Original Research Article Party Politics ‘Party position, electoral incentives, and 2023, Vol. 0(0) 1–11 © The Author(s) 2023 Article reuse guidelines: attention to European Union issues in the sagepub.com/journals-permissions DOI: 10.1177/13540688231224059 journals.sagepub.com/home/ppq 2019 European Parliament election’ Nuria Font  Universitat Autònoma de Barcelona, Spain Abstract Under what conditions do Members of the European Parliament (MEPs) give attention to EU issues in European Parliament (EP) election campaigns? This study examines how party positioning on European integration and individual electoral incentives affected MEPs’ attention to EU issues in the 2019 EP election campaign. Based on data collected from Twitter, the findings show that pro-EU parties and individual vote-seeking incentives increased attention to EU issues. More specifically, they show that MEPs from pro-EU mainstream parties had incentives to debate about EU issues, especially when they competed for votes. The findings help to unfold the interplay between party-level and legislator-level factors and its effects on attention to EU issues and aim to contribute to research on EU issue salience and EP election campaigns. Keywords European parliament, elections, salience, members of the European parliament, political parties Introduction were advanced by Hooghe and Marks’ (2009) claim of a post-functionalist turn in European integration theory. Ar- Under what conditions do Members of the European Par- guing that gradual transfers of authority to the EU in the liament (MEPs) give attention to EU issues in EP election decades following the Maastricht Treaty increased the sa- campaigns? Research on EU issue salience and European lience of European issues, they illuminated a path to a fertile elections has often addressed the question of whether EP field of research. Therein, scholars have since explored the elections are about Europe. For decades, due to being drivers of EU salience and politicization both in the do- dominated by domestic concerns instead of European is- mestic (e.g. Grande and Hutter 2016; Hutter and Kriesi sues, EP elections have been categorised as second-order 2019; Kriesi 2016) and European arenas (e.g. Braun et al., national elections (Reif and Schmitt 1980). This catego- 2016; Braun and Grande 2021; Spoon 2012), as well as risation has recently been challenged by the idea that EU politicization’s normative and empirical repercussions (e.g. issues are increasingly important in EP elections (e.g. Adam Braun and Grande 2021; Vasilopoulou and Gattermann and Maier 2016; Braun and Grande 2021; Braun and 2021). Such studies have commonly conceptualised polit- Schäffer 2022; Eugster et al., 2021; Hobolt and De Vries icization as a process through which European integration 2016). The claim that EP elections are about Europe becomes more visible and contested (e.g. De Wilde et al., probably found its highest expression in the 2019 election, 2016). As such, politicization is conceived to be composed when high levels of Europeanness of the electorates (Braun of three dimensions: EU issue salience (that is, the public 2021), European issue salience (Braun and Grande 2021), and politicization (Braun and Schäfer 2022) were observed. Research on European elections has shown that EU issues have become increasingly salient for parties and Paper submitted 23 March 2023; accepted for publication 12 December provided accounts of parties’ emphasis on European topics. 2023 Recent studies on EU issue salience have commonly been Corresponding author: conducted under the research umbrella of the politicization Nuria Font, Political Science and Public Law, Universitat Autònoma de of European integration. From this broader perspective, Barcelona, Edifici B, Barcelona 08193, Spain. theoretical debates on European integration’s politicization Email: nuria.font@uab.es 2 Party Politics 0(0) visibility of EU issues), the polarisation of opinions and issue differently, Nai et al. (2022) show that Eurosceptic debates on European integration, and actor expansion (De parties are more likely than Europhile ones to campaign in Wilde et al., 2016; Grande and Hutter 2016; Hutter and harsh ways and to benefit electorally from it. Last, taking a Kriesi 2019). In our study we focus on the salience of EU multilevel election system perspective, Braun and Schmitt issues during the 2019 EP election, examining how un (2020) show that parties put less emphasis on EU issues in under what conditions EU issues were emphasized and national elections than in EP elections, although they take given visibility in the public debate. We share Grande and similar positions on the EU in both electoral contests. Our Hutter’s (2016) assumption that salience is not the sole study engages with those theoretical debates by investi- dimension of politicization but is the most basic one and gating how pro-EU mainstream parties were more incen- cannot be replaced by any other dimension. They also tivized to emphasize EU issues than Eurosceptic ones in the contend that only issues raised by political actors and de- 2019 election. Our study also connects with studies on both bated in public can be regarded as being politicised (Grande the polity-policy dimension of EU issue emphasis and and Hutter 2016). Although we share their view that only political elites’ communication tone. the study of the three dimensions can provide a full picture Our study focuses on MEPs in their capacity as national of the politicization of European integration, we also believe party members and representatives in the EP (e.g. Obholzer that some underexplored aspects of EU issue salience in EP and Daniel 2016). This approach enables us to explore how elections, namely the role of pro-EU mainstream parties and party-level and individual-level incentives influence EU the link between party positioning and individual vote- issue salience. This is relevant because it is unclear whether seeking considerations, still need to be surveyed. the MEPs’ emphasis on EU issues in EP election campaigns Research on EU issue salience has revealed that parties is primarily driven by their parties’ orientation towards emphasize and de-emphasize EU issues to gain advantage in European integration or also influenced by individual vote- electoral competition (e.g. Braun and Grande 2021; Braun seeking incentives. Previous research shows that Euro- and Schmitt 2020). Theoretical accounts of EU issue sa- sceptic parties emphasise European issues (Braun and lience have emphasised the attributes of parties and party Grande 2021; Hooghe and Marks 2009), that individual systems, the nature of the European issues