Abstract
This article experimentally examines how content cues politicising the European Union (EU) influence citizens’ EU attitudes, assessing the pessimist and optimist top-down models of politicisation. The former approach posits that politicisation emphasises criticism of the EU and weakens citizens’ EU support. The latter model asserts that politicisation fosters greater EU support when focusing on the EU's positive aspects. Results align with the pessimist model: cues with a negative EU tone reduce EU support. Conversely, the optimist model is not backed: cues with a positive tone have no effect on EU attitudes. Moreover, cues also interact with individuals’ national vs. European identity and interest in politics. Overall, the results suggest an asymmetrical impact of politicisation, with messages critical of the EU having a greater influence on citizens’ attitudes.
Keywords
Introduction
It has become a truism to note the politicisation of the European Union (EU). As a recent study succinctly puts it, ‘Scholars agree that the EU has become politicised’ (Anders, 2021: 175). This reflects an increased political salience and contestation of European integration, influencing citizens’ attitudes towards the EU. In its initial decades, the EU was largely ignored by the mass electorate (Lindberg and Scheingold, 1970: 42), making European integration a ‘non-issue’ (Hooghe and Marks, 2009: 6). The Maastricht Treaty is seen as bringing increased politicisation of the EU, fostering not only greater public attention but also more sceptical attitudes towards European integration (Hooghe and Marks, 2009).
The effects of this politicisation percolate to broader EU dynamics. Citizens’ initial disengagement from the EU was seen as generating a ‘permissive consensus’ for European integration, with political elites insulated from citizen pressures (Lindberg and Scheingold, 1970: 42). The greater public contestation over the EU since the 1990s is seen as shifting this to a ‘constraining dissensus’, with elites restricted by greater domestic scrutiny over European integration (Hooghe and Marks, 2009).
A key claim in the literature on EU politicisation concerns the role of top-down (or supply-side) factors in shaping citizens’ attitudes towards the EU, with public opinion being ‘particularly susceptible to construction’ and ‘cueing’ regarding European integration (Hooghe and Marks, 2009: 13). In this perspective, politicisation is instigated by political actors and entrepreneurs who articulate conflicts in the public debate, with the cues they provide fostering greater mass attention to the EU (Grande and Hutter, 2016; Wiesner, 2021: 36). These cues may emphasise criticism of the EU (pessimist politicisation) or its benefits (optimist politicisation), with potentially distinct implications for how citizens evaluate European integration.
Within this literature, one relatively underexplored aspect is the role of content (or message) cues in shaping citizens’ attitudes towards the EU. Content cues refer to the information or arguments in a message. In the EU context, their role in top-down politicisation has received limited attention. This article contributes to this field, addressing the following research question: How do content cues regarding the EU affect citizens’ attitudes towards the EU?
To answer this question, we use a survey experiment with a representative sample of the Portuguese population. While there is extensive research on how parties politicise and provide cues on EU issues (Hutter and Kriesi, 2019), there is a relative absence of experimental causal assessments of how political cues shape citizens’ attitudes towards the EU (Nielsen, 2016). In this study, participants were exposed to billboards featuring political messages with different tone and scope regarding the EU. These simulate real-world campaign materials and present short slogans – the type of political message that citizens may encounter in a campaign. To assess the effect of content cues, the materials include only message content without any source attribution.
Overall, we find that content cues influence attitudes towards the EU, but in an asymmetrical way. Negative tone cues significantly reduce support for the EU, lending support to the pessimist top-down model of politicisation. By contrast, positive tone cues do not significantly affect attitudes towards the EU, providing no empirical support for the optimist model. This effect is stronger when cues emphasise the EU's impact on the member-state (domestic scope), though negative messages referring to Europe as a whole (supranational scope) also have a statistically significant impact. In addition, these effects are moderated by respondents’ characteristics: those with a stronger national identity and lower interest in politics are more susceptible to negative content cues.
Politicisation, cueing and attitudes towards the EU
Politicisation is a multilevel concept: one that has both ‘a deep critical theoretical and even philosophical meaning’ but also refers to ‘concrete acts that can be usefully measured in empirical research’ (Wood, 2016: 527). For this author, the study of politicisation involves three levels: the macro-level, which focuses on its philosophical and theoretical underpinnings; the meso-level, which considers how politicisation can be analysed and, finally, the micro-level, which concerns its empirical assessment. These levels complement each other, yielding diverse but ‘equally legitimate’ perspectives regarding politicisation (Wood, 2016: 531–532).
At the macro level, there is consensus that politicisation entails bringing an issue or object into the realm of politics. Thus, Zürn (2014: 50) defines politicisation as ‘the demand for or the act of transporting an issue into the field of politics – making previously unpolitical matters political’. Grande and Hutter (2016) characterise politicisation as bringing a matter into the scope of political conflict; and Wiesner (2022: 23) conceives it as ‘the act of marking an issue as political within a polity’.
