition to the EU. In Caiani M., Guerra S. (Eds.), Euroscepticism, democracy and the media  (pp. 49– 72). Basingstoke: Palgrave Macmillan. Google Scholar CrossRef Search ADS   Guerra S. ( 2013). Does familiarity breed contempt? Determinants of public support for European integration and opposition to it before and after accession. Journal of Common Market Studies , 51, 38– 50. doi: 10.1111/j.1468-5965.2012.02300.x Google Scholar CrossRef Search ADS   Hobolt S. B., Tilley J. ( 2014). Blaming Europe? Responsibility without accountability in the European Union . Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   Hooghe L., Marks G. ( 2005). Calculation, community and cues: Public opinion on European integration. European Union Politics , 6, 419– 443. doi: 10.1177/1465116505057816 Google Scholar CrossRef Search ADS   Hox J. J. ( 2010). Multilevel analysis: Techniques and applications . New York, NY: Routledge. Karp J. A., Bowler S. ( 2006). Broadening and deepening or broadening versus deepening: The question of enlargement and Europe's ‘hesitant Europeans’. European Journal of Political Research , 45, 369– 390. doi: 10.1111/j.1475-6765.2006.00302.x Google Scholar CrossRef Search ADS   Kepplinger H. M., Koehler C., Post S. ( 2015). Do political events change national views? In: Picard R. G. (Ed.), The Euro Crisis in the media. Journalistic coverage of economic crisis and European institutions  (pp. 45– 62). London; New York, NY: I.B. Tauris. Kleinen-von Königslow K. ( 2012). Europe in crisis? Testing the stability and explanatory factors of the Europeanization of national public spheres. International Communication Gazette , 74, 443– 463. doi: 10.1177/1748048512445153 Google Scholar CrossRef Search ADS   Kopecký P., Mudde C. ( 2002). The two sides of Euroscepticism: Party positions on European integration in East Central Europe. European Union Politics , 3, 297– 326. doi: 10.1177/1465116502003003002 Google Scholar CrossRef Search ADS   Kriesi H. ( 2014). The political consequences of the economic crisis in Europe: Electoral punishment and popular protest. In Bermeo N., Bartels L. M. (Eds.), Mass politics in tough times. Opinions, votes, and protest in the Great Recession  (pp. 297– 333). New York, NY: Oxford University Press. Google Scholar CrossRef Search ADS   Lane P. R. ( 2012). The European Sovereign Debt Crisis. The Journal of Economic Perspectives , 26, 49– 67. doi: 10.1257/jep.26.3.49 Google Scholar CrossRef Search ADS   Lecheler S., De Vreese C. H. ( 2011). Getting real: The duration of framing effects. Journal of Communication , 61, 959– 983. doi: 10.1111/j.1460-2466.2011.01580.x Google Scholar CrossRef Search ADS   Maestas C. D., Atkeson L. R., Croom T., Bryant L. A. ( 2008). Shifting the blame: Federalism, media, and public assignment of blame following Hurricane Katrina. Publius: The Journal of Federalism , 38, 609– 632. doi: 10.1093/publius/pjn021 Google Scholar CrossRef Search ADS   McLaren L. M. ( 2004). Opposition to European integration and fear of loss of national identity: Debunking a basic assumption regarding hostility to the integration project. European Journal of Political Research , 43, 895– 912. doi: 10.1111/j.0304-4130.2004.00179.x Google Scholar CrossRef Search ADS   Nienstedt H. W., Kepplinger H. M., Quiring O. ( 2015). What went wrong and why? Roots, responsibilities and solutions of the Euro Crisis in European newspapers. In: Picard R. G. (Ed.), The Euro Crisis in the media. Journalistic coverage of economic crisis and European institutions  (pp. 19– 45). London; New York, NY: I.B. Tauris. Noelle-Neumann E. ( 1973). Return to the concept of the powerful mass media. Studies in Broadcasting , 9, 67– 112. Paldam M., Nannestad P. ( 2000). What do voters know about the economy? A study of Danish data, 1990-1993. Electoral Studies , 19, 363– 391. doi: 10.1016/S0261-3794(99)00057-8 Google Scholar CrossRef Search ADS   Picard R. G. (Ed.) ( 2015a) The Euro Crisis in the media. Journalistic coverage of economic crisis and European institutions . London; New York, NY: I.B. Tauris. Picard R. G. ( 2015b). Understanding the crisis. In: Picard R. G. (Ed.), The Euro Crisis in the media. Journalistic coverage of economic crisis and European institutions  (pp. 1– 19). London; New York, NY: I.B. Tauris. Schuck A. R., Vliegenthart R., de Vreese C. ( 2016). Matching theory and data: Why combining media content with survey data matters. British Journal of Political Science , 46, 205– 213. doi: 10.1017/S0007123415000228 Google Scholar CrossRef Search ADS   Semetko H. A., Valkenburg P. M. ( 2000). Framing European politics: A content analysis of press and television news. Journal of Communication , 20, 93– 109. doi: 10.1111/j.1460-2466.2000.tb02843.x Google Scholar CrossRef Search ADS   Serricchio F., Tsakatika M., Quaglia L. ( 2013). Euroscepticism and the global financial crisis. Journal of Common Market Studies , 51, 51– 64. doi: 10.1111/j.1468-5965.2012.02299.x Google Scholar CrossRef Search ADS   Sniderman P. M., Theriault S. M. ( 2004). The dynamics of political argument and the logic of issue framing. In Saris W. E., Sniderman P. M. (Eds.), Studies in public opinion. Attitudes, nonattitudes, measurement error, and change  (pp. 133– 165). Princeton, NJ: Princeton University Press. Toshkov D. ( 2011). Public opinion and policy output in the European Union: A lost relationship. European Union Politics , 12, 169– 191. doi: 10.1177/1465116510395043 Google Scholar CrossRef Search ADS   Vliegenthart R., Schuck A. R., Boomgaarden H. G., de Vreese C. ( 2008). News coverage and support for European integration, 1990-2006. International Journal of Public Opinion Research , 20, 415– 439. doi: 10.1093/ijpor/edn044 Google Scholar CrossRef Search ADS   © The Author(s) 2018. Published by Oxford University Press on behalf of The World Association for Public Opinion Research. All rights reserved. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal of Public Opinion Research Oxford University Press

I Want a Savior, Not a Victim. The Impact of Media Representations of the EU and Economic Shocks on Citizens’ Supportive Attitudes on the European Integration

Loading next page...
 
/lp/ou_press/i-want-a-savior-not-a-victim-the-impact-of-media-representations-of-NYF7aptuIv
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of The World Association for Public Opinion Research. All rights reserved.
ISSN
0954-2892
eISSN
1471-6909
D.O.I.