being examined, vote-seeking motivations influence legislators’ behaviour in and the national and transnational campaign arenas. One of the parliamentary and electoral arenas (Müller and Strøm the key arguments is that Eurosceptic challenger parties 1999), and that electoral systems impact parties’ and po- have increased the salience of European issues and are the litical elites’ behaviour and campaigns in the European main drivers of political conflict over Europe (e.g. Hutter context (Farrell and Scully 2010; Hermansen and Pegan and Kriesi 2019; Braun and Grande 2021) and that pro-EU 2023). We investigate attention to EU issues from the mainstream parties have de-emphasized and even blurred perspective of MEPs’ role in election campaigns to better EU issues in response to Eurosceptic challengers (e.g. understand the interplay between party and individual Koedam 2021). Relatedly, Braun and Grande (2021) show electoral drivers of attention to EU issues in EP elections. that mainstream parties de-emphasise European issues in EP We put the spotlight on MEPs for two main reasons. First, as election campaigns in response to the intensification of party members, MEPs are expected to align with their party political conflict over Europe by radical Eurosceptic parties. position on EU issues. Second, because not all MEPs stand They also observe this trend in the 2019 elections, when for re-election, their individual electoral incentives differ. mainstream parties were forced to open the debate around Such differential status raises the underexplored question of European issues only in countries with high levels of po- how MEPs’ attention to EU issues in EP election campaigns liticization (Braun and Grande 2021). Likewise, Adam and is driven by party considerations as well as by individual, Maier (2011) observe that EU issues are more prominent on vote-seeking incentives. By investigating the effect of the the EP campaign agenda in party systems with many Eu- interplay between party and electoral considerations on rosceptic parties. Spoon (2012) describes variation in the attention to EU issues through the lenses of MEPs, we aim salience of European issues across parties and over time and to contribute to debates on EU issue salience and EP demonstrates that domestic politics, namely parties and election campaigns. party systems, influence the salience of European issues in Our study is based on Twitter content published by MEPs parties’ Euromanifestos. With a different focus, Adam and during the 2019 EP pre-election period. Relying on Twitter Maier (2016) and Eugster et al. (2021) find that parties with data is justified, as prior research on EU issue salience—and internal dissent about the EU avoid European issues. Fo- more generally on politicization—and EP election cam- cusing on a polity-policy perspective, Braun et al. (2016) paigns has primarily relied on sources such as Euro- argue that European issue salience is affected by party manifestos (Braun et al., 2016; Braun and Schmitt 2020; position and party system polarisation, with parties’ em- Spoon 2012), mass media coverage (Braun and Grande phasis on Europe focused on policies more than on fun- 2021; Grande and Hutter 2016; Hutter and Kriesi 2019), damental EU issues or domestic topics. Approaching the party communications (Eugster et al., 2021), or a Font 3 combination of these sources (e.g. Koedam 2021; Maier whether MEPs, beyond party lines, have individual vote- et al., 2019). However, digital innovation has led parties and seeking motivations when deciding to emphasize or de- political elites to increasingly use social media, especially emphasize EU issues in EP campaigns. Our interest is Twitter, as a key communication tool in election campaigns. investigating attention to EU issues in the 2019 EP These platforms enable political elites to provide person- election focusing on MEPs in order to unfold the effects alised content (Nulty et al., 2016) and emphasise (or de- of the interplay between party-level and individual-level emphasise) policy issues (Fazekas et al., 2021). Research electoral considerations on attention to EU issues. Ulti- has explored several dimensions of candidates’ and MEPs’ mately, our findings aim to contribute to understanding usage of social media in the EP (e.g. Daniel et al., 2019). better the drivers of EU issue salience in European More specifically, Fazekas et al. (2021) find that politicians election campaigns. adopt a less engaging style when discussing EU issues on Twitter during EP election campaigns, while Stier et al. Theoretical framework (2021) highlight national rather than transnational linkages among EP candidates on Twitter in the 2019 election. Our Theories on issue salience argue that parties strategically study also contributes to research on social media cam- emphasise certain issues and de-emphasise others to paigns by analysing attention to EU issues on Twitter during maximise their electoral prospects (Koedam 2021; Spoon the EP campaign. 2012). Building on research on EU issue salience and EP Building on theoretical debates on EU issue salience and campaigns, we present arguments based on party positions EP campaigns, we aim to elucidate how party EU position, on European integration, individual vote-seeking goals, and standing for re-election, and the ballot system affect MEPs’ the ballot structure. Hypotheses at the party, individual, and attention to EU issues. Our study is based on a dataset of electoral system levels are proposed. 61,648 tweets from MEPs from all member states during the four weeks preceding the 2019 election. Our study shows that pro-EU mainstream parties did not avoid discussing Party position on European integration about EU issues and that, among MEPs from pro-EU Research on EU issue salience and, more generally, on the parties, those competing for votes were more inclined to domestic politicization of European integration, has emphasize EU issues. Our study also shows that MEPs from commonly shown that extreme ideological parties are EU-friendly parties focus on specific policy issues and that more incentivised than mainstream moderate ones to Eurosceptics employ a more negative tone when talking politicise the EU in national electoral contests (e.g. about Europe. Our findings contribute to research on EU Hobolt and De Vries 2015; Hooghe and Marks 2009; issue salience and on EP elections and aim to shed light on Hutter and Kriesi 2019). Similarly, research focused on the interplay between party-level and legislator-level factors the European context has also suggested that emphasizing in shaping EU issue attention. EU issues is strategically advantageous