As Wiesner (2022) notes, these descriptions belie an additional question: what is meant by politics, with implications at the meso, conceptual, level. In this article, we adopt the widely used definition of Grande and Hutter (2016), which draws on Schattschneider's (1960) view that conflict is at the heart of politics. From this, the authors identify three central elements in politicisation: issue salience, actor expansion and polarisation. Thus, an issue is politicised when it is raised within the public debate (salience) by a widening range of elite actors (expansion) that present rival perspectives on it (polarisation), with salience as the core element (Grande and Hutter, 2016: 7–10).
This perspective puts the key emphasis on top-down politicisation. As the authors put it: ‘If an issue is not debated in public, it can only be politicised to a very limited extent – if at all’ (Grande and Hutter, 2016: 8). This is not to say that politicisation cannot occur within the ‘private’ sphere, in a bottom-up direction (Wiesner, 2022: 34). Indeed, as Hay (2007: 79–82) notes, politicisation can occur in different spheres, namely the private, public and governmental levels. This perspective of Hay acknowledges politicisation within the private sphere, as Wiesner argues; however, it also recognises that there is greater politicisation in the public and governmental spheres, consistent with Grande and Hutter.
When we turn to the micro-level empirical assessment of politicisation, a key driver of politicisation is cueing. A cue can be defined as ‘a piece of information that allows individuals to make inferences without drawing on more detailed knowledge’ (Druckman et al., 2010: 137). Cues thus provide information shortcuts, allowing citizens to draw complex inferences and navigate modern politics (Druckman and Lupia, 2016: 16). They play important roles in individuals’ political decision-making, affecting how they vote, form issue attitudes and process information (Conroy-Krutz et al., 2016). At the same time, it is important to note that cueing is not a feature of the information itself, but rather a function of how individuals interpret it. The same information may be processed heuristically by some, while others may engage with it more systematically (Chaiken, 1980).
A focus in the literature on EU politicisation concerns how this process is driven by political cues (Hooghe and Marks, 2005). Political cues can take on different forms (for an overview, see Hooghe and Marks, 2005: 428–429). One key subtype pertains to party cues (Hooghe, 2007; Hooghe and Marks, 2005). Echoing the meso-level debates on politicisation, this can entail bottom-up cueing, with voters’ orientations creating incentives for parties to take specific positions on the EU; and top-down cueing, with parties providing cues that citizens then follow (Hooghe, 2007). The existing empirical literature provides fairly strong support for top-down party cueing (Duarte, 2024; Maier et al., 2012; Pannico, 2020; Torcal et al., 2018), even if this can coexist with bottom-up dynamics (Steenbergen et al., 2007). As Carrieri (2021: 34) summarises, the EU's complexity makes voters particularly reliant on party cues.
The bulk of the literature on the micro-level assessment of politicisation in general, and on the effects of top-down cues on citizens’ attitudes towards the EU, has relied on observational data. However, as Nielsen (2016) notes, there is a considerable margin to explore these questions through experimental methods, complementing observational studies by assessing causality more directly and overcoming endogeneity issues.
This article contributes to the emerging, yet comparatively scant, experimental research on how top-down cues politicising European integration can affect citizen attitudes towards the EU. Part of this literature has assessed the effect of party cues in shaping attitudes towards EU policies (Pannico, 2017), playing a larger role in shaping citizens’ opinions and issue competence on EU policies than ideological cues (Torcal et al., 2018).
There is also some experimental research on how cues influence attitudes towards the EU. Maier et al. (2012) assess the effect of partisan cues and of two thematic issue cues, identity and economy, on attitudes towards the EU. They find an effect of negative economic cues, as well a moderating effect of identifying with the party that is providing the cues. More recent experimental research by the same authors examines the effects of source cues on EU attitudes. Overall, they find that Eurosceptic messages with a mainstream right-wing party as the source have a greater effect on EU attitudes than the same messages with a radical-right party as the source, notably among the in-partisans of the former; and that, for out-partisans, this effect is mediated by the perceived credibility of the message (Maier et al., 2017).
The existing experimental research thus points to the importance of cues in shaping attitudes towards the EU. At the same time, there remains an important gap in this literature, which this article addresses, which pertains to the distinction between content (or message) cues and source cues (Hoewe and Barton, 2024; Reinhard and Sporer, 2010).
Content cues refer to the information or arguments presented and are thus related to the concept of framing. Framing refers to how information is presented – namely, what aspects are emphasised or downplayed – which can influence individuals’ opinions, while cueing refers to how individuals use information shortcuts to make judgements and decisions about a subject (Druckman et al., 2010). A frame thus serves as a content cue, ‘insofar as a frame constitutes information that individuals use to simplify the decision-making process’ (Druckman et al., 2010: 137).
