Periodic Reporting for period 4 - EUROPINIONS (Causes and consequences of change in public opinion about Europe)
Reporting period: 2020-03-01 to 2020-08-31
The EUROPINIONS project conceptualizes citizens’ attitudes towards the European Union as multi-dimensional. Our conceptualization of EU attitudes builds on the PI’s work with colleagues on EU attitudes (Boomgaarden et al., 2011; de Vreese et al., 2018). Our research conclusions can be organized along three lines of research – with the red thread being the nuanced conceptualization of EU attitudes.
The first group of studies focuses on the effects of media and political information on EU attitudes. The panel set-up of our data collection allowed us to zoom in on specific events and to isolate the effects of these events at the micro-level. We found that learning about an EU summit specifically affects evaluations of EU performance (one of the five dimensions), and that learning about an EU-level political malpractice affair negatively impacts trust in EU institutions – while leaving evaluations of national institutions unaffected. Crucially, these studies show that effects of knowledge gains can have a key impact on the specific institutions they relate to. We also conducted an experiment on EU advertising on social media which shows how the impact of different types of political information is moderated by citizens’ emotions towards the EU. The results of this study underscore that emotional reactions to political EU-messages on social media are conditional upon respondents’ own political orientation and the post’s position.
The second group of studies focuses on voting behaviour and political competition in relation to EU attitude dimensions. During the data-collection phase of the project, we took advantage of existing datasets to assess the effect of different EU attitude dimensions on voting behaviour. On the basis of the Europinions panel survey data, we continued this line of research by studying 1. The impact of the five EU dimensions on voting in the 2019 EP elections across 10 countries, and 2. The structure of party and citizen-level polarization on EU issues. Finally, we have looked into the question how campaign engagement during the EP2019 campaign impacts turnout in these EP elections, thereby connecting the first and second lines of research in the project. The findings from our panel study in the context of the 2019 elections show that unidimensional categorizations along voters’ degree of engagement with campaign information are insufficient; instead, types of media formats, discussions and content matter for effects.
The third line of research looked at the complexity of EU attitudes from a different angle: namely by studying citizens’ future EU preferences. Starting from open answers to a question on what respondents want for the future of Europe, we developed a set of eight future EU scenarios which were subsequently presented to the respondents of the Europinions panel survey. By having the respondents rank the scenarios, we were able to obtain a detailed image of citizens’ preferences for the future of the EU (both in the Netherlands and in a five-country study). The key take-away is, again, that citizens have much more nuanced views on the EU than a simple yes-no divide would suggest. Rather than wanting to leave the Union, citizens are equally spread across different intermediate scenarios of integration, which are explainable from their ideological preferences as well as from country differences.
Finally, next to the core focal points of the Europinions project, we used the flexibility of the panel survey for advancing other pressing research questions. As the debate about mis- and disinformation took off in Europe, both in policy circles and in politics, we developed a scale for measuring perceptions of dis- and misinformation, and validated these scales across ten countries.
The first group of studies focuses on the effects of media and political information on EU attitudes. The panel set-up of our data collection allowed us to zoom in on specific events and to isolate the effects of these events at the micro-level. We found that learning about an EU summit specifically affects evaluations of EU performance (one of the five dimensions), and that learning about an EU-level political malpractice affair negatively impacts trust in EU institutions – while leaving evaluations of national institutions unaffected. Crucially, these studies show that effects of knowledge gains can have a key impact on the specific institutions they relate to. We also conducted an experiment on EU advertising on social media which shows how the impact of different types of political information is moderated by citizens’ emotions towards the EU. The results of this study underscore that emotional reactions to political EU-messages on social media are conditional upon respondents’ own political orientation and the post’s position.
The second group of studies focuses on voting behaviour and political competition in relation to EU attitude dimensions. During the data-collection phase of the project, we took advantage of existing datasets to assess the effect of different EU attitude dimensions on voting behaviour. On the basis of the Europinions panel survey data, we continued this line of research by studying 1. The impact of the five EU dimensions on voting in the 2019 EP elections across 10 countries, and 2. The structure of party and citizen-level polarization on EU issues. Finally, we have looked into the question how campaign engagement during the EP2019 campaign impacts turnout in these EP elections, thereby connecting the first and second lines of research in the project. The findings from our panel study in the context of the 2019 elections show that unidimensional categorizations along voters’ degree of engagement with campaign information are insufficient; instead, types of media formats, discussions and content matter for effects.