for Eurosceptic parties (e.g. Braun and Grande 2021; Braun et al., 2016; Spoon 2012). Along those lines, Braun and Grande Attention to EU issues and EP elections (2021) argue that mainstream parties de-emphasize Eu- This study addresses some underexplored aspects of EU ropean issues in response to Eurosceptics’ mobilization. issue salience in EP elections, namely the role of pro-EU They also observe that trend in the 2019 elections and mainstream parties in the 2019 EP elections and the link show that mainstream parties were forced to open the between party positioning and individual vote-seeking debate about Europe only in countries with high levels of considerations. It bridges two strands of research. On the politicization (France and United Kingdom). We engage one hand, literature on EU salience shows that emphasis with these theoretical debates by hypothesizing that, in on EU issues varies over time (Braun and Grande 2021) the 2019 EP election, MEPs from Europhile parties did and across parties (Braun and Schmitt 2020; Braun et al., not avoid EU issues and were more incentivized to pay 2016; Eugster et al., 2021; Spoon 2012). A common attention to EU issues than MEPs from Eurocritical ones. assumption in related research is that parties and partisan The underlying argument brings into consideration dif- entrepreneurship, most notably within national party ferences in the type and intensity of preferences on competition but also in the EU contest, are the main European integration between pro-EU mainstream drivers of EU salience and politicization (Braun and parties, and anti-EU challenger parties. Grande 2021; Kriesi 2016). On the other hand, com- Regarding pro-EU mainstream parties, electoral pro- parative research on social media campaigns shows that jections in the run-up to the election predicting a surge of campaigns are not exclusively centralised in party radical right parties and a close election nourished the leadership but also decentralised in individual candidates perception among mainstream pro-EU parties that the (e.g. Balmas et al., 2014). This trend raises the question of 2019 election was decisive because the future of Europe 4 Party Politics 0(0) was at stake (Braun and Schäfer 2022). Borrowing from in giving saliency and politicizing European issues. Van Middelaar (2016, p. 495), we assume that main- Relatedly, the role of party elites, specifically MEPs, in stream, pro-EU political forces perceived the EP election adopting their parties’ strategies on EU issues is also as a ‘moment of truth’. Beyond country variations, worth exploring. MEPs seeking re-election may em- mainstream parties were incentivised to engage in efforts phasize certain issues in order to gain votes, while non- of pro-European mobilisation (Braun and Grande 2021; candidates may also emphasize them to campaign for Braun and Schäfer 2022), also by appealing to relevant their national party in pursuit of office benefits (Müller transnational policy issues to which voters attribute and Strøm 1999). However, the latter may also distin- greater importance (Braun and Schäfer 2022). guish themselves in the absence of individual vote- On the Eurosceptic side, anti-EU parties have been seeking goals or under distinct career incentives. We weak in increasing the salience of EU issues in EP assume that attention to EU issues is affected not only by election debates over the past elections (Braun and the party, but also by individual vote-seeking incentives. Grande 2021). Arguably, anti-EU parties’ motivations If, as hypothesized, MEPs competing for votes are more to give attention to EU issues might have been of a incentivized to emphasize European issues, then talking different nature in the 2019 campaign. Recent work by about Europe is meant to mobilize European voters. Braun et al. (2016) on EU salience based on Euro- Our argument partly builds on the theoretical frame- manifestos can help to tailor our expectation, particularly work elaborated by Müller and Strøm (1999) concerning their argument that Eurosceptic parties highlight con- parties’ and politicians’ political behaviour. They dis- stitutive, polity issues instead of issues concerning EU tinguish between office-seeking, policy-seeking, and policy (Braun et al., 2016). We assume that Eurosceptic vote-seeking behaviours (Müller and Strøm 1999). As we parties, at least from the EU27, might have had reasons to seek to discern whether individual electoral incentives lessen the visibility of (primarily constitutive) issues in affect attention to EU issues in European elections, vote- the 2019 campaign. If a country’s EU membership was a seeking objectives are at the centre of our argument. primary constitutive concern (Hutter et al., 2016) for anti- Relatedly, previous research shows that the EP election EU parties, then their campaign strategies might have cycle affects parties and MEPs’ behaviour (e.g. Koop been influenced by the potential repercussions of Brexit et al., 2018) and that, when MEPs campaign in highly negotiations. Although the result of the Brexit referen- politicised contexts, national parties tend to exert control dum initially led Eurosceptics in other member states to over their MEPs (Koop et al., 2018). At the same time, call for membership referendums in their own countries, MEPs standing for re-election might not only emphasise Brexit acted as a deterrent for other member states during their party brand but also, as comparative studies show, the prolonged negotiations (Hobolt et al., 2022). Because adopt a personalised campaign (Balmas et al., 2014). As the uncertainty associated with leaving the EU increased candidate MEP’s future office is directly affected by EP support for European integration, EU membership began elections (Koop et al., 2018), they may have motivations to look relatively favourable (De Vries 2017). Moreover, to pay more attention to EU issues as a way to place value the unity among the 27 member states throughout Brexit on and give visibility to their legislative work within the negotiations allowed the EU to maintain a firm position EU arena, call for votes in the EP election, and try to vis-à-vis the UK government (Dooley 2022), highlighting improve their re-election prospects. By contrast, MEPs the arduous path towards withdrawal. Therefore, if the not participating in the election may have weaker in- idea of one’s country leaving the EU became increasingly centives to campaign for their party. Thus, we expect that unattractive to citizens and possibly less beneficial for MEPs with re-election goals have higher incentives to Eurosceptic parties (in EU27) during the 2019 election, it emphasize EU