10.1093/ijpor/edx023
Publisher site
See Article on Publisher Site

Abstract

Abstract The Euro crisis substantially diminished citizens’ support for the European integration. Therefore, we intend to analyze the independent effects of both the depiction of the Euro crisis in national news media and macroeconomic contexts on citizens’ support for a deeper integration. Based on a combination of Eurobarometer data, results from a content analysis, and data on the economic performance of citizens’ home regions, we show that presenting the European Union (EU) as a victim impedes citizens’ supportive attitudes concerning the European integration, whereas depicting the EU as responsible for solving the crisis has a strengthening effect. Additionally, sharp economic developments in citizens’ home regions also have significant effects on their prointegration attitudes, suggesting a differential influence of close and remote economic indicators. From the Maastricht Treaty in 1992 to the Lisbon Treaty in 2009, the development of the European Union (EU) was constantly driven toward ever-deeper integration. Public support by the citizens of the EU member countries, however, provided but a fragile basis for such far-reaching institutional transformations (Eichenberg & Dalton, 2007; Serricchio, Tsakatika, & Quaglia, 2013; Toshkov, 2011). The political and economic disruptions of the Euro crisis have further nourished public skepticism toward the EU both as a political institution and as a community of countries (European Commission, 2012; Kriesi, 2014). In explaining citizens’ attitudes concerning the European integration, research has traditionally focused on determinants such as subjective and objective utilitarianism (Gabel, 1998; Guerra, 2013), perceptions of the national political system, or citizens’ identity (Hooghe & Marks, 2005; McLaren, 2004). Recently, the mass media coverage of EU-related issues, namely, the specific valence of framing, has been shown to be crucial in explaining citizens’ EU-related thoughts (de Vreese, 2004) and attitudes (Desmet, van Spanje, & de Vreese, 2015; de Vreese & Boomgaarden, 2006; Vliegenthart, Schuck, Boomgaarden, & de Vreese, 2008). It is even argued that pronounced negativity in the media debate on EU issues may result in a downward spiral of Euroscepticism (Galpin & Trenz, 2017). Beyond the mere valence of EU-related frames, however, it can be assumed that the specific manner, in which the media characterizes the EU as an actor handling problematic situations such as the Euro crisis, will be considered by recipients when forming an opinion on the future of the European project. For example, if the media coverage presents the EU as responsible for solving the crisis, the sovereignty of the EU is emphasized. Whereas presenting the EU as passive, and its possibilities to react as constraint (i.e., as a victim of the crisis), may question its capacity to manage critical situations. In consequence, such attributions of different roles to the EU can be assumed to influence citizens’ opinions on the further development of the EU. Additionally, as the Euro crisis is mainly of an economic nature, it seems plausible to consider the macroeconomic context next to the attitudinal determinants and the media coverage as a contributor to citizens’ perceptions and evaluation of the further development of the EU. The sharp economic downturn with its serious social consequences may represent a “sudden frustration of expectations” (Easton, 1975, p. 445), which is assumed to erode even the sturdiest supportive attitudes toward a political system. Therefore, pronounced economic developments during the Euro crisis in citizen’s immediate environment (e.g., when many people in the local region lose their jobs in a short period of time) may serve as powerful predictors of citizens’ opinions on the European future in times of economic crises. Therefore, it is the aim of our study to enhance our understanding of the media coverage and economic shocks as potential determinants shaping citizens’ supportive attitudes toward the European future. We seek to investigate the independent effects of both the depiction of the EU in the national news media coverage about the Euro crisis and outstanding economic developments in citizens’ immediate everyday environment (i.e., their home region) during the crisis on public attitudes toward a further integration of the EU. At the same time, we control for the impact of attitudinal and evaluative determinants traditionally studied in the context of public support for the EU (i.e., utilitarianism, perceptions of the national and European political system). In doing so, we merge data from (1) a content analysis of the media coverage on the Euro crisis in 10 European countries, and (2) indicators of economic shocks on a regional level with (3) data from the Eurobarometer Survey (November 2012) in a multilevel analysis. Political Attitudes on Europe’s future in the aftermath of the Euro crisis Since 2007, Europe has been troubled by multilayered financial crises: First, the burst of the U.S. housing bubble and the bankruptcy of the Lehman Brothers investment bank caused severe disturbances on the international financial markets, forcing European governments to impose extensive policies to stabilize capital markets. Accordingly, the public debt of several European countries rose dangerously, resulting in these countries experiencing difficulties in refinancing their national deficit (i.e., European Sovereign Debt Crisis). To avoid the collapse of the common currency, the European community established huge bailout packages to help the countries in trouble (cf., Lane, 2012). Severe social and political consequences accompanied the economic and fiscal turmoil (Kriesi, 2014) giving rise to Eurosceptic attitudes all over Europe (Armingeon & Ceka, 2014; Braun & Tausendpfund, 2014; Clements, Nanou, & Verney, 2014). According to David Easton (1965, 1975), one may conclude that such a decline in supportive attitudes may even endanger the stability of the EU. Based on his theory of political systems, Easton (1965, 1975) claims that democratic political systems—national as well as supranational regimes such as the EU—can only be persistent in the long run, if they rest on the support of their citizens. Citizens’ supportive attitudes can be divided into specific and diffuse support (Easton, 1965, 1975). Whereas specific support is related to the short-term output of a political system and its incumbent political authorities, diffuse support is more basic and stable in nature and stems from a generalized evaluation of political systems’ vital objects. These are a sense of trust in and legitimacy for the political regime and the political authorities in general (independent of current incumbents), as well as a sense of belonging to the political community (Easton, 1965, 1975). Drawing on Easton’s general concept of support for political systems, Kopecký and Mudde (2002) conceptualize Euroscepticism in terms of diffuse and specific support toward the European integration. Whereas the first refers to the “ideas of European integration”, the latter applies to “the European Union as the current embodiment of these ideas” (Kopecký & Mudde, 2002, p. 300). Diffuse support can be deemed crucial for the persistence of a political system, as it can compensate for a short-term lack of specific support (Easton, 1965). Hence, while European policymakers can withstand a short-term lack of specific support, pronounced Eurosceptic attitudes in terms of a lack in diffuse support endanger the fundamental supportive basis for the EU’s further development. Therefore, it is the aim of our study to investigate the determinants explaining citizens’ diffuse support for the European integration in times of economic crises. European integration can be conceived in multiple terms (Karp & Bowler, 2006); hence, a further specification seems to be expedient. Similar to Easton’s general concept of trust in political regimes, which means “to believe that the regime’s goals (…) can be trusted to provide outcomes equitable for all” (Easton, 1975, p. 448), diffuse support toward the European integration implies a citizen’s approval of the general idea of sharing sovereignty between the member countries of the EU (Kopecký & Mudde, 2002). Sharing of sovereignty can be understood either in terms of broadening the European community (i.e., enlargement with more countries sharing their sovereignty) or in terms of deepening the cooperation between the member countries in different policy fields such as economic and fiscal policy (Karp & Bowler, 2006). Especially during the Euro crisis, issues such as intensifying the solidarity between the existing member countries dominated the debate among political elites and the general public. Therefore, we focus our analysis on the deepening aspect of European integration. To establish citizens’ approval of the idea of deepening the European integration, we rely on one prominent scenario for the further development of the EU: the notion of a political union (Beck & Grande, 2011). Such a development implies an integration process moving away from the ‘Confederacy of European States’, which is indicated by a nations’ almost full independence from other countries, and stepping closer together into a federation of nation states, sharing more sovereignty in vital parts of the polity (Börzel & Risse, 2000). Citizens’ diffuse support for a political system builds on the general meaning they ascribe to it and on the experience they have made with it (Easton, 1975). In the following, we shall argue how the presentation of the EU in the national media coverage and outstanding economic shocks may serve as sources for such experiences with and impressions of the European integration. Formation of Attitudes Toward the European Integration and the Economic Context Diffuse support for a political system may be based not only on socialization processes but also on experiences with its political objects. Citizens develop generalized attitudes toward the political system in terms of a “product of spill-over effects from evaluations of a series of outputs and of performance” produced by the political objects (Easton, 