Source cues, on the other hand, pertain to the identity of the source that is providing the cue. In other words, content cues concern the effect of the message on the recipient, while source cues pertain to the effect of the messenger. There is a long-standing debate about the relationship between the two, though it is fairly consensual that source cues impinge, to a greater or lesser extent, on the effect of content cues (Druckman, 2001). The former leads citizens to focus at least in part on the messenger, instead of the message, because a source, such as a party, permits citizens to be confident with lower effort levels (Bullock, 2011: 498). In the context of EU politicisation and the top-down party cues that were outlined above, content cues centre on the position being presented regarding the EU in a particular statement; while source cues would concern the political party that is presenting this position.
While we know that cues matter, there is relatively little assessment of the effects of content cues and the extent to which these can drive top-down politicisation. The very partial exceptions are the works of Pannico (2020) and Vössing (2021). Both consider the separate effects of policy information (content) and party (source) cues. The former finds that source effects outweigh content ones, with respondents more likely to support an EU policy position from their preferred party (Pannico, 2020). The latter finds that both party cues and policy justifications have an impact on attitudes towards EU policies, albeit dependent on participants’ political engagement and on the degree of political competition around the issue. However, in both cases, the dependent variable pertains to participants’ support for specific EU policies: the role of the EU in energy policies (Pannico, 2020) or support for a Euro-tax (Vössing, 2021). Neither study examines how cues affect participants’ attitudes towards the EU in general. As such, these studies do not explore the effects of content cues in driving top-down politicisation with regards to EU support, even if they provide relevant pointers.
Within this overall backdrop, we test the following five hypotheses. The first assesses the Hooghe and Marks (2009) perspective of EU politicisation. Wiesner (2021: 36) defines this as a pessimist top-down model of politicisation, whereby EU politicisation emphasises criticism of European integration and its dangers, leading citizens to then withdraw support from the EU. This perspective finds ample echo in the literature. Negative messages on the EU from political elites (Bauhr and Charron, 2023), parties (Juratic et al., 2023) and the media (Desmet et al., 2015) have been reported to foster more negative attitudes towards the EU.
By counterpoint to this pessimist perspective, Wiesner (2021: 37) posits an optimist top-down model of politicisation, whereby EU politicisation focuses on the positive aspects of the EU, consequently increasing EU support among citizens. Carrieri (2021) finds evidence of pro-EU politicisation by mainstream parties, influencing voting behaviour. Likewise, there is some evidence that a positive tone regarding the EU in the media improves citizens’ attitudes towards the EU (Desmet et al., 2015).
At the same time, the effects of cues do not occur in a vacuum. Rather, they potentially interact with the characteristics of the recipient of the cue. As Iyengar and Kinder (1987) demonstrate in their seminal work, political messages are interpreted through the lens of an individual's prior beliefs and identities. We thus assess whether the effect of cues is moderated by the following individual-level characteristics. First, we consider an individual's national vs. EU identity. Existing research suggests that stronger national identities amplify negative reactions to EU politicisation, while a European-oriented identity enhances receptiveness to transnational political messaging (Hooghe and Marks, 2005, 2009).
Second, we consider the moderating role of individuals’ external efficacy with regards to the EU, i.e. the perception that the EU is responsive to their concerns and preferences. Recent EU-wide comparative studies have found that those who perceive the EU as being responsive to their preferences have higher support for the EU (Mcevoy, 2016), a pattern found in the southern European context also (Teperoglou and Belchior, 2024). These results echo the pioneering work of Rohrschneider (2002), which shows that perceptions of a lack of political representation in the EU lead to weaker support for the EU. Building on this, we expect that individuals who feel that the EU does not listen to their concerns will rely more on external cues to form their opinions.
Finally, we assess whether the effect of cues is moderated by interest in politics. Individuals with less interest and awareness in politics rely more on party cues and heuristic processing in general (Kam, 2005). This is also confirmed in the context of the EU, with Pannico (2020) finding that less aware citizens are more easily influenced by cues than citizens with a better knowledge of EU politics. Awareness is typically measured by individuals’ levels of political knowledge. However, as Gherghina and Silagadze (2023) note, knowledge presupposes prior motivation, which is measured through political interest. Political interest thus constitutes ‘a precondition for both level of knowledge and political awareness – the factors that decide what mechanisms of opinion formation are at play’ (Gherghina and Silagadze, 2023: 584). In this article, we explore the effect of the interest in politics, following existing research (Gherghina and Silagadze, 2023; Linde, 2018).
Data and methods
This article assesses the effect of content cues on citizens’ attitudes towards the EU. To test this, we conducted an online population-based experiment with a representative sample of the Portuguese population.