issues than non-candidate MEPs. More- is reasonable to expect that MEPs from anti-EU parties over, we expect that the effect of EU positioning on found emphasising (typically constitutive) EU issues less attention to EU issues varies depending on whether MEPs electorally advantageous. stand for re-election. MEPs from pro-EU parties are expected to devote more attention to EU issues when H1. MEPs from pro-EU parties give more attention to standing for re-election than when not. While giving EU issues in the EP election campaign than MEPs from attention to EU issues may be strategically advantageous anti-EU parties for pro-EU parties in an election viewed as being deci- sive, standing for re-election provides additional incen- tives to emphasise European topics. In essence, Vote-seeking incentives emphasizing the benefits of European integration in an Previous research has extensively examined strategies election considered to be critical may be particularly employed by parties with opposing views on European advantageous for MEPs from pro-EU parties who try to integration and has shown that parties play a crucial role secure re-election. Font 5 H2. MEPs standing for re-election give more attention to We use MEPs’ Twitter handles that appeared in their EU issues in the EP election campaign than MEPs not profiles on the EP website. This strategy may have limi- standing for re-election tations as MEPs may use other Twitter handles—for in- H3. MEPs from pro-European parties give more atten- stance, a party and/or a private one—either alternatively or tion to EU issues in the EP election campaign when they in complement to the official EP one. By relying on the stand for re-election officially published handles, we can be certain that the results of our analysis are not biased due to the potential over- or underrepresentation of certain parties and MEPs. Electoral list systems Our data show that MEPs’ tweeting activity varies Research has extensively investigated the effects of elec- considerably. The most active MEPs on Twitter are Inés toral list systems on political elites’ behaviour during Ayala Sender (PSOE, Spain, 2344 tweets) and Molly Scott election campaigns (e.g. Obholzer and Daniel 2016) as well Cato (Green Party, UK, 2273 tweets), whereas the least as on legislative behaviour and activity within the EU active are Kay Swinburne (Conservative Party, UK, (Farrell and Scully 2010; Däubler and Hix 2018). As 2 tweets) and Edouard Martin (Parti socialiste, France, comparative literature addressing the effects of electoral 2 tweets). To ensure that the results are not driven by the systems on legislators’ behaviour shows, open list systems, most active users, we perform robustness checks removing usually known as ‘candidate-centred’, allow voters to ex- their tweets. Table A1 in the Appendix lists the most active press preferences among candidates, whereas closed lists, tweeting MEPs. usually known as ‘party-centred’, have voters cast their Given that tweets are short texts and that our aim is to votes for a party list (Carey 2007). In the EP context, investigate the conditions under which MEPs give at- electoral rules, although based on proportional represen- tention to EU issues, we operationalise the dependent tation, present variations in the ballot structure (Farrell and variable, attention to EU issues, as a binary variable Scully 2010). In candidate-centred system, candidates have indicating whether the tweet includes EU content or not. incentives to perform personalised campaigns (Farrell and Tweets coded as having EU content contain both general Scully 2010; Hermansen and Pegan 2023) and emphasise references to the EU and Europe, EU institutions and issues relevant to their local constituencies. MEPs from bodies, and top positions. To facilitate coding, we candidate-centred systems face fewer party constraints and elaborated a dictionary of keywords that unequivocally tend to prioritise issues that are important for their local refer to those topics (see Table A2 in the Appendix). We constituency over EU topics. Conversely, in party-centred assume that giving attention to EU issues is not incom- systems, candidates align with their national parties due to patible with giving attention to member states. Thus, party leadership’s controls of legislators’ re-selection. Thus, tweets referring to both the EU and the MEP’s country or they have fewer incentives for personalised, locally oriented nationality are coded as having EU content. Our coding campaigns, and may more readily reflect their party’s po- does not discern between political and non-political sition on EU topics. content, in contrast to Fazekas et al. (2021). We as- sume that when MEPs use the Twitter handle published in H4. MEPs from party-centred systems give more at- their official EP profile, they inherently engage in tention to EU issues in the EP election campaign than communication of a political nature with their audience. MEPs from candidate-centred systems Tweets containing at least one keyword in the dictionary take the value of 1, and 0 otherwise. Coding was per- formed automatically. To check the validity of automatic coding, a reliability test consisted of manually coding a Data and methods randomly selected sample of 500 tweets. The percent Our analysis is based on an original dataset including agreement between the automatic and manual coding was Twitter content posted by MEPs from national parties .92 (Krippendorff’s alpha = .84). In descriptive terms, during the 2019 EP campaign, covering the period from 49.4% of the tweets mentioned EU issues. Figure A1 in 29 April to 26 May 2019. Tweets were extracted from the Appendix displaying the means by country shows that MEP’s Twitter during the week following the election and attention to EU issues is higher in Estonia (.79), Croatia translated into English using Google Translate. The dataset (.70), and Germany (.66), and lower in Slovakia (.18), ultimately includes 61,648 tweets published by 379 MEPs Hungary (.26), and Austria (.29). Attention to EU issues from 150 parties in all 28 member states. It includes tweets in France (.56) and the United Kingdom (.53) is slightly (32.7%), retweets (53.8%), and replies (13.5%). In line with higher than the EU average. Table A3 in the Appendix previous Twitter-based research (Fazekas et al., 2021), presents the percentage of tweets with EU content in individual tweets (covering the three types) are our unit of relation to those mentioning MEPs’ own country or analysis. nationality. As a robustness check, we operationalise 6 Party Politics 0(0) attention to EU issues restrictively by assigning the value systems (open lists