1975, p. 446). As the European integration was originally driven by the idea of enhancing welfare by way of economic cooperation between the EU member countries, the economic performance may well serve as a basis for citizens’ evaluation of the same. Accordingly, studies drawing on the concept of economic voting (Downs, 1957) show that the economic context in terms of the national economic performance influences citizens’ evaluation of the EU. A pronounced growth of the gross domestic product (GDP), low interest rates, low unemployment rates, and intense intra-EU trade has strengthening effects on citizens’ supportive attitudes (Armingeon & Ceka, 2014; Eichenberg & Dalton, 2007; Serricchio et al., 2013). Most of these studies draw on aggregate economic figures that represent the latest national economic status quo (e.g., most recent GDP growth rate) (Eichenberg & Dalton, 2007). However, such a conceptualization of the economic context does not reflect citizens’ scope of economic experience. Citizens cannot directly observe minor changes of the real economy, as they are displayed by slight increases or decreases of key macroeconomic figures on an aggregate level (e.g., unemployment rate, public debt rate, GDP). Yet, when the macroeconomic context manifests itself in citizens’ immediate environment (e.g., fear of job losses may come up in everyday talk, cost-cutting measures in major companies may result in wage cuts among family members or acquaintances), it may well be perceivable for ordinary citizens (Paldam & Nannestad, 2000). Therefore, especially, the economic consequences in citizens’ local region are supposedly influential. Likewise, citizens have no possibility to evaluate (in absolute terms) the status of the economic situation at one specific point in time (e.g., in 2012). However, sharp economic developments, as they occur during an economic crisis (e.g., the youth unemployment rate in Spain reached a high of 55% at the end of 2012 and Greece experienced a 9% downturn of the GDP in 2011; Eurostat, 2015a), may well leave a lasting impression (e.g., when many people lose their jobs in a brief period of time) (Armingeon & Ceka, 2014). Such sharp economic turbulences may also represent what Easton (1975, p. 445) calls “a sudden frustration of expectations” that has the potential to erode even the more stable entity of diffuse support. Therefore, we assume: H1: The more pronounced the macroeconomic consequences of the Euro crisis were at the regional level, the less supportive will a citizen living in this region be toward the European integration. H2: Outstanding economic developments in a citizen’s direct environment will have a stronger impact on this person's support for a deeper European integration as compared with the aggregate current economic development. Yet, even when relying on economic indicators located on a level that is close to a citizen’s life (i.e., the local region), some economic indicators are directly relevant to their everyday experience (e.g., unemployment, prices). On the other hand, measures of the overall growth of the companies’ output in a region (i.e., GDP) represent more remote aspects of a person's economic experience, as they do not have as great an impact on the everyday economic conditions (Paldam & Nannestad, 2000). We take these considerations into account and assume: H3: The impact of macroeconomic indicators close to citizens’ everyday experience will be more pronounced than the impact of macroeconomic indicators remote from citizens’ everyday life. National Media Coverage and Interpretations of the Crisis Via Role Attributions In its most basic sense, diffuse support “refers to evaluations of what an object is or represents—to the general meaning it has for a person” (Easton, 1975, p. 444). Such a general impression of the European integration may not only be impacted by a sudden frustration of citizens’ expectations on the economic output of the EU. Yet, also the mass media, as the central transmission mechanism between the real world and the pictures in a person's head, is supposed to shape such general impressions of the European integration by presenting European political actors or policies during the Euro crisis in a specific way (Picard, 2015b). Journalists usually try to contextualize and interpret issues using coherent storylines that help recipients make sense of these issues (Galpin & Trenz, 2017). Such basic storylines are called ‘frames’ (Entman, 1993; Sniderman & Theriault, 2004). At the core of framing theory stands the idea that journalists are never able to tell every aspect and every conceivable interpretation of an issue. Rather, they emphasize some sets of considerations while excluding or downplaying others. Consequently, only those aspects that are (frequently) referred to in the media discourse have a notable chance of gaining salience in the minds of the recipients (de Vreese, 2010; Sniderman & Theriault, 2004). Research has accumulated extensive evidence for an influence of media framing on citizens’ EU-related attitudes. In particular, the valence of news frames has been shown to guide recipients’ reasoning about EU matters (Desmet et al., 2015; de Vreese & Boomgaarden, 2006; Vliegenthart et al., 2008). However, research has widely ignored the possible impact of specific content dimensions of the media coverage on the recipients’ EU-related attitudes. One possible set of such content dimensions that help recipients to make sense of the complex developments during economic crises is to identify clearly the roles of the actors involved: the identification of primary victims (Who is suffering most from the crisis?) is a possibility for rendering the impact of a crisis more tangible. The attribution of a treatment responsibility (Who is responsible for solving the crisis?) points to actors potentially capable of solving the crisis. Research has shown that a responsibility frame (relying on the attribution of causal and treatment responsibility) and the victim's perspective were key interpretations in the news discourse about EU policies in general (Semetko & Valkenburg, 2000) and the media coverage about the Euro crisis in particular (Nienstedt, Kepplinger, & Quiring 2015). Additionally, such role attributions in the media coverage influence citizens’ attitudes toward political actors in general (Maestas, Atkeson, Croom, & Bryant, 2008) and the EU in particular (Hobolt & Tilley, 2014). Yet, research has mainly focused on the attribution of a causal responsibility for different political or economic issues (e.g., unemployment). In contrast to causal attributions, which imply a failure of policymakers (Hobolt & Tilley, 2014), presenting the EU as responsible for solving the economic turbulences (treatment responsibility) underscores the EU’s sovereignty to manage problematic situations and may—in consequence—strengthen citizens’ approval of sharing more sovereignty with other members of the EU. Identifying the EU as suffering from the crisis may instead draw a picture of a helpless and overcharged actor and can potentially thwart the idea of a further integration of the EU. Finally, such storylines presenting the EU as suffering from or responsible for solving the economic crisis are not coherently conveyed in different national media systems. Still today, one cannot speak of a European public sphere with pan-European mass media that are used by all or at least the majority of EU-citizens (de Vreese, 2007). Therefore, citizens almost entirely rely on the mediated representations of European issues in their respective country’s national mass media (European Commission, 2004). As the national media discourses hardly take a genuine European perspective either (Kleinen-von Königslow, 2012), the media coverage in different EU countries differs in terms of the stand on European issues (e.g., the Euro crisis) and the pattern of attributing the victim status or the treatment responsibility to the EU (Kepplinger, Koehler, & Post, 2015). Therefore, we focus in our study on the specific impact exerted by the attribution of a victim status and treatment responsibility to the EU in the national news media coverage on citizens’ prointegration attitudes: H4: Attributing the treatment responsibility in the national news media coverage to the EU will have a positive impact on citizens’ support for a deeper European integration. H5: Attributing the victim status in the national news media coverage to the EU will have a negative impact on citizens’ support for a deeper European integration. Besides the impact of the media representations of the EU and citizens’ regional economic context, research suggests several attitudinal factors on the individual level to be powerful in explaining citizens’ supportive attitudes toward the European integration. To avoid biased estimations because of omitted factors, we include predictors representing objective and subjective utilitarianism (Gabel, 1998; Gabel & Whitten, 1997), evaluations of the national political system derived from cue theory (Armingeon & Ceka, 2014) as well as general perceptions of the EU (Serricchio et al., 2013) into our analysis. Although identity-based concepts have been proven influential as well (Hooghe & Marks, 2005; McLaren, 2004), we are not able to study such aspects because of limitations of the used Eurobarometer survey. Method Data and Sampling As it is our aim to analyze the impact of the real economic development during the Euro crisis and the attribution of different roles to the EU in the national news media on citizens’ attitudes toward a deeper integration of the EU, we merged data from three different sources. We draw on (1) data from a comparative quantitative content analysis of media coverage during the Euro crisis in 10 European countries (Belgium, Finland, France, Germany, Greece, Italy, The Netherlands, Poland, Spain, and the United Kingdom) (cf., Supplementary Appendix and Picard, 2015a). The selection of countries was guided by the idea to include nations with different roles during the crisis (countries with vs. without sovereign debt problems, members vs. nonmembers of the Eurozone), as one may expect different positions of the national media coverage on crucial issues regarding the crisis