The third line of research looked at the complexity of EU attitudes from a different angle: namely by studying citizens’ future EU preferences. Starting from open answers to a question on what respondents want for the future of Europe, we developed a set of eight future EU scenarios which were subsequently presented to the respondents of the Europinions panel survey. By having the respondents rank the scenarios, we were able to obtain a detailed image of citizens’ preferences for the future of the EU (both in the Netherlands and in a five-country study). The key take-away is, again, that citizens have much more nuanced views on the EU than a simple yes-no divide would suggest. Rather than wanting to leave the Union, citizens are equally spread across different intermediate scenarios of integration, which are explainable from their ideological preferences as well as from country differences.
Finally, next to the core focal points of the Europinions project, we used the flexibility of the panel survey for advancing other pressing research questions. As the debate about mis- and disinformation took off in Europe, both in policy circles and in politics, we developed a scale for measuring perceptions of dis- and misinformation, and validated these scales across ten countries.
In the project we performed the work as planned and expanded activities when possible and needed.
We completed the empirical work (including multi-wave surveys, flash surveys, content analyses, and a series of experiments). The outcomes are summarised above and more is coming in manuscripts currently under review.
The PhD student completed her dissertation on time and is now a post doc. The post doc completed their work and have successfully taken up new positions.
Project team members were active in conferences and workshops. Project events, updates, conference participation etc was communicated via social media (each team member has a Twitter handle).
Major milestones were listed via this website: https://www.polcomm.org/research/erc-europinions/
Team members appeared in several news items relaying information from the project.
We completed the empirical work (including multi-wave surveys, flash surveys, content analyses, and a series of experiments). The outcomes are summarised above and more is coming in manuscripts currently under review.
The PhD student completed her dissertation on time and is now a post doc. The post doc completed their work and have successfully taken up new positions.
Project team members were active in conferences and workshops. Project events, updates, conference participation etc was communicated via social media (each team member has a Twitter handle).
Major milestones were listed via this website: https://www.polcomm.org/research/erc-europinions/
Team members appeared in several news items relaying information from the project.
In addition to the substantive achievements addressed above, we also moved beyond the state of the art in the design. The Europinions project used a multi-method research design with various data sources collected in the course of the project. At the core was a large-scale survey data collection across ten EU member countries and including up to seven waves covering a time of almost two years. This enabled not only a cross-sectional analysis, but also the detection of changes in European citizens’ attitudes over time and in response to ongoing political events. As a second main data source we conducted an extensive manual analysis of media content from newspapers, TV news, and online news in the same ten European countries, for which we trained 37 coders, who were all native speakers of the respective languages. Both the survey and manual content analysis were designed to enable a linkage between the two sources, that is, to link respondents’ news diet with respective media content to enable a better understanding of the media’s influence on EU public opinion.
We complemented these main components with various additional methods:
We collected newspaper articles about the EU and its institutions in twelve countries, from 2000-2018. This newspaper coverage was systemized and analysed using automated methods of content analysis. For more short-term effects, we conducted flash surveys immediately after key events. In order to establish causal effects in more rigorous ways, we also ran various experiments to examine the influence of mediated information for individuals’ attitudes and behavioral intentions. The latter experimental studies are particularly innovative as they take a longitudinal perspective and investigate effects over time.
Furthermore, we designed a new so-called scenario-ranking approach to examine respondents’ preferences for the future of the EU. Such new types of survey variables required innovative and less common methods to analyse the data in the right way, in this case for instance the running of a multidimensional unfolding analysis (Goldberg, van Elsas & de Vreese, 2020).
Finally, we combined our substantial interest in the topic with methodological challenges such as the order of survey questions. We incorporated this in our research by testing the reliability and validity of commonly used survey questions about institutional trust by conducting a number of survey experiments, in which we varied the order and closeness of questions about national and EU institutions. We found that the levels of reported trust can, to an extent, indeed depend on questionnaire context (Brosius et al, 2020).
We complemented these main components with various additional methods:
We collected newspaper articles about the EU and its institutions in twelve countries, from 2000-2018. This newspaper coverage was systemized and analysed using automated methods of content analysis. For more short-term effects, we conducted flash surveys immediately after key events. In order to establish causal effects in more rigorous ways, we also ran various experiments to examine the influence of mediated information for individuals’ attitudes and behavioral intentions. The latter experimental studies are particularly innovative as they take a longitudinal perspective and investigate effects over time.
Furthermore, we designed a new so-called scenario-ranking approach to examine respondents’ preferences for the future of the EU. Such new types of survey variables required innovative and less common methods to analyse the data in the right way, in this case for instance the running of a multidimensional unfolding analysis (Goldberg, van Elsas & de Vreese, 2020).
Finally, we combined our substantial interest in the topic with methodological challenges such as the order of survey questions. We incorporated this in our research by testing the reliability and validity of commonly used survey questions about institutional trust by conducting a number of survey experiments, in which we varied the order and closeness of questions about national and EU institutions. We found that the levels of reported trust can, to an extent, indeed depend on questionnaire context (Brosius et al, 2020).