and single transferable vote) take the 1 only when tweets make direct reference to Europe and value of 0. the EU, and 0 otherwise. Several controls are included. As being in government To better describe our dependent variable, and partly limits the extent to which parties can obfuscate their position building on previous research on the polity-policy di- on European integration (Koedam 2021), we control for mensions of EU issues and the tone of EP campaigns, we whether the party was in the national government when the examine whether MEPs from EU-friendly parties, when tweet was published (=1) or not (=0). Data were taken from talking about Europe, focus on specific policy issues ParlGov (Döring and Manow 2019). Following Spoon (Braun et al., 2016) and whether Eurosceptics employ a (2012) and Braun et al. (2016), we also control for intra- harsher tone (Nai et al., 2022). To that end, we first party dissent over European integration. Data were taken created three dictionaries of keywords related to the most from the CHES question on intra-party dissent on European important policy issues for Europeans, namely economy, integration, with values ranging from 0 (party completely climate change and environment, and immigration, re- united) to 10 (party extremely divided). As a curvilinear spectively. The selection of issues was based on Euro- relationship between intra-party dissent and EU issue sa- barometer data (EB 91, spring 2019) and previous lience could be assumed, we also take the squared term (e.g. research on issue salience (e.g. Abou-Chadi et al., 2020; Braun et al., 2016). The national economic situation is also Braun et al., 2016). The dictionaries can be found in Table controlled for by including the relative GDP per capita and A4 in the Appendix. We then checked whether tweets unemployment rate in 2019, with data taken from Eurostat. with EU content referred to any of these three major We also control for MEPs’ gender (female=1) and age, with policy issues. Although the results need to be interpreted data taken from each MEP’s profile on the EP website. cautiously, they suggest that, when talking about Europe, Because tweets, replies, and retweets may reflect different MEPs from pro-EU parties pay more attention than communication strategies in EP campaigns (Fazekas et al., Eurosceptics to specific policy issues. Second, a senti- 2021), we control for the type of tweet. As research on ment analysis based on the well-known Linguistic In- communication strategies on Twitter in EP campaigns may quiry and Word Count database dictionary (Boyd et al., assume that original content (tweets and replies) matters 2022) shows that MEPs from Eurosceptic parties employ more than an endorsement (retweets) (Fazekas et al., 2021), a more negative tone than pro-EU ones when talking we perform robustness checks without retweets. Table A6 in about EU. In both analyses, the differences are statisti- the Appendix summarises the descriptive statistics of the cally significant (p < .05). dependent, independent, and control variables. Independent variables are operationalised as follows. For party position on European integration, we use data from the 2019 Chapel Hill Expert Survey (CHES) (Bakker et al., Analysis 2020), that provide scores on party position on European Our dataset is hierarchically structured, as tweets are nested integration based on expert judgement, with scores ranging in MEPs, who are nested in national parties, which in turn from 1 (strongly opposed to European integration) to 7 are nested in countries. To avoid violating the assumption of (strongly in favour of European integration). We use MEPs’ the independence of observations, multi-level regression party affiliation at the end of the 2014-2019 legislature. For models with MEP, party, and country levels are performed. parties not covered in the 2019 wave, we alternatively use As the dependent variable is dichotomous, multi-level scores from the 2014 CHES (Polk et al., 2017). For parties mixed effects logistic regression is the most adequate not coded in either CHES wave due to party splits or model. The results are reported in Table 1 as odd ratios subnational versions of national parties, we assign the value (Table A7 in the Appendix reports the p-values). We esti- of the original or national party, respectively (see Table A5 mate four models. Model 1 includes all independent vari- in the Appendix). Consistent with established research ables, to which Model 2 adds controls. Model 3 includes the practice, we exclude MEPs from the remaining 23 parties independent variables and the interaction and Model 4 adds not included in CHES data and those without party affili- controls. ation. As prior research highlights that extreme parties on The results corroborate H1 on the effect of EU position the left and right are drivers of the politicization of EU on attention to EU issues. MEPs from parties with pro-EU issues, we perform robustness checks to assess whether positions are significantly more likely to pay attention to attention to EU issues is affected by ideological extremism. EU issues than those from Eurosceptic parties. Based on As for the other independent variables, a binary variable Model 1, we calculate the predicted probability of at- indicates whether MEPs stand for re-election (=1) or not tention to EU issues at 95% CIs, keeping continuous (=0). For the electoral list system, our operationalisation is variables at their means and categorical variables at their based on Däubler and Hix (2018). Party-centred systems modes. The probability of giving attention to EU issues (closed lists) take the value of 1, whereas candidate-centred increases from .41 to .58 at the minimum and maximum Font 7 Table 1. Attention to EU issues in the EP19 campaign (odds ratio). M1 M2 M3 M4 Independent Variables EU position (H1) 1.15*** (0.04) 1.14*** (0.05) 1.13** (0.07) 1.1 (0.07) Candidate (H2) 2.22*** (0.24) 2.41*** (0.26) 1.97* (0.81) 1.76 (0.73) EU position * candidate (H3) 1.02 (0.07) 1.06 (0.08) Electoral list system (H4) 1.34* (0.24) 1.27 (0.22) 1.34 (0.24) 1.26 (0.22) Controls National government 1.18 (0.14) 1.19 (0.14) Intra-party dissent 0.99 (0.13) 0.97 (0.13) Intra-party dissent2 1.01 (0.02) 1.01 (0.02) GDP per capita 1.01 (0.01) 1.01 (0.01) Unemployment rate 0.99 (0.02) 0.99 (0.02) Gender (female) 1.05 (0.1) 1.05 (0.1) Age 1.01** (0.01) 1.01** (0.01) Tweet type (ref. = Reply) Tweet 2.28*** (0.07) 2.28*** (0.07) Retweet 3.06*** (0.09) 3.06*** (0.09) Random effects Country 0.09 (0.05) 0.07 (0.05) 0.09 (0.05) 0.07 (0.05) Party 0.04 (0.05) 0.07 (0.06) 0.04 (0.05) 0.07 (0.06) MEP 0.59 (0.07) 0.56 (0.06) 0.59 (0.07) 0.55 (0.06) Constant 