because of these different roles. In each country, the coverage of four major newspapers (tabloid, quality papers, financial press1) was analyzed in the context of 11 major political and economic events (e.g., change of EU contract; Mario Monti’s appointment as Italy’s prime minister) between February 2010 and July 2012 (1 week before and after the events; total number of articles: 13,718). These events were chosen as they were closely linked to actions taken by European institutions in the course of the crisis (Picard, 2015b). The selection of newspapers was based on the leading papers of each category in each country. Although the media sample is not representative for the media landscape in the different countries under analysis, the discourse of the national opinion-forming newspapers with respect to different strata of the public (economic specialist, liberal and conservative quality press, tabloids) was captured (comparable with the procedure of Vliegenthart et al., 2008). The sample period begins when the Greek government admitted difficulties in refinancing their budget deficit on the international financial markets (Clements et al., 2014). It ends in July 2012 with the broadest spread of the crisis throughout Europe. The data on the media coverage were combined with (2) data from the Eurobarometer survey, November 2012 (EB 78.1; ZA5685; European Commission, 2015). The Eurobarometer survey is a biannual face-to-face survey on public opinion in Europe, representative for the national populations aged ≥15 years (European Commission, 2015). In total, the survey covers 10,967 participants in the 10 countries considered in the content analysis. Of those, 3,298 had to be excluded because of missing values on one or more of the variables under investigation. Hence, the final sample consisted of 7,669 participants. As the media coverage indicators were weighted by measures of individual media use, the survey and media content data comprise the individual level of our analysis. We supplemented the individual-level data with (3) macroeconomic indicators on the economic performance of the local regions of the participants, using data from Eurostat. As de Vreese (2010, p. 206) claims that “[f]uture research must address the cumulative effects of repeated exposure to consonant news framing,” we analyze the cumulative effects of the media presentation of the Euro crisis over a period of almost 1.5 years on citizens’ attitudes 4 months after the end of this sample period (see Figure 1 for the time frame of the analysis). Existing research from single-shot experiments suggests that the impact of frames may last for almost 2 weeks—although these effects lose considerable strength (cf., Lecheler & de Vreese, 2011). However, long-standing and profound attitudinal effects are supposed to be based on the cumulatively and repeatedly presented distillate of the mass media discourse (Noelle-Neumann, 1973), which has led researchers to request more realistic designs in which to study potential longstanding effects (Lecheler & de Vreese, 2011). Hence, we build our analysis on a combination of content analysis and survey data on an aggregate level to capture such cumulative influences (instead of frames in single news items). The comparably long-time lag between our content analysis data and the survey data ensures that potential applicability effects from the role attributions in the media coverage about the Euro crisis on citizens’ opinion-building have already been consolidated (Baden & Lecheler, 2012), so that we do not only capture the impact of the most recent media coverage. To assess whether media coverage from specific time periods would allow for a better prediction of the effects on citizens’ attitudes toward the European integration, we calculated alternative models containing media coverage from different time periods (cf., Supplementary Table A1). The comparison of model fits between our main effects model (cf., Table 2) and the alternative models reveals that the alternative models are not significantly better (cf., Nagelkerke’s R2 values). Table 1 Average of the Maximum Increase in Regional Unemployment Rates and the Minimum Increase in Regional GDP (2010-2012) per Country   Belgium  Finland  France  Germany  The Netherlands  Poland  The United Kingdom  Noncrisis countries  Spain  Greece  Italy  Crisis countries  N = 66 regions  n = 3  n = 1  n = 8  n = 16  n = 4  n = 6  n = 12  n = 50  n = 7  n = 4  n = 5  n = 16  Ø of max. increase in unemployment rate  0.77  0.10  0.78  -0.31  1.02  1.42  0.61  0.47  3.39  6.25  2.46  3.81  Ø of min. GDP growth rate  −0.40  −1.74  0.3  0.62  −2.08  1.10  0.17  0.19  −1.78  −8.02  −2.54  −3.58    Belgium  Finland  France  Germany  The Netherlands  Poland  The United Kingdom  Noncrisis countries  Spain  Greece  Italy  Crisis countries  N = 66 regions  n = 3  n = 1  n = 8  n = 16  n = 4  n = 6  n = 12  n = 50  n = 7  n = 4  n = 5  n = 16  Ø of max. increase in unemployment rate  0.77  0.10  0.78  -0.31  1.02  1.42  0.61  0.47  3.39  6.25  2.46  3.81  Ø of min. GDP growth rate  −0.40  −1.74  0.3  0.62  −2.08  1.10  0.17  0.19  −1.78  −8.02  −2.54  −3.58  Note. Differences between countries with versus without sovereign debt issues—in unemployment rate: t(16.54) = −7.31, p < .001; in GDP growth rate: t(17.03) = 5.321, p < .001. GDP = gross domestic product. Table 1 Average of the Maximum Increase in Regional Unemployment Rates and the Minimum Increase in Regional GDP (2010-2012) per Country   Belgium  Finland  France  Germany  The Netherlands  Poland  The United Kingdom  Noncrisis countries  Spain  Greece  Italy  Crisis countries  N = 66 regions  n = 3  n = 1  n = 8  n = 16  n = 4  n = 6  n = 12  n = 50  n = 7  n = 4  n = 5  n = 16  Ø of max. increase in unemployment rate  0.77  0.10  0.78  -0.31  1.02  1.42  0.61  0.47  3.39  6.25  2.46  3.81  Ø of min. GDP growth rate  −0.40  −1.74  0.3  0.62  −2.08  1.10  0.17  0.19  −1.78  −8.02  −2.54  −3.58    Belgium  Finland  France  Germany  The Netherlands  Poland  The United Kingdom  Noncrisis countries  Spain  Greece  Italy  Crisis countries  N = 66 regions  n = 3  n = 1  n = 8  n = 16  n = 4  n = 6  n = 12  n = 50  n = 7  n = 4  n = 5  n = 16  Ø of max. increase in unemployment rate  0.77  0.10  0.78  -0.31  1.02  1.42  0.61  0.47  3.39  6.25  2.46  3.81  Ø of min. GDP growth rate  −0.40  −1.74  0.3  0.62  −2.08  1.10  0.17  0.19  −1.78  −8.02  −2.54  −3.58  Note. Differences between countries with versus without sovereign debt issues—in unemployment rate: t(16.54) = −7.31, p < .001; in GDP growth rate: t(17.03) = 5.321, p < .001. GDP = gross domestic product. Table 2 Multilevel Models of Support for a Deeper European Integration Model 1: Main effects model  Unstandardized coefficients  Standard errors  Intercept  1.499****  (0.043)      AIC  20,814.26      Individual-level controls          Gender (1 = female)  −0.064***  (0.025)      Age  0.0003  (0.001)      Education  0.007**  (0.003)      Utilitarian evaluation: national economy  −0.033**  (0.015)      Utilitarian evaluation: Individual economic situation  0.029*  (0.016)      Satisfaction with national democracy  −0.006  (0.014)      General attitudes toward EU  0.435****  (0.030)      Δ Deviance  806.25****        AIC  20,022      Contextual macrolevel (outstanding economic development)          Max Δ unemployment (2010–2012)  0.062**  (0.031)      Min Δ GDP (2010–2012)  0.060***  (0.021)      Control: Dummy for countries with sovereign debt issues  0.280***  (0.091)      Δ Deviance  14.21***        AIC  20,013.80      Microlevel independent variables: media coverage          Responsibility: EU  0.084***  (0.029)      Victim: EU  −0.210**  (0.093)      Δ Deviance  8.77**        AIC  20,009.03        Nagelkerke’s Pseudo R2  0.107      Model 2: Alternative contextual macro level variables (most recent economic developments)          Unemployment rate in 2012  0.013***  (0.005)      GDP growth in 2012  0.050**  (0.019)      AIC  20,009.14        Nagelkerke’s Pseudo R2  0.107    Model 1: Main effects model  Unstandardized coefficients  Standard errors  Intercept  1.499****  (0.043)      AIC  20,814.26      Individual-level controls          Gender (1 = female)  −0.064***  (0.025)      Age  0.0003  (0.001)      Education  0.007**  (0.003)      Utilitarian evaluation: national economy  −0.033**  (0.015)      Utilitarian evaluation: Individual economic situation  0.029*  (0.016)      Satisfaction with national democracy  −0.006  (0.014)      General attitudes toward EU  0.435****  (0.030)      Δ Deviance  806.25****        AIC  20,022      Contextual macrolevel (outstanding economic development)          Max Δ unemployment (2010–2012)  0.062**  (0.031)      Min Δ GDP (2010–2012)  0.060***  (0.021)      Control: Dummy for countries with sovereign debt issues  0.280***  (0.091)      Δ Deviance  14.21***        AIC  20,013.80      Microlevel independent variables: media coverage          Responsibility: EU  0.084***  (0.029)      Victim: EU  −0.210**  (0.093)      Δ Deviance  8.77**        AIC  20,009.03        Nagelkerke’s Pseudo R2  0.107      Model 2: Alternative contextual macro level variables (most recent economic developments)          Unemployment rate in 2012  0.013***  (0.005)      GDP growth in 2012  0.050**  (0.019)      AIC  20,009.14        Nagelkerke’s Pseudo R2  0.107    Note. ****p < .001; ***p < .01; **p < .05; *p < .1. Test of significant difference in deviance is based on χ2 test. Robust standard errors in parenthesis; nonstandardized coefficients, FML estimation, grand-centered variables; macrolevel N = 66; microlevel N = 7,977. Multilevel models allow to calculate an R2 only approximately. Thus, we compared the different subordinated models by the more reliable measures of differences in deviances (tested by χ2) and the Akaike information criterion (AIC) (Hox, 2010). R2 values are only reported for the final models to compare different—not subordinated—models with each other. For Model 2 only, the relevant figures (most recent economic events) are displayed to keep the presentation of results clear. EU = European Union; GDP = gross domestic product. Table 2 Multilevel Models of Support for a Deeper European Integration Model 1: Main effects model  Unstandardized coefficients  