The experiment randomly assigned participants to groups exposed to different treatments. These derive, first, from the exposure to positive and negative cues regarding the EU. We term this the tone of the cue, reflecting how the EU is portrayed (Lengauer et al., 2012). A positive tone depicts the EU as an opportunity with positive impacts and characteristics, e.g. in terms of ‘political success, problem solutions, achievement, improvement, advance, prosperity, accomplishment, enthusiasm, hope, benefit, gain, sustainability, gratification or accomplishment’ (Lengauer et al., 2012: 195; Maier et al., 2012). Conversely, a negative tone cue portrays the EU as a threat with negative characteristics, associating it with political failure, crisis or disappointment, inter alia (Lengauer et al., 2012: 195; Maier et al., 2012). It should be noted that this concept of positive and negative tones in party cues towards the EU has been applied more widely in the literature (e.g. Braun et al., 2019).
In addition to the tone, the experiment also tests for the context or scope of EU politicisation. As Silva et al. (2022) note, we can speak of a ‘differentiated politicisation’ of the EU in public debates, reflecting the context in which the EU is framed. As the authors note, politicisation can indicate a more Europeanised public space when the focus is placed at a supranational rather than intergovernmental level, and vice-versa.
In this experiment, we assess the effects of this differentiated politicisation. To that end, we distinguish between two types of EU scope. The first pertains to the EU being politicised in terms of its positive or negative effects on the member-state where the cue is being provided, e.g. how the EU negatively or positively affects the country's economy, culture, public administration or other dimensions. Adapting Hurrelmann et al. (2015), we term this as a domestic scope: the effects of EU membership at the national level. An example of a message with a domestic scope and a positive tone in the experiment is: ‘The European Union strengthens economic growth in Portugal’.
By contrast, the supranational scope refers to the EU being politicised in terms of its broader effects on the EU and its member-states collectively. Rather than linking the EU directly to the outcomes of a specific member-state, messages with a supranational scope highlight consequences for Europe as a whole. For instance, a positive tone message with a supranational scope in the experiment reads: ‘The European Union strengthens economic growth across Europe’.
These examples illustrate how a message can present the EU's impact either at the national level (domestic scope) or at the broader European level (supranational scope). These two scopes carry different normative implications for how the EU is understood and may influence citizen attitudes in distinct ways. However, the existing literature on EU top-down politicisation has not examined their different effects, instead conflating both under the classification of EU scope (Eugster et al., 2021).
The combination of the positive and negative tone with a domestic and supranational scope yields a 2 × 2 factorial design, summarised in Table 1.
Factorial design of the experimental treatments.
EU: European Union.
In addition to these four experimental groups, the study encompasses a group that was exposed to campaign messages without any EU tone or scope. This serves as a control group to measure the impact of the experimental stimuli, enabling us to assess the effect of EU cues in comparison.
The online survey consisted of an initial questionnaire comprising socio-demographic and attitudinal questions. Afterwards, participants were randomly assigned to the treatment groups. The stimuli were artificial billboards – large outdoor campaign posters. Participants were informed that they would see ‘maquettes of billboards for the recent European election campaign of May 2019’ and were asked to observe these attentively, which they saw on their screen. Exposure time was not forced.
Each participant was exposed to four different billboards. These framed the EU with the randomised treatment of a positive or negative tone and a domestic or supranational scope on four different issue topics: economy; identity and culture; environment; and immigration. Following Carmines and Stimson's (1980) distinction between ‘easy’ and ‘hard issues’, these topics were selected to include only easy issues, with which the general public tends to be familiar with. Moreover, these are issue topics that have been highlighted in the literature on EU politicisation. Economic and identity issues are seen as central in the politicisation of the EU (Hooghe and Marks, 2005), with immigration (Carmo Duarte, 2024) and environmental issues (Wiesner, 2022) also playing an important role.
By including different topics, the experiment minimises the effect of participants attributing different levels of importance to different issues, which can bias the results of a single-issue design. Moreover, doing so also helped generate longer exposure periods. Given that participants were exposed to campaign materials with relatively short written information, this ensured that participants spent more time exposed to the selected treatment. The order by which the issues were presented to participants was randomised, to mitigate effects caused by the sequencing of the topics. The campaign materials were presented in the format of artificially designed outdoor billboards, simulating the type of real-world information that might be conveyed in political billboards.
Table 2 outlines the textual stimuli presented, highlighting the application of a consistent slogan for each treatment, aligned with the intended scope and tone. Graphically, all the materials follow a similar template and include the context given to participants, stating that these were for the 2019 European elections. Crucially, the stimuli excluded any reference to a source, such as a political party, enabling us to isolate the effect of content alone. Examples of the graphical template for each treatment are provided in the Online appendix.
Textual stimulus used.
EU: European Union.