0.27*** (0.07) 0.06*** (0.03) 0.29*** (0.11) 0.07*** (0.11) N 61,648 61,648 61,648 61,648 ***p < 0.01, **p < 0.05, *p < 0.1; standard errors in parentheses. values of EU position, respectively (see Figure A2 in the 95% CIs. Although no significant difference is found for Appendix). Statistically significant differences are ob- parties strongly opposed to European integration, dif- served at the 95% CIs when comparing values ≤ 2 and ferences are significant for parties with scores of 3.5 or those ≥ 6 on a 1-7 scale. The results suggest that for more on a 7-point scale (see Figure 1). For parties parties with a pro-EU stance, highlighting EU issues in an strongly in favour of European integration, the proba- election viewed as being decisive was advantageous, bility of giving attention to EU issues increases from whereas Eurosceptic parties, faced with the uncertainty of .45 when MEPs do not stand for re-election to .66 when Brexit, preferred to downplay EU issues. Complementing they do. That is, for MEPs from pro-EU parties, giving the analysis, Figure A3 in the Appendix graphs the re- attention to EU topics is particularly advantageous when lationship between EU position and attention to EU is- they have vote-seeking motivations. By contrast, indi- sues, by country. The association is strong and positive vidual electoral incentives do not matter for Eurosceptics’ for most countries with parties that, on average, have high attention to EU issues. Note that the non-significance of pro-EU positions (e.g. Germany, Croatia, Belgium, Es- party EU position in the presence of the interaction term tonia) and negative for the one with the strongest Euro- suggests that the relationship between EU position and critical stance (Hungary). attention to EU issues is not uniform across different H2 concerning how standing for re-election affects ideological stances. This relationship is contingent on giving attention to EU issues also finds support. MEPs whether MEPs hold critical or pro-EU positions, which running for re-election are more likely to emphasize EU possibly indicates that standing for re-election does not issues than those not running. Based on Model 1, the allow the effect of EU positioning to manifest in a sig- probability of giving attention to EU issues increases nificant way. This suggests that the moderating role of from .41 for non-candidates to .58 for candidates at 95% standing for re-election is more pronounced for Europhile CIs (see Figure A2 in the Appendix). The results confirm parties. the expectation that MEPs from pro-European parties H4 on the effects of electoral list systems on attention to give more attention to EU issues in the EP election EU issues is not corroborated. Although the coefficient campaign when they stand for re-election (H3). We suggests that MEPs from party-centred systems are more calculate the predicted probabilities based on Model 3 at likely to talk about EU issues than those from candidate- 8 Party Politics 0(0) extremism. Using data from the CHES, and following Maier et al. (2019), a binary variable takes the value 1 if the party position on the left–right spectrum is between 0 and 2 or between 8 and 10 on an 11-point scale ranging from 0 (extreme left) to 10 (extreme right). As ideological ex- tremism and EU position are strongly correlated, models 1 and 2 are run without EU position. Extremism signifi- cantly decreases attention to EU issues in Model 1 (p < .05) and Model 2 (p < .1). Considering that Eurosceptic parties are placed at both ends of the ideological spectrum, the results are consistent with our main findings. Last, all models are performed including random effects for parties and MEPs and fixed effects for countries. Overall, the re- sults remain unaffected. Conclusion Our study explores MEPs’ emphasis on EU issues in EP elections. Using their communication on Twitter during the 2019 election campaign, we analyse party positions on European integration and individual electoral incentives. Figure 1. Predicted probability of attention to EU issues, 95% Drawing from research on European issue salience and EP CIs. election campaigns, we show that pro-EU mainstream parties emphasized EU issues and that MEPs from those parties pay more attention to EU issues when running for re- centred systems, the difference is not statistically significant election. We test our hypotheses with a novel dataset of (see Figure A2 in the Appendix). MEPs’ tweets posted in the four weeks before the day of the Regarding controls, MEPs’ age significantly increases election. Results contribute to research on EU issue salience attention to EU issues, although the effect’s magnitude is and EP campaigns and more indirectly on political elites’ quite low. Last, when MEPs post tweets or retweets, they are social media campaigning. more likely to pay attention to EU issues, compared with Our findings show that MEPs from pro-EU parties replies. were incentivized to emphasize EU issues, whereas Eurosceptic parties did not appear to benefit the most from highlighting them. Our interpretation of the results Robustness checks relies on the different type and intensity of the preferences We perform robustness checks on the sample, the dependent of pro- and anti-EU parties. Mainstream, pro-European variable, and the model. To check that the results are not parties’ perception that the 2019 election was a funda- driven by the most active MEPs on Twitter, we exclude mental moment motivated MEPs from those parties to tweets posted by the two MEPs with the most tweets (N = make claims about Europe to mobilise voters. By con- 57,031) as well as without the 10% most active ones (N = trast, Eurosceptic parties’ motivations to emphasise EU 55,303). Given the exceptional circumstances of the EP issues were toned down considering that Brexit negoti- election in the UK due to Brexit, we perform all models ations, as noted by Hobolt et al. (2022), functioned as a without United Kingdom MEPs (N = 49,972). The analysis deterrent for other member states. Our findings contribute is also conducted without retweets (N = 28,500). Results to debates on the relationship between party orientation remain largely unchanged in all cases. We run all models towards European integration and EU issue salience. with a restrictive operationalisation of the dependent var- They contrast with Braun and Grande’s (2021) findings iable. Partly based on Lehmann’s (2022) approach, it takes that, in the 2019 election, mainstream parties responded the value 1 only when tweets make direct reference to to Eurosceptic challengers by de-emphasizing European Europe or the EU. Another restrictive operationalisation