Standard errors  Intercept  1.499****  (0.043)      AIC  20,814.26      Individual-level controls          Gender (1 = female)  −0.064***  (0.025)      Age  0.0003  (0.001)      Education  0.007**  (0.003)      Utilitarian evaluation: national economy  −0.033**  (0.015)      Utilitarian evaluation: Individual economic situation  0.029*  (0.016)      Satisfaction with national democracy  −0.006  (0.014)      General attitudes toward EU  0.435****  (0.030)      Δ Deviance  806.25****        AIC  20,022      Contextual macrolevel (outstanding economic development)          Max Δ unemployment (2010–2012)  0.062**  (0.031)      Min Δ GDP (2010–2012)  0.060***  (0.021)      Control: Dummy for countries with sovereign debt issues  0.280***  (0.091)      Δ Deviance  14.21***        AIC  20,013.80      Microlevel independent variables: media coverage          Responsibility: EU  0.084***  (0.029)      Victim: EU  −0.210**  (0.093)      Δ Deviance  8.77**        AIC  20,009.03        Nagelkerke’s Pseudo R2  0.107      Model 2: Alternative contextual macro level variables (most recent economic developments)          Unemployment rate in 2012  0.013***  (0.005)      GDP growth in 2012  0.050**  (0.019)      AIC  20,009.14        Nagelkerke’s Pseudo R2  0.107    Model 1: Main effects model  Unstandardized coefficients  Standard errors  Intercept  1.499****  (0.043)      AIC  20,814.26      Individual-level controls          Gender (1 = female)  −0.064***  (0.025)      Age  0.0003  (0.001)      Education  0.007**  (0.003)      Utilitarian evaluation: national economy  −0.033**  (0.015)      Utilitarian evaluation: Individual economic situation  0.029*  (0.016)      Satisfaction with national democracy  −0.006  (0.014)      General attitudes toward EU  0.435****  (0.030)      Δ Deviance  806.25****        AIC  20,022      Contextual macrolevel (outstanding economic development)          Max Δ unemployment (2010–2012)  0.062**  (0.031)      Min Δ GDP (2010–2012)  0.060***  (0.021)      Control: Dummy for countries with sovereign debt issues  0.280***  (0.091)      Δ Deviance  14.21***        AIC  20,013.80      Microlevel independent variables: media coverage          Responsibility: EU  0.084***  (0.029)      Victim: EU  −0.210**  (0.093)      Δ Deviance  8.77**        AIC  20,009.03        Nagelkerke’s Pseudo R2  0.107      Model 2: Alternative contextual macro level variables (most recent economic developments)          Unemployment rate in 2012  0.013***  (0.005)      GDP growth in 2012  0.050**  (0.019)      AIC  20,009.14        Nagelkerke’s Pseudo R2  0.107    Note. ****p < .001; ***p < .01; **p < .05; *p < .1. Test of significant difference in deviance is based on χ2 test. Robust standard errors in parenthesis; nonstandardized coefficients, FML estimation, grand-centered variables; macrolevel N = 66; microlevel N = 7,977. Multilevel models allow to calculate an R2 only approximately. Thus, we compared the different subordinated models by the more reliable measures of differences in deviances (tested by χ2) and the Akaike information criterion (AIC) (Hox, 2010). R2 values are only reported for the final models to compare different—not subordinated—models with each other. For Model 2 only, the relevant figures (most recent economic events) are displayed to keep the presentation of results clear. EU = European Union; GDP = gross domestic product. Figure 1 View largeDownload slide Time frame of analysis: sample periods of content analysis, Eurobarometer survey, and macroeconomic indicators Figure 1 View largeDownload slide Time frame of analysis: sample periods of content analysis, Eurobarometer survey, and macroeconomic indicators Dependent Measure Citizens’ diffuse support for the idea of a deeper integration of the EU in terms of a higher amount of shared sovereignty (Kopecký & Mudde, 2002) was measured by one item from the Eurobarometer survey (EB 78.1; ZA5685, QA20a), which represents the idea of strengthening the EU as a political union. “The EU should develop further into a federation of nation states” ranging from 0 = totally disagree to 3 = totally agree (M = 1.47; SD = 0.95). Contextual Macrolevel Independent Measures To investigate the effects of the real economic development during the Euro crisis on citizens’ attitudes, we used data on unemployment rates (close macroeconomic indicator) and the GDP (remote macroeconomic indicator) provided by Eurostat (2015a). As we intend to include macroeconomic indicators into our analysis that represent citizens’ scope of immediate economic experience, the economic indicators were adapted according to two specific aspects: first, we used regional economic indicators for the local regions of our participants that represent the people's immediate environment (instead of national data). Second, we identified the most pertinent development of both the unemployment rate and the GDP during the sample period to represent outstanding changes in the economic situation (highest increase/decrease in percentage as compared with the previous year in 2010, 2011, and 2012) (maximum increase in unemployment rate: M = 1.28, SD = 1.79; minimum increase of the GDP: M = −0.72, SD = 2.36). To test these indicators against the most recent economic developments in our participants’ local region, we also calculated a model containing the most recent economic measures—the unemployment rate and the GDP growth rate of 2012 (unemployment 2012: M = 10.99, SD = 6.88; GDP 2012: M = -0.38, SD = 1.98). Microlevel Independent Measures: Media Coverage Indicators The depiction of the Euro crisis in the national news media is operationalized by two broad categories (both dummy-coded): the attribution of a treatment responsibility (i.e., who is responsible for solving the crisis) and the attribution of a victim status (i.e., who is suffering from the crisis). For both the treatment responsibility and the victim status, we calculated the average percentage of articles (compared with the total number of articles) in every country that identifies the EU as suffering from/responsible for solving the crisis. Reliability for the coding in Germany was Holsti = .69 for victim status, and Holsti = .69 for treatment responsibility. Estimated reliabilities for the overall sample are Holsti = .75 for victim status, and Holsti = .82 for treatment responsibility.2 We calculated respondents’ potential exposure to the attribution of victim status and treatment responsibility to the EU in the national media, weighting the media content variables by each respondent’s individual media diet (the frequency of using the written press; EB 78.1, ZA5685, QE3: “Could you tell me to what extent you read the written press?” 0 = never; 5 = everyday; M = 3.36; SD = 1.81). The procedure is comparable with that of Desmet et al. (2015). Controls Following established concepts in political science, we included different political values as well as demographic and evaluative determinants as controls on the microlevel into our analysis. First, we incorporated respondents’ education (i.e., age at which you finished school, EB 78.1, ZA5685, D8; M = 19.2 SD = 4.64) into our analysis to control for citizens’ general sociodemographic position (objective utilitarian theory). Furthermore, citizens’ subjective utilitarian evaluations of the national economic performance and the personal economic situation were included into our analysis (i.e., “How would you judge the current situation in each of the following?” 0 = very bad; 3 = very good; national economy: M = 1.05; SD = 0.80; personal economic situation: M = 1.69; SD = 0.77; EB 78.1, ZA5685, QA3a). Citizens’ attitudes toward their national democratic system were established using the item “On the whole, are you very satisfied, fairly satisfied, not very satisfied, or not at all satisfied with the way democracy works in your country?” ranging from 0 = not at all satisfied to 3 = very satisfied (M = 1.48; SD = 0.85; EB 78.1, ZA5685, QA19a). Citizens’ general attitudes on the EU were operationalized using four items, combined into a composite index (“The interests of our country are well taken into account in the EU”; “My voice counts in the EU”; both EB 78.1, ZA5685, QA20a; “Would you say that you are very optimistic, fairly optimistic, fairly pessimistic or very pessimistic about the future of the EU?,” EB 78.1, ZA5685, QA23; “You feel you are a citizen of the EU”; M = 1.44; SD = 0.67; EB 78.1, ZA5685, QD2; α = .773). Additionally, we included participant’s age (M = 51.52; SD = 16.81; EB 78.1, ZA5685, D11) and gender (dummy coded with 1 = female, M = .5; EB 78.1, ZA5685, D10). Finally, we controlled on the macrolevel for the influence that may stem from the fact that our analysis covers regions in different countries, which occupied different roles during the Euro crisis: Some countries were notably concerned with sovereign debt problems, with social and economic consequences having been especially prevalent in countries such as Greece, Spain, or Italy (cf., Lane, 2012). Therefore, we included a dummy-coded variable representing regions in the group of countries, which were affected by the crisis (=1 for regions from Greece, Italy, or Spain, M = 0.24, SD = 0.43). Analysis The impact of the real economic indicators, of citizens’ potential exposure to the different characteristics of the media coverage, as well as the impact of the control variables on citizens’ support for a deeper European integration was analyzed by calculating hierarchical linear models using the software HLM (version 7). Random and fixed effects were estimated by full-maximum-likelihood (FML) procedure to be able to compare the model fits (Hox, 2010). The unit of analysis at the contextual macrolevel was a participant’s home region (NUTS4 1 classification of the EU, Eurostat, 2015b; N = 66). The unemployment indicators and the GDP indicators (growth rate in 2012; maximum increase between 2010 and 2012) were established on the regional level. On the individual level (microlevel), the single participants are the unit of interest. Participants were assigned their demographic and attitudinal characteristics as controls, their potential exposure to the attribution of treatment responsibility and victim status in their national media, and