The visual information in the billboards served to reinforce the scope and tone of the message. Thus, those with a supranational scope present images that represent the EU transnationally. In those with a positive tone, the EU is symbolised by its flag, reflecting the treatment's slogan of ‘More Union, Better Europe’, while those with a negative tone present the flags of member-states replacing the EU flag, mirroring the treatment's slogan of ‘Less European Union, Stronger Nations’. In those with a domestic scope, the positive tone visual image depicts the Portuguese flag being joined with the EU one, reflecting this treatment's slogan of ‘More Union, Better Portugal’; while, conversely, the negative tone sees the Portuguese flag replacing the EU one, consistent with the treatment's slogan of ‘Less European Union, Better Portugal’. Finally, the billboards for the control group do not make any reference to the EU, either textually or visually. To maintain consistency with the other stimuli, they address the same four issues but do so solely from a national perspective, without linking them to the EU. Given the absence of EU references in the presentation of the issues in this treatment, we do not expect it to affect participants’ pre-existing EU attitudes.
After seeing the posters, participants were asked about their attitudes towards the EU. Attitudes towards the EU are multidimensional (Boomgaarden et al., 2011: 244; De Vreese et al., 2019). Research on EU attitudes thus benefits from greater specification on the dependent variable, distinguishing between the various dimensions that comprise EU support (Boomgaarden et al., 2011: 260). This is a concern that this article addresses, by disaggregating the dependent variable of EU support on four different dimensions, detailed in Table 3. These four dimensions draw from standard Eurobarometer items and measure distinct dimensions of EU support, namely the EU's image, with item 1 (Van den Hoogen et al., 2022); distinct dimensions of utilitarianism, with items 2 and 3 (Boomgaarden et al., 2011; De Vreese et al., 2019), with item 2 used by some studies to measure EU broad support or diffuse support (Beaudonnet and Franklin, 2016; Clark, 2020), while item 3 is used to measure national benefit (Hooghe and Marks, 2005); and, finally, EU performance, with item 4 (Boomgaarden et al., 2011; De Vreese et al., 2019). To enable comparisons among variables, all item responses were measured on a five-point Likert scale ranging from 1 to 5, where 1 stands for the highest degree of support for the EU, 5 for the lowest.
Attitudes towards the European Union (measures).
EU: European Union.
In terms of the moderators, external political efficacy is measured with agreement to the following question: ‘My voice counts in the EU’, ranging from 1 (‘I disagree completely’) to 5 (‘I agree completely’).
Participants’ national vs. European identity is obtained by aggregating participants’ agreement with the following 10 statements on a scale of 1 (disagree completely) to 5 (agree completely):
I identify with the Portuguese; I am happy to be a Portuguese citizen; Being Portuguese is important to me; I share Portuguese ideals and values; In general, I am similar to the Portuguese. I identify with Europeans; I'm happy to be a European citizen; Being European is important to me; I share European ideals and values; In general, I am similar to Europeans.
Following the logic of Maier et al. (2012), these were aggregated by subtracting participants’ degree of identification with Portugal from their degree of identification with Europe, with the measure having values from +20, signifying an exclusively European identity, to −20 for an exclusively national identity. Finally, the level of interest in politics was divided into low vs. high interest in politics (low: ‘not’ and ‘little interested’; high: ‘reasonably’ and ‘very interested’). 1
An international market research company, Cint, was recruited for data collection using the computer assisted web interviewing methodology, ensuring that the panel met standard criteria for online research and participant recruitment. A quota sampling was drawn from the panel to be representative of the Portuguese adult population regarding gender, age and geographical region according to the nomenclature of territorial units for statistics II classification of Eurostat, with fieldwork taking place between 15 January and 11 March 2020. The experiment was validated and approved by the Ethics Committee of the University of Aveiro (decision no. 06-CED/2019), ensuring that all procedures met the required ethical standards.
Initially, 2166 participants were randomly assigned to each of the five different groups, with a minimum of 400 participants per group. After the data collection, manipulation checks were carried out to ensure participants apprehended the instructions and were attentive to the experimental stimulus, following Kane and Barabas (2019). This removed 357 responses, leaving a total of 1804 participants, with at least 300 per group. The experimental groups are not statistically different in a set of relevant variables (for test details and results, see the Online appendix).
Finally, as noted, the experiment was carried out for a representative sample of the Portuguese population. Portugal constitutes an interesting case to examine the effects of political cues, for two reasons. First, the generally high levels of EU support. The country has long been among the most pro-EU member-states, with relatively high levels of support for European integration that, after the Eurozone crisis, were above the pre-crisis levels (Scotto di Vettimo, 2022). In the Summer 2020 Standard Eurobarometer 93 (European Union, 2020), which took place shortly after the fieldwork for this experiment, Portugal presented the highest level of positive image of the EU in the EU-27 countries (tied with Poland), with 55% of respondents having a positive image, well above the EU average of 40%. Moreover, 35% of respondents had a neutral image of the EU, not far below the EU-27 average (40%).