of issues and were forced to open a debate only in countries the dependent variable takes the value 1 when tweets with high levels of EU politicization. Our findings, based mention EU issues (based on our dictionary) but do not on Twitter communication by MEPs from 150 parties in mention MEPs’ country or nationality. In both cases, the all member states, demonstrate that pro-EU mainstream results remain almost the same. We also check whether parties did not avoid EU issues and had greater moti- MEPs’ attention to EU issues is affected by ideological vations than Eurosceptic parties to emphasize EU issues. Font 9 While our analysis does not perform a country-level Furthermore, bridging the study of these dimensions with analysis of politicization, the results are robust in research on legislative behaviour and activity throughout showing that pro-EU mainstream parties expected to gain the EP legislature also merits in-depth study. more benefits than Eurosceptics by emphasizing EU is- More generally, our study engages with broader debates sues. The results also show that, when talking about on the nature of EP elections. Our findings add to recent European issues, pro-EU parties were more focused on literature challenging arguments about the second-order specific policy issues, as suggested by Braun et al. (2016), election contest and emphasising the idea that Europe while Euro-critical parties employed a harsher tone, as matters in EP elections. First, EU issues received attention shown by Nai et al. (2022). in roughly half of the tweets posted by MEPs in their Our analysis also reveals that electoral incentives derived communication strategies in the 2019 campaign, reflecting from MEPs’ re-election aspirations matter for attention to the salience of European issues in the public conversation. EU issues in EP elections. MEPs competing for votes are Second, MEPs from pro-EU parties considered that talking more incentivised to talk about EU issues than those not about the EU was advantageous, whereas those from Eu- seeking re-election. For candidate MEPs, emphasizing EU rosceptic parties downplayed the European card. Third, as issues can serve as a way to enhance their parliamentary mainstream parties treated the EU as a natural venue for activity, call for votes in the European campaign, and in- relevant policies, they potentially contributed to reinforcing crease re-election prospects. Moreover, MEPs pro-EU EU legitimacy. Finally, MEPs from pro-EU parties per- parties are more inclined to emphasize EU issues when ceived that giving visibility to EU issues was especially seeking re-election. These results have several implications: advantageous when seeking re-election. The findings pro- pro-EU party MEPs’ motivations may vary depending on vide novel evidence suggesting that a European electoral individual electoral incentives, and emphasizing EU issues logic was at work in the 2019 EP election and align with the is meant to mobilize voters. claim that EP elections are about Europe. Our findings unfold the interplay between party-level and legislator-level factors and its effects on attention to EU Declaration of conflicting interests issues in EP elections. More generally, they demonstrate The author(s) declared no potential conflicts of interest with re- that analysing attention to EU issues through the lenses of spect to the research, authorship, and/or publication of this article. MEPs has methodological advantages. Focusing on MEPs allows examining EU issue attention in EP elections via party members from all member states who simultaneously Funding serve as representatives in the EP. Moreover, focusing on The author(s) received no financial support for the research, au- MEPs does not take individual re-election incentives for thorship, and/or publication of this article. granted and helps to unveil how party- and individual-level factors matter for attention to EU issues. Beyond that, by ORCID iD putting the spotlight on MEPs, our contribution comple- ments previous research based on Euromanifestos and party Nuria Font  https://orcid.org/0000-0002-2629-6829 communication. Our study is not without its limitations. Analysing an EP Supplemental Material election considered to be the most politicised ever raises Supplemental material for this article is available online. concerns about the extent to which the results are context- driven. Future research would benefit from comparing the 2019 election with other elections and verifying whether our References results suggest relatively stable trends. Second, although Abou-Chadi T, Green-Pedersen C and Mortensen P (2020) Parties’ party members are assumed to share positions on the key policy adjustments in response to changes in issue saliency. dimensions structuring political conflict, it might be that West European Politics 43(4): 749–771. party members who are MEPs are more inclined to em- Adam S and Maier M (2011) National parties as politicizers of EU phasise EU issues than those operating in domestic politics. integration? Party campaign communication in the run-up to Further studies could explore EU issue attention on social the 2009 European Parliament election. European Union media platforms in EP elections by both national and EU- Politics 12(3): 431–453. level actors. Additionally, our study is about EU issue Adam S and Maier M (2016) Campaigning on or downplaying salience—the most basic dimension of politicization—but European Union integration? Explaining the salience parties disregards polarisation and actors’ engagement. Developing attach to EU matters in European Parliamentary election expectations about how the interplay between party and campaigns. In: van der Brug W and Vreese de CH (eds) Un) individual electoral incentives affect other dimensions of intended Consequences of European Parliamentary Elec- politicization may be a fruitful line of research. tions. Oxford: Oxford University Press, 148–170. 10 Party Politics 0(0) Bakker R, Hooghe L, Jolly S, et al. (2020) ‘2019 Chapel Hill Fazekas Z, Adrian Popa S, Schmitt H, et al. (2021) Elite-public Expert Survey’. Version 2019.1. Available on Chesdata.eu. interaction on twitter: EU issue expansion in the campaign. (Chapel Hill, NC: University of North Carolina). European Journal of Political Research 60(2): 376–396. Balmas M, Rahat G, Sheafer T, et al. (2014) Two routes to per- Grande E and Hutter S (2016) Introduction: European integration sonalized politics: centralized and decentralized personali- and the challenge of politicisation. In: Hutter S, Grande E and zation. Party Politics 20(1): 37–51. Kriesi H (eds) Politicising Europe. Integration and Mass Boyd RL, Ashokkumar A, Seraj S, et al. (2022) The Development Politics. Cambridge: Cambridge University Press, 3–31. and Psychometric Properties of LIWC-22. Austin, TX: Hermansen SSL and Pegan A (2023) Blurred lines between University of Texas at Austin. https://www.liwc.app electoral and parliamentary representation: the use of con- Braun D (2021) The Europeanness of the 2019 EP elections and stituency staff among members of the european parliament. the mobilizing power of European issues. Politics 41(4): European Union Politics. Online ahead of print DOI: 10. 