the dependent variable (N = 7,669) (see Figure 2 for the data structure). The models were calculated using robust standard errors for all variables centered on their grand mean (except of gender and the dummy for countries with sovereign debt problems; Hox, 2010). Figure 2 View largeDownload slide Structure of data Figure 2 View largeDownload slide Structure of data Results Descriptive Analysis Before analyzing the determinants of citizens’ support for a deeper European integration, we first conducted a descriptive analysis of the percentage of articles in the national media coverage of each country attributing the victim role and the treatment responsibility to the EU (cf., Figure 3). Figure 3 View largeDownload slide Descriptive analysis of role attributions in the national media coverage. Differences between countries with versus without sovereign debt issues—attribution of treatment responsibility: χ2(1, N = 13,718) = 1.86, p = .09; attribution of victim role: χ2(1, N = 13,718) = 26.38, p < .001; N = 13,718 articles Figure 3 View largeDownload slide Descriptive analysis of role attributions in the national media coverage. Differences between countries with versus without sovereign debt issues—attribution of treatment responsibility: χ2(1, N = 13,718) = 1.86, p = .09; attribution of victim role: χ2(1, N = 13,718) = 26.38, p < .001; N = 13,718 articles On average, 10% of the articles in all countries claimed that the EU is suffering from the crisis, whereas 21% attributed the responsibility for solving the crisis to the EU. The coverage differed significantly (for treatment responsibility only by tendency), but only marginally between countries, which were affected by the crisis as compared with those who were not affected. Nevertheless, the media coverage in those countries being affected economically and socially by the Euro crisis showed a relatively homogeneous pattern for both attributing the victim role and the treatment responsibility to the EU. The coverage in countries not concerned with sovereign debt problems was, in contrast, more heterogeneous. Additionally, we calculated the average of the maximum increase in regional unemployment rates and the minimum regional GDP growth rates (2010–2012) per country (cf., Table 1). While media coverage differed only slightly between countries with and without sovereign debt problems, the average regional economic indicators revealed pronounced gaps: on average, the regions in Greece, Italy, and Spain were more affected by increased unemployment and a reduction in the economic power as compared with their counterparts in Belgium, The Netherlands, Finland, Germany, the United Kingdom, Poland, and France. Determinants of Citizens’ Support for a Deeper European Integration In a first instance, the empty model was computed to separate the variance components of the individual versus the regional level. The intraclass correlation coefficient, calculated on the basis of the empty model (not shown in the table; ɣ00 = 1.519 (SE = 0.042); σu0 = 0.108; σe = 0.882), indicated that 11% of the total variance in participants’ support for a deeper European integration was explained by the macro level (i.e., the region—though not necessarily by the macroeconomic indicators). Hence, the region in which participants live contributed substantially to explaining their attitudes toward the European integration. Subsequently, the individual-level predictors as well as the macrolevel indicators were included into the model step by step to test, whether and to which extent the single variables contribute to the overall model fit (cf., Table 2; differences in deviance and AIC values are based on the single step-by-step models). Finally, models with random slopes were estimated to test for any possible interaction effects between the real economic indicators for participants’ local regions and the media coverage indicators. The results of this final step will not be reported because of the lack of significant cross-level interactions. Conclusively, our results do not suggest that the media effects depend on the severity of economic shocks sustained by the local region of the participants. With respect to the control variables, the results were mainly in line with previous research: gender had a significant effect, with men being more supportive of the European integration. Education had a positive impact on the support of the further integration of the EU. Age had no effect, and the impact of the egocentric utilitarian evaluation was only marginally significant, with positive evaluations of one’s own economic situation resulting in more support for a deeper integration of the EU. However, citizens’ satisfaction with the national political system had no significant effect, whereas general attitudes toward the EU had a highly significant and positive impact on their support of the further integration of the EU. The comparison of the AIC between the empty model and the model containing the control variables underscored the strong impact of these political evaluations and attitudes as well as the demographic characteristics on citizens’ prointegration attitudes. The contextual macrolevel as well as the participants’ potential exposure to the media coverage improved the model fit significantly, though not to the extent of the control variables. Impact of the Macroeconomic Indicators In H1, we assumed that pronounced macroeconomic consequences of the crisis in participants’ home region would have a negative impact on their support of a deeper European integration. The effect of the minimum increase of the regional GDP supported this assumption: participants living in a region with a moderate minimum of the regional GDP growth between 2010 and 2012 were more supportive toward the EU compared with those living in a region with a low minimum of the GDP growth rate. Yet, concerning unemployment rates, a reversed pattern revealed that the worse the economic downturn in one’s home region in terms of high unemployment between 2010 and 2012, the more supportive a person in terms of a deeper European integration (H1 partly confirmed). Although not directly connected to H1, this pattern of results could also be observed in case of the dummy for countries with sovereign debt issues: participants living in countries that have experienced social and economic disturbances were on average more supportive toward the European integration than did participants living in countries without sovereign debt issues. Also on the individual level, participants’ evaluation of their national economic performance produced a comparable effect: the worse they subjectively evaluated their national economy, the more they supported a deeper integration. The results regarding H1 already suggested that the impact of the macroeconomic context was differential with respect to close and remote macroeconomic indicators as assumed in H3. Therefore, we turn to H3 before analyzing H2. We proposed that the impact of economic indicators that can be experienced immediately in everyday live should be more pronounced than the influence of remote economic indicators. The results did not support our assumption, as the effect of the maximum increase in unemployment (close indicator) and the effect of the minimum increase in the regional GDP (remote indicator) were almost identical in strength (H3 rejected). Yet, the effects pointed at opposite directions. Finally, in H2, we assumed that sharp economic developments would better explain citizens’ attitudes toward the European integration than most recent economic developments. To test this assumption, we calculated an alternative to our main effects model (Model 2 in Table 1; only the relevant fixed effects are displayed) that contained the most recent instead of the steepest economic developments during our sample period. First, the comparison of the Nagelkerke’s R2 values indicate that Model 2, containing the most recent economic predictors, did not prove to be a better fit to our data as compared with Model 1, containing the most pertinent economic developments. Both economic indicators for the most recent development in 2012 had a significant impact but were—giving the same scaling—not as influential as the maximum increase in unemployment and the minimum increase in GDP in Model 1. Media Coverage and Citizens’ Attitudes Toward a Deeper European Integration Including the media coverage indicators into the model improved the model fit significantly. This implies that it was not solely the control variables and the macroeconomic context, but also the media coverage of the crisis in the national news media that shaped citizens’ attitudes toward the European integration. The attribution of responsibility for solving the crisis to the EU in the national news media coverage about the Euro crisis had a significant impact on citizens’ support for a deeper European integration. The direction of influence suggests that if the national media identified the EU as being responsible for solving the crisis, citizens’ support for a deeper European integration would be enhanced (supporting H4). Identifying the EU as suffering from the crisis in the national news media was also significant with respect to citizens’ prointegration attitudes. The impact of presenting the EU as a victim of the economic turbulences on citizens’ attitudes was negative (H5 confirmed). As the attribution of a treatment responsibility as well as a victim status were denominated in the same unit of analysis (percentage of articles weighted by individual media use), one can conclude that the identification of a victim status in the national news media exerted a stronger (negative) impact on citizen’s attitudes toward the European integration. Discussion and Conclusion The aim of our study was to investigate the impact of the depiction of the Euro crisis in the national news media and the macroeconomic context on citizens’ approval of a deeper European integration, while controlling for the influence of attitudinal factors. Our analysis suggests that both the media coverage indicators as well as the macroeconomic context were important in explaining citizens’ supportive attitudes toward the European integration. With respect to the impact of