Second, Portugal presents a party system that has been predominantly in favour of European integration, with a weak politicisation of the EU. Opposition to the EU has been restricted to smaller parties, particularly in the hard Eurosceptic Communist Party; and, though to a lesser extent, in the soft Eurosceptic radical-left Left Bloc, which is more divided on the EU (Cunha, 2021; Lisi, 2024). Euroscepticism gained salience in party discourses during the Eurozone crisis, but this fell again with the end of the country's bailout (Lisi, 2024). This weaker salience was accentuated by the collaboration of the Eurosceptic radical left parties with the pro-EU Socialist Party in the government solution of 2015–2019, which further weakened Eurosceptic politicisation (Lisi, 2024). Indeed, even the radical-right populist Chega – which won 1.3% of the vote and a single parliamentary seat in 2019 – adopted a low salience soft Eurosceptic position vis-à-vis Europe (Cunha, 2021: 154). Overall, then, there emerges a ‘depoliticised consensus surrounding Portuguese integration in Europe’ (Standring, 2019: 137).
Consistent with this, in the October 2019 legislative elections, held shortly before this article's fieldwork, parliamentary parties with a negative position on European integration, based on the Manifesto Project's (MP) European integration position indicator (Lehmann et al., 2024) received a combined 18% of the vote and 14% of parliamentary seats. Mean EU salience, drawing on the MP European integration salience indicator, was 1.05 in this election, well below the mean of 2.35 across all EU member-states in the MP dataset, using the election closest to the 2019 Portuguese election (Lehmann et al., 2024).
This makes Portugal a relevant case for assessing how political cues shape citizens’ attitudes towards the EU. Given the relative absence of politicisation, the effect of the experimental cues is less influenced by broader party system dynamics. Likewise, the generally high levels of support for the EU in the population allow us to better isolate how content cues resonate in an environment that lacks strong anti-integration attitudes. As such, as well as assessing the effects of cues in the Portuguese context, given the representative nature of the sample, this study also informs our broader understanding of their effect more generally.
Results and discussion
We begin by testing the effect of the cues on the four dimensions of support for the EU. The results are summarised in Table 4 (full models are reported in the Online appendix). The treatment is significant in two of the models, with the marginal effects of these presented in Figure 1.
Marginal effects of cues on dimensions of support (95% CIs). 95% CI: 95% confidence interval.
Effect of treatment vs. control cues on EU support (ordered logit regressions with robust standard errors).
DV: Dependent Variable; EU: European Union.
***p < 0.01, **p < 0.05.
These results further our understanding of the mechanisms of EU politicisation. They confirm the top-down perspective of politicisation, with the cues provided in this experiment influencing citizens’ attitudes towards the EU. Moreover, they indicate that content cues can influence support for the EU, even when devoid of source (party) information.
In terms of our hypotheses, the results support the pessimist top-down politicisation hypothesis (H1). The two models that present a significant effect of cues both pertain to negative tones. As Figure 1 indicates, exposure to negative tones leads to less positive attitudes towards the EU. Exposure to the negative tone in Figure 1(a) reduces the odds of a positive evaluation of the EU by 24%, with an odds ratio (OR) of 0.756 compared to the control. In Figure 1(b), exposure to the negative tone reduces the odds of a positive evaluation of the EU by 35%, corresponding to an OR of 0.653. Conversely, the optimist top-down politicisation hypothesis (H2) is not supported by these results, with positive cues having no statistically significant effect on participants’ support for the EU.
The results thus crucially point to an imbalance in terms of the effects of politicisation, with attitudes towards the EU more influenced by negative than positive tone cues. Consistent with the pessimist top-down perspective of Hooghe and Marks (2009), we find that, when EU politicisation criticises the EU, citizens withdraw support from the EU. The optimist perspective of Wiesner (2021), on the other hand, is not supported: positive cues regarding the EU had no significant bearing on attitudes towards the EU.
Politicisation, then, does not occur on a level playing field: anti-EU messages are more likely to influence support for the EU than pro-EU messages. This result is consistent with the recent study of Beaudonnet and Gomez (2024), which finds an imbalanced effect of politicisation that favours Eurosceptic parties and the constraining dissensus. Examining the effects of politicisation of the EU on the propensity to vote for different parties, they find that politicisation creates a ‘one-way street’ that benefits Eurosceptic parties more than their Europhile counterparts. As such, the attempts by pro-EU parties to depoliticise the EU found in some of the literature (Adam et al., 2017) may be a rational response to a process where Eurosceptic parties stand to gain more than their Europhile counterparts, both electorally and in terms of shaping public opinion towards the EU.