451–466. 1177/146511652211499. Braun D and Grande E (2021) Politicizing Europe in elections to Hobolt SB and de Vries C (2015) Issue entrepreneurship and the European parliament (1994–2019): the crucial role of multiparty competition. Comparative Political Studies 48(9): mainstream parties. Journal of Common Market Studies 1159–1185. 59(5): 1124–1141. Hobolt SB and de Vries CE (2016) Turning against the Union? The Braun D and Schäfer C (2022) Issues that mobilize Europe. The impact of the crisis on the Eurosceptic vote in the 2014 Eu- role of key policy issues for voter turnout in the 2019’, ropean Parliament elections. Electoral Studies 44: 504–514. European Parliament election. European Union Politics 3(1): Hobolt SB, Popa SA, Van der Brug W, et al. (2022) The Brexit 120–140. deterrent? How member state exit shapes public support for the Braun D and Schmitt H (2020) Different emphases, same positions? European Union. European Union Politics 23(1): 100–119. The election manifestos of political parties in the EU multilevel Hooghe L and Marks G (2009) A post-functionalist theory of electoral system compared. Party Politics 26(5): 640–650. European Integration: from permissive consensus to con- Braun D, Hutter S and Kerscher A (2016) What type of Europe? straining dissensus. British Journal of Political Science 39(1): The salience of polity and policy issues in European Par- 1–23. liament elections. European Union Politics 17(4): 570–592. Hutter S and Kriesi H (2019) Politicizing Europe in times of crisis. Carey JM (2007) Competing principals, political institutions and Journal of European Public Policy 26(7): 996–1017. party unity in legislative voting. American Journal of Po- Hutter S, Braun D and Kerscher A (2016) Constitutive issues as litical Science 51(1): 92–107. driving forces of politicisation? In: Hutter S, Grande E and Däubler T and Hix S (2018) Ballot structure, list flexibility and Kriesi H (eds) Politicising Europe. Integration and Mass policy representation. Journal of European Public Policy Politics. Cambridge: Cambridge University Press, 25(12): 1798–1816. 137–155. Daniel W, Obholzer L and Hurka S (2019) Static and dynamic Koedam J (2021) Avoidance, ambiguity, alternation: position incentives for Twitter usage in the European Parliament. blurring strategies in multidimensional party competition. Party Politics 25(6): 771–781. European Union Politics 22(4): 655–675. De Vries C (2017) Benchmarking Brexit: how the British decision Koop C, Reh C and Bressanelli E (2018) When politics prevails: to leave shapes EU public opinion. Journal of Common parties, elections and loyalty in the European Parliament. Market Studies 55(S1): 38–53. European Journal of Political Research 57: 563–586. De Wilde P, Leupold A and Schmidtke H (2016) Introduction: the Kriesi H (2016) The politicization of European integration. differentiated politicisation of European governance. West Journal of Common Market Studies 54: 32–47. European Politics 39(1): 3–22. Lehmann F (2022) Talking about Europe? Explaining the salience Döring H and Manow P (2019) Parliaments and Governments of the european union in the plenaries of 17 national par- Database (ParlGov): Information on Parties, Elections and liaments during 2006–2019. European Union Politics. Online Cabinets in Modern Democracies. Development version. ahead of print DOI: 10.1177/14651165221143962. Dooley N (2022) Frustrating Brexit? Ireland and the UK’s con- Maier M, Bacherle P, Adam S, et al. (2019) The interplay between flicting approaches to Brexit negotiations. Journal of Euro- parties and media in putting EU issues on the agenda: a pean Public Policy, Online ahead of print DOI: 10.1080/ temporal pattern analysis of the 2014 European Parliamentary 13501763.2022.2050780 election campaigns in Austria, Germany and the United Eugster B, Jalali C, Maier M, et al. (2021) When do European Kingdom. Party Politics 25(2): 167–178. election campaigns become about Europe?’ West European Müller W and Strøm K (1999) Policy, office or votes? In: How Politics 44(7): 1425–1454. Political Parties in Western Europe Make Hard Decisions. Farrell DM and Scully R (2010) The European Parliament: one Cambridge: Cambridge University Press. parliament, several modes of political representation on the Nai A, Medeiros M, Maier M, et al. (2022) Euroscepticism and the ground? Journal of European Public Policy 17(1): 36–54. use of negative, uncivil and emotional campaigns in the Font 11 2019 European Parliament election: a winning combination. Spoon J-J (2012) ‘How salient is Europe? An analysis of European European Union Politics 23(1): 21–42. election manifestos, 1979–2004’, European Union Politics Nulty PY, Theocharis SA, Popa OP, et al. (2016) Social media and 13(4): 558–579. political communication in the 2014 elections to the European Stier S, Froio C and Schünemann WJ (2021) Going transnational? Parliament. Electoral Studies 44: 429–444. Candidates’ transnational linkages on Twitter during the Obholzer L and Daniel WT (2016) An online electoral con- 2019 European Parliament elections. West European Politics nection? How electoral systems condition representatives’ 44(7): 1455–1481. social media use. European Union Politics 17(3): Van Middelaar L (2016) The return of politics – the European 387–407. Union after the crises in the Eurozone and Ukraine. Journal of Polk J, Rovny J, Bakker JR, et al. (2017), ‘Explaining the salience Common Market Studies 54(3): 495–507. of anti-elitism and reducing political corruption for political Vasilopoulou S and Gattermann K (2021) Does politicization matter for parties in Europe with the 2014 Chapel Hill Expert Survey EU representation? A comparison of four European parliament data,’ Research & Politics Online ahead of print (January- elections. Journal of Common Market Studies 59(3): 661–678. March): 1–9. Reif K and Schmitt H (1980) Nine second-order national elections - a conceptual framework for the analysis of Author biography European election results. European Journal of Political Nuria Font, Associate Professor of Political Science at the Research 8(1): 3–44. Universitat Autònoma de Barcelona.