the macroeconomic indicators, the assumptions derived from the economic voting approach (Downs, 1957) were only supported in terms of a low minimum in the GDP growth rate. However, the fact that one lives in a region, which is located in Italy, Greece, or Spain (countries affected by sovereign debt crisis), or which was affected by pertinent increases in unemployment, enhanced the support for the concept of a deeper European integration. The same is true for a poor subjective evaluation of one’s national economic performance. These effects stood in contrast to what the economic voting approach and previous studies on supportive attitudes toward the European integration suggest. Although the effects of the unemployment rate as well as the GDP growth rate were equal in strength, the opposed directions may indicate that the impact of close and remote economic indicators has to be differentiated. From a more general perspective, these differential effects, stemming from the diverse economic indicators (subjective evaluation, role of one’s home region as part of a crisis country, close vs. remote indicator of economic performance), may suggest that citizens process and evaluate the economic conditions surrounding them by applying different schemas and heuristics (Paldam & Nannestad, 2000). This, in turn, results in different conclusions regarding whether to support the European integration. Finally, our analysis provides evidence that the specific conceptualization of macroeconomic indicators is important when studying the impact of the economic context on citizens’ attitudes on the European integration in times of economic crises. Indeed, the most recent value for the unemployment rate and the GDP growth rate had a significant effect on citizens’ attitudes toward the European integration. However, the impact of these indicators was lower as compared with the most pertinent economic developments in one’s home region. Therefore, one could conclude that economic indicators should represent citizens’ scope of economic experience in terms of sharp economic developments instead of using most recent economic developments that cannot be as easily observed or experienced in everyday life. Additionally, further research should expand its focus toward different economic phases (boom, other crises) and more economic indicators (interest rates, inflation) to identify a general pattern of influence exerted by economic developments on citizens’ supportive attitudes. With respect to the influence of the potential exposure to media attributions of different roles—the attribution of responsibility and the primary victim status to the EU—results indicate that the media impact on citizens’ attitudes is not only connected to the valence of framing but also to specific content dimensions that are presented in the media coverage about the crisis. Hence, rather than focusing on the attribution of causal responsibility only, future research should consider more diverse categories of role attributions to (collective) political actors to systematically identify the impact of different content dimensions on citizens’ political attitudes. Additionally, it may be worthwhile to differentiate such media attributions with respect to different issues. As the economic situation in general and unemployment and public debt in particular were at the top of citizens’ minds during the Euro crisis (European Commission, 2012), these specific media presentations of the EU (e.g., as capable of acting or as weak and suffering) may fall on fertile soil. However, such effects may change or differ when other issues, such as security or crime, become more salient in citizens’ perceptions. When interpreting our results, some limitations have to be considered: we assumed that the media coverage about the Euro crisis exerted a cumulative influence on citizens’ attitudes toward the EU. The underlying assumption suggested that citizens were able to perceive and process the basic messages of news media coverage over an extended period of time and applied those perceptions when evaluating and judging the EU. As our design did not permit the analysis of such dynamics over time, we were not able to ensure that the media influence was in fact based on the cumulative distillate of how the crisis was represented in the media. A time series approach or a long-term experimental design would better suit such purposes and be more appropriate to test the assumed causality of influences (Schuck, Vliegenthart, & de Vreese, 2016). Furthermore, citizens’ potential exposure to the media frame elements was calculated using the average amount of articles representing the specific content dimensions in one country weighted by citizens’ individual use of the written press. First, we cannot report the more rigid and commonly used Krippendorff’s α as an indicator of coding reliability, as the original study reports Holsti measures only (Picard, 2015a). Yet, the Holsti values for the variables as they were used in our analysis show an acceptable level. Additionally, we did not include citizens’ actual exposure to the news media coverage about the Euro crisis, but computed proxies that can be understood in terms of citizens’ potential exposure to the aggregate presentation of the Euro crisis in the national news media. A dynamic approach considering citizens’ actual media use would take into account several aspects: first, media outlets differ in the picture they draw of the crisis. Second, media use may change over time on the individual level (e.g., more intense media use in times of crises or saturation effects because of long issue carriers). This leads to a general conclusion regarding the methodological assets of our study: as we conducted a secondary analysis of already existing data, we had to choose a still appropriate, yet approximate way of investigating the phenomenon of interest. We perceive our study as an attempt of better understanding the possibilities as well as the limitations of connecting survey data with content analytic data beyond the pure amount and tone of raw news items as suggested by Schuck and colleagues (Schuck et al., 2016). Our study implies that the specific content dimensions of the media representation of EU issues as well as citizens’ immediate social environment of everyday life should not be neglected when studying the diverse antecedents of citizens’ EU-related attitudes. In times of enduring political, social, and economic turbulences, such a holistic and more diversified research approach seems to be expedient. Supplementary Data Supplementary Data are available at IJPOR online. Christina Koehler is a research assistant at the University of Mainz, Germany. Her research is mainly concerned with questions from the political communication field addressing the structure of political and economic news as well as its impact on citizens’ as well as political elites’ attitudes and behavior. Mathias Weber is a research associate at the University of Mainz, Germany. His research mainly focuses on media socialization during adolescence (social contexts of media use; uses and effects of potentially harmful media content) and economic aspects of social communication. Research has been published, for example, in the International Journal of Press/Politics, the Journal of Children and Media, and the Journal of Consumer Behavior. Oliver Quiring is professor for communication studies at the University of Mainz, Germany. His research interests include social, economic and political communication as well as integrative qualitative-quantitative methodology. Especially the effects of mediated political and economic reality on individuals and the functioning of modern societies are at the heart of his research focus. Research has been published e.g. in the International Journal of Press/Politics, Journal of Communication and the Communication Yearbook. Footnotes 1In France, Greece, Italy, and Spain, no national tabloids exist. In these cases, three national quality papers representing different editorial slants (liberal, centrist, and conservative) were included in the analysis. 2The reliability measures for the original variables established in the course of the comparative content analysis are as follows: Holsti = .60 for treatment responsibility; Holsti = .53 for victim status. However, we did not use the categories, as they were originally established, but dichotomized the variables as to represent only the differentiation between “EU as responsible/victim” and “others are responsible/victim.” A recalculation of the reliability was only possible for the German data. Reliability for attribution of victim status was improved by .21. If we use this change in reliability as a rough proxy for the overall reliabilities of this category for all countries, one could expect an improvement of Holsti to about .75. Applying the same procedure to the category attribution of treatment responsibility, a Holsti value of .82 can be expected. These new approximate inter-coder reliability values indicate an acceptable quality for the data used in this analysis. 3A principal component analysis (PCA) with oblique rotation of these four items reveals a single component solution (Kaiser-Mayer-Olkin index = .77; explained variance 59.41%; factor loadings are all above .75), which justifies the calculation of an average index. Also calculating PCAs for single countries reveals single-component solutions in all countries under analysis with factor loadings being roughly the same for all countries. Correlation between index “general attitudes towards EU” and DV “support for deeper European integration”: r = .31***. 