Finally, the results provide two additional results of note. First, the cues affect different dimensions of EU support. Specifically, evaluations regarding the benefits of EU membership and the EU's performance were more sensitive to cues, while perceptions of the EU's overall image and the broader question of whether membership is a good thing remained more resistant. This pattern suggests that cues influence specific, concrete evaluations – such as institutional performance and benefits – while more general perceptions of the EU's image and evaluations of whether EU membership is a good thing are less affected, proving resilient in this experimental design.
Second, in terms of differentiated politicisation, the results suggest that attitudes are affected by negative tones both when framed with a domestic and supranational scope. The effect of cues is stronger with a domestic scope, when the EU is framed as having a negative impact on the member-state (Figure 1(b)). However, interestingly, a negative tone impacts attitudes also when framed with a supranational scope. This suggests that criticisms of the EU can significantly shape public attitudes even when articulated at the European level.
We next examine whether the effect of cues is moderated by prior characteristics of the participants, namely their national vs. EU identity; their external political efficacy and, finally, their interest in politics. Of these three, we find no moderation effect for external political efficacy in any of the models, with the effect of the cues on EU attitudes not varying significantly across different levels of external political efficacy (full models are reported in the Online appendix). As such, we can reject the political efficacy moderation hypothesis (H4). However, we do find significant moderation effects of participants’ identity, as shown in Figure 2.
Marginal effects of negative tone with domestic scope on EU support, moderated by participants’ national vs. EU identity (95% CIs). EU: European Union; 95% CI: 95% confidence interval.
The results indicate a substantial interaction between identity and negative tones. Consistent with H3a, cues with a negative EU tone have a more negative effect on EU support among individuals with a stronger national identity. This is true for three different dimensions of EU support: the image of the EU; evaluations of EU membership as a good thing for the member-state; and evaluations of EU performance. In all of these, as identity shifts from exclusively national to exclusively EU, the effect of the cue on EU support decreases. The effect is statistically significant for individuals with a strong national identity and loses significance as identity becomes more EU-oriented. The one exception in this latter regard is shown in Figure 2(b), where exposure to the negative tone in fact reinforces EU support among those with a stronger EU identity. A strong EU identity thus mitigates the effects of cues with a negative EU tone, while those with a stronger national identity are more likely to be influenced by these cues. However, we do not find support for H3b, as cues with a positive tone about the EU do not have a more positive effect on EU support among individuals with a stronger EU identity.
Let us now turn to H5, which examines the moderating effect of participants’ interest in politics. Here, we find a significant interaction effect of interest in politics with the negative tone and domestic scope treatment, presented in Figure 3 (full models are reported in the Online appendix).
Marginal effects, interaction between interest in politics and negative tone with domestic scope (95% CIs). 95% CI: 95% confidence interval.
The results indicate that this interaction significantly moderates participants’ assessment of whether EU membership is a good thing. Consistent with H5, participants that have a low interest in politics are significantly more likely to disagree with the notion that EU membership is a good thing (Figure 3(a)) when exposed to a negative tone regarding the EU's impact on the member-state. On the other hand, the cue has no effect on those with a high interest in politics (Figure 3(b)). Thus, it appears that having high levels of political interest and awareness can indeed insulate against the effect of cues, as posited by Kam (2005).
Examining these moderating effects as a whole, three ideas emerge. First, significant moderating effects occur only with respect to treatments with a negative tone. This reinforces the notion that politicisation affects public opinion in a skewed manner, with a greater effect of negative tones, which particularly affect those with less interest in politics and a stronger national identity. Once again, there are no interaction effects for positive EU cues, consistent with the notion that politicisation is imbalanced towards Euroscepticism. Second, the interaction effects are only significant in treatments with a domestic scope. This suggests that, when we consider differentiated politicisation, the effects are greater for treatments that focus on the effects of the EU on the member-state, particularly for those who are less interested in politics and have a more national identity. Third, when we examine the different dimensions of EU support, we find significant moderating effects on the more general perceptions of the EU's image and of whether EU membership is a good thing, suggesting that cues affect across the different dimensions of EU support when the moderating effects of a low interest in politics and stronger national identity are considered.
Conclusion
This article examined how top-down cues politicising the EU affect citizens’ attitudes towards the EU. It does so by experimentally testing the effects of content cues on different dimensions of EU support, isolating them from source (party) cues. This empirical approach is used to assess the debate between the pessimist and optimist top-down models of politicisation (Hooghe and Marks, 2009; Wiesner, 2021). The former posits that EU politicisation emphasises criticism of the EU, leading to weaker citizen support for the EU; the latter that politicisation can foster greater citizen support for the EU by focusing on the EU's positive aspects.
Overall, we find that top-down content cues do impact support for the EU. However, not all cues have the same impact. Consistent with the pessimist top-down model of politicisation, we find that EU support is significantly weakened by cues with a negative tone towards the EU. By contrast, and against the optimist model, we do not find any significant effects of a positive tone on attitudes towards the EU.
