Media personalization during European elections: The 2019 election campaigns in

Disclaimer/Complaints regulations If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible.


Introduction
The European Parliament (EP) placed great hopes in the Spitzenkandidaten (lead candidates) procedure, which was first employed in the 2014 European elections and entails that European party families nominate pan-European lead candidates for the president of the European Commission (EC).Following the nomination and subsequent election of former lead candidate Jean-Claude Juncker as Commission president in 2014, the EP evaluated the procedure as successful and argued, among other things, that the procedure 'fosters the political awareness of European citizens in the run-up to the European elections'. 1  However, the election of Ursula von der Leyen as Commission president in 2019 led to the provisional abandonment of the procedure and casted doubt on the impact of the procedure on European Union (EU) politics in the long run.Although many of these concerns address inter-institutional relations (see Hobolt, 2014), low levels of recognising lead candidates (Gattermann and de Vreese, 2020;Hobolt, 2014) as well as their limited potential for mobilization (Schmitt et al., 2015;Gattermann and Marquart, 2020) indicate that the Spitzenkandidaten have not yet fully resonated with European citizens.Why is this the case?
This contribution sheds light on this phenomenon by examining it through the lens of the personalization of politics at the EU level.Personalization implies that individual politicians come increasingly into focus at the expense of political parties and institutions (Rahat and Sheafer, 2007, p