4NUTS (Nomenclature of Territorial Units for Statistics). References Armingeon K., Ceka B. ( 2014). The loss of trust in the European Union during the great recession since 2007: The role of heuristics from the national political system. European Union Politics , 15, 82– 107. doi: 10.1177/1465116513495595 Google Scholar CrossRef Search ADS   Baden C., Lecheler S. ( 2012). Fleeting, fading, or far-reaching? A knowledge-based model of the persistence of framing effects. Communication Theory , 22, 359– 382. doi: 10.1111/j.1468-2885.2012.01413.x Google Scholar CrossRef Search ADS   Beck U., Grande E. ( 2011). Empire Europe: Statehood and political authority in the process of regional integration. In Neyer J., Wiener A. (Eds.), Political theory of the European Union  (pp. 21– 46). Oxford: Oxford University Press. Börzel T. A., Risse T. ( 2000). Who is afraid of a European Federation? How to constitutionalise a multi-level governance system. In Joerges C., Mény Y., Weiler J. H. H., (Eds.), What kind of constitution for what kind of polity  (pp. 46– 58). Florence: European University Institute. Braun D., Tausendpfund M. ( 2014). The impact of the Euro Crisis on citizens' support for the European Union. Journal of European Integration , 36, 231– 345. doi:10.1080/07036337.2014.885751 Google Scholar CrossRef Search ADS   Clements B., Nanou K., Verney S. ( 2014). ‘We no longer love you, but we don’t want to leave you': The Eurozone Crisis and popular Euroscepticism in Greece. Journal of European Integration , 36, 247– 265. doi: 10.1080/07036337.2014.885753 Google Scholar CrossRef Search ADS   de Vreese C. ( 2004). The effects of frames in political television news on issue interpretation and frame salience. Journalism and Mass Communication Quarterly , 81, 36– 52. doi: 10.1177/107769900408100104 Google Scholar CrossRef Search ADS   de Vreese C. ( 2007). The EU as a public sphere. Living Reviews in European Governance , 2. Retrieved from http://www.livingreviews.org/lreg-2007-3 de Vreese C. ( 2010). Framing the economy: Effects of journalistic news frames. In D'Angelo P., Kuypers J. A. (Eds.), Doing news framing analysis. Empirical and theoretical perspectives  (pp. 187– 214). New York, NY: Routledge. de Vreese C., Boomgaarden H. ( 2006). Media effects on public opinion about the enlargement of the European Union. Journal of Common Market Studies , 44, 419– 436. doi: 10.1111/j.1468-5965.2006.00629.x Google Scholar CrossRef Search ADS   Desmet P., van Spanje J., de Vreese C. ( 2015). Discussing the democratic deficit: effects of media and interpersonal communication on satisfaction with democracy in the European Union. International Journal of Communication , 9, 3177– 3198. Downs A. ( 1957). An economic theory of democracy . New York, NY: Harper. Easton D. ( 1965). A system analysis of political life . New York, NY: John Wiley & Sons. Easton D. ( 1975). A re-assessment of the concept of political support. British Journal of Political Science , 5, 435– 457. http://dx.doi.org/10.1017/S0007123400008309 Google Scholar CrossRef Search ADS   Eichenberg R. C., Dalton R. J. ( 2007). Post-maastricht blues: The transformation of citizen support for European integration, 1973–2004. Acta Politica , 42, 128– 152. doi: 10.1057/palgrave.ap.5500182 Google Scholar CrossRef Search ADS   Entman, R. M. ( 1993). Framing: Toward clarification of a fractured paradigm. Journal of Communication , 43, 51– 58. CrossRef Search ADS   European Commission ( 2004). Eurobarometer Spring 2004. Public Opinion on Europe. Retrieved from ec.europa.eu/public_opinion/archives/eb/eb61/eb61_en.pdf European Commission ( 2012). Public Opinion in the EU. Standard Eurobarometer 78, autumn 2012. Retrieved from http://ec.europa.eu/COMMFrontOffice/PublicOpinion/index.cfm/Survey/getSurveyDetail/yearFrom/1973/yearTo/2012/search/Eurobarometer/surveyKy/1069 European Commission ( 2015). Eurobarometer 78.1 (2012). TNS opinion, Brussels [producer]. GESIS Data Archive, Cologne. ZA5685 Data file Version 2.0.0. https://dbk.gesis.org/dbksearch/sdesc2.asp?no=5685&db=e Eurostat ( 2015a). Database. Retrieved from http://ec.europa.eu/eurostat/data/database Eurostat ( 2015b). NUTS, Retrieved from http://ec.europa.eu/eurostat/ramon/nomenclatures/index.cfm?TargetUrl=LST_NOM_DTL&StrNom=NUTS_2013&StrLanguageCode=EN&IntPcKey=&StrLayoutCode=HIERARCHIC Gabel M. ( 1998). Public support for European integration. An empirical test of five theories. Journal of Politics , 60, 333– 354. doi: 10.2307/2647912 Google Scholar CrossRef Search ADS   Gabel M., Whitten G. D. ( 1997). Economic conditions, economic perceptions, and public support for European integration. Political Behavior , 19, 81– 96. doi: 10.1023/A:1024801923824 Google Scholar CrossRef Search ADS   Galpin C., Trenz H. J. ( 2017). The spiral of Euroscepticism: Media negativity, framing and opposition to the EU. In Caiani M., Guerra S. (Eds.), Euroscepticism, democracy and the media  (pp. 49– 72). Basingstoke: Palgrave Macmillan. Google Scholar CrossRef Search ADS   Guerra S. ( 2013). Does familiarity breed contempt? Determinants of public support for European integration and opposition to it before and after accession. Journal of Common Market Studies , 51, 38– 50. doi: 10.1111/j.1468-5965.2012.02300.x Google Scholar CrossRef Search ADS   Hobolt S. B., Tilley J. ( 2014). Blaming Europe? Responsibility without accountability in the European Union . Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   Hooghe L., Marks G. ( 2005). Calculation, community and cues: Public opinion on European integration. European Union Politics , 6, 419– 443. doi: 10.1177/1465116505057816 Google Scholar CrossRef Search ADS   Hox J. J. ( 2010). Multilevel analysis: Techniques and applications . New York, NY: Routledge. Karp J. A., Bowler S. ( 2006). Broadening and deepening or broadening versus deepening: The question of enlargement and Europe's ‘hesitant Europeans’. European Journal of Political Research , 45, 369– 390. doi: 10.1111/j.1475-6765.2006.00302.x Google Scholar CrossRef Search ADS   Kepplinger H. M., Koehler C., Post S. ( 2015). Do political events change national views? In: Picard R. G. (Ed.), The Euro Crisis in the media. Journalistic coverage of economic crisis and European institutions  (pp. 45– 62). London; New York, NY: I.B. Tauris. Kleinen-von Königslow K. ( 2012). Europe in crisis? Testing the stability and explanatory factors of the Europeanization of national public spheres. International Communication Gazette , 74, 443– 463. doi: 10.1177/1748048512445153 Google Scholar CrossRef Search ADS   Kopecký P., Mudde C. ( 2002). The two sides of Euroscepticism: Party positions on European integration in East Central Europe. European Union Politics , 3, 297– 326. doi: 10.1177/1465116502003003002 Google Scholar CrossRef Search ADS   Kriesi H. ( 2014). The political consequences of the economic crisis in Europe: Electoral punishment and popular protest. In Bermeo N., Bartels L. M. (Eds.), Mass politics in tough times. Opinions, votes, and protest in the Great Recession  (pp. 297– 333). New York, NY: Oxford University Press. Google Scholar CrossRef Search ADS   Lane P. R. ( 2012). The European Sovereign Debt Crisis. The Journal of Economic Perspectives , 26, 49– 67. doi: 10.1257/jep.26.3.49 Google Scholar CrossRef Search ADS   Lecheler S., De Vreese C. H. ( 2011). Getting real: The duration of framing effects. Journal of Communication , 61, 959– 983. doi: 10.1111/j.1460-2466.2011.01580.x Google Scholar CrossRef Search ADS   Maestas C. D., Atkeson L. R., Croom T., Bryant L. A. ( 2008). Shifting the blame: Federalism, media, and public assignment of blame following Hurricane Katrina. Publius: The Journal of Federalism , 38, 609– 632. doi: 10.1093/publius/pjn021 Google Scholar CrossRef Search ADS   McLaren L. M. ( 2004). Opposition to European integration and fear of loss of national identity: Debunking a basic assumption regarding hostility to the integration project. European Journal of Political Research , 43, 895– 912. doi: 10.1111/j.0304-4130.2004.00179.x Google Scholar CrossRef Search ADS   Nienstedt H. W., Kepplinger H. M., Quiring O. ( 2015). What went wrong and why? Roots, responsibilities and solutions of the Euro Crisis in European newspapers. In: Picard R. G. (Ed.), The Euro Crisis in the media. Journalistic coverage of economic crisis and European institutions  (pp. 19– 45). London; New York, NY: I.B. Tauris. Noelle-Neumann E. ( 1973). Return to the concept of the powerful mass media. Studies in Broadcasting , 9, 67– 112. Paldam M., Nannestad P. ( 2000). What do voters know about the economy? A study of Danish data, 1990-1993. Electoral Studies , 19, 363– 391. doi: 10.1016/S0261-3794(99)00057-8 Google Scholar CrossRef Search ADS   Picard R. G. (Ed.) ( 2015a) The Euro Crisis in the media. Journalistic coverage of economic crisis and European institutions . London; New York, NY: I.B. Tauris. Picard R. G. ( 2015b). Understanding the crisis. In: Picard R. G. (Ed.), The Euro Crisis in the media. Journalistic coverage of economic crisis and European institutions  (pp. 1– 19). London; New York, NY: I.B. Tauris. Schuck A. R., Vliegenthart R., de Vreese C. ( 2016). Matching theory and data: Why combining media content with survey data matters. British Journal of Political Science , 46, 205– 213. doi: 10.1017/S0007123415000228 Google Scholar CrossRef Search ADS   Semetko H. A., Valkenburg P. M. ( 2000). Framing European politics: A content analysis of press and television news. Journal of Communication , 20, 93– 109. doi: 10.1111/j.1460-2466.2000.tb02843.x Google Scholar CrossRef Search ADS   Serricchio F., Tsakatika M., Quaglia L. ( 2013). Euroscepticism and the global financial crisis. Journal of Common Market Studies , 51, 51– 64. doi: 10.1111/j.1468-5965.2012.02299.x Google Scholar CrossRef Search ADS   Sniderman P. M., Theriault S. M. ( 2004). The dynamics of political argument and the logic of issue framing. In Saris W. E., Sniderman P. M. (Eds.), Studies in public opinion. Attitudes, nonattitudes, measurement error, and change  (pp. 133– 165). Princeton, NJ: Princeton University Press. Toshkov D. ( 2011). Public opinion and policy output in the European Union: A lost relationship. European Union Politics , 12, 169– 191. doi: 10.1177/1465116510395043 Google Scholar CrossRef Search ADS   Vliegenthart R., Schuck A. R., Boomgaarden H. G., de Vreese C. ( 2008). News coverage and support for European integration, 1990-2006. International Journal of Public Opinion Research , 20, 415– 439. doi: 10.1093/ijpor/edn044 Google Scholar CrossRef Search ADS   © The Author(s) 2018. Published by Oxford University Press on behalf of The World Association for Public Opinion Research. All rights reserved.

Journal

International Journal of Public Opinion ResearchOxford University Press

Published: Feb 12, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off