The effect of cues is also moderated by individuals’ characteristics. In particular, those with a stronger national identity, as well those with a lower interest in politics, are more susceptible to having their attitudes towards the EU affected by cues with a negative tone towards the EU. Conversely, a stronger European identity and greater interest in politics appear to insulate against the pessimist top-down model of politicisation.
Another contribution of this research is testing the effects of differentiated politicisation. Specifically, it examines two different levels at which the impact of the EU can be politicised. One pertains to the EU's impact on the member-state where politicisation is taking place, defined as a domestic EU scope. The other refers to the EU's impact on the EU and its member-states as a whole, termed as a supranational EU scope. Overall, we find that cues with a domestic scope have a greater effect on attitudes towards the EU. However, albeit to a lesser extent, we also find a statistically significant effect for a negative tone treatment with a supranational scope, which suggests that criticisms of the EU can influence citizens’ attitudes even when framed at the European level.
This experimental research was carried out with a representative sample of the Portuguese population, thus providing an assessment of these effects in the Portuguese context. Moreover, Portugal constitutes a relevant context for examining the causal effects of cues more generally, given the weak politicisation of European integration in the party system and the high levels of EU support. Nonetheless, future research can assess cue effects in contexts with different levels of EU politicisation and support, examining how these moderate the effect of cues. Likewise, it can seek to overcome a limitation of this research, namely that, as an experiment with artificial stimuli, it does not fully replicate real-world conditions and involves a short exposure to cues, which potentially limits their impact. To that end, an experiment with a longer exposure period and based on actual and manipulated political materials could be devised. Finally, while political actors can shape the public's views through politicisation, they also react to public opinion, albeit in a strategic rather than direct manner (Soo et al., 2023). Future research can thus explore the interaction between bottom-up dynamics and top-down politicisation, whereby public responses shape how and whether elites choose to politicise EU issues.
The results shed light on the causal mechanisms driving the mass public's constraining dissensus within the EU. They also indicate that top-down politicisation does not play out on an even playing field: anti-EU messages are more likely to impel negative attitudes towards the EU than pro-EU messages are to generate positive attitudes. This asymmetry means that further politicisation may accentuate, rather than counter, the constraining dissensus.
Supplemental Material
sj-docx-1-eup-10.1177_14651165251364768 - Supplemental material for Which cues matter: The effect of content cues on European Union support
Supplemental material, sj-docx-1-eup-10.1177_14651165251364768 for Which cues matter: The effect of content cues on European Union support by Carlos Jalali, Patrícia Silva, Patrício Costa and Filipe Falcão in European Union Politics
Supplemental Material
sj-zip-2-eup-10.1177_14651165251364768 - Supplemental material for Which cues matter: The effect of content cues on European Union support
Supplemental material, sj-zip-2-eup-10.1177_14651165251364768 for Which cues matter: The effect of content cues on European Union support by Carlos Jalali, Patrícia Silva, Patrício Costa and Filipe Falcão in European Union Politics
Supplemental Material
sj-zip-3-eup-10.1177_14651165251364768 - Supplemental material for Which cues matter: The effect of content cues on European Union support
Supplemental material, sj-zip-3-eup-10.1177_14651165251364768 for Which cues matter: The effect of content cues on European Union support by Carlos Jalali, Patrícia Silva, Patrício Costa and Filipe Falcão in European Union Politics
Footnotes
Acknowledgements
We thank Michelle Serralva Macêdo and Teresa Forte for their assistance in the preparation of the survey, as well as Tiago Silva, José Santana-Pereira and João Moniz for their suggestions and comments. We are also grateful to the three anonymous reviewers and the journal's executive editor, Gerald Schneider, for their insightful and comprehensive comments and suggestions.
Author contributions
CJ, PS, PC and FF contributed to conception and research design. CJ drafted the article. PS contributed to the literature review section and critically revised the article. CJ performed the data analysis. PC verified the data analysis. FF designed the experimental stimuli.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by the Research Project ‘Changing European Elections: The impact of Eurozone bailouts on European Parliament election campaigns’, funded by the Portuguese National Foundation for Science and Technology, FCT (PTDC/IVC-CPO/3481/2014), the Operational Programme ‘Competitiveness and Internationalization’ (COMPETE 2020) and the Lisbon Regional Operational Programme (POCI-01-0145-FEDER-016887). It also received funding from the Research Unit on Governance, Competitiveness and Public Policies (UIDB/04058/2020) + (UIDP/04058/2020), funded by national funds through the Portuguese National Foundation for Science and Technology, FCT.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data availability statement
The data and replication code are available as part of the Supplementary Material at the DOI of the article.
Supplemental material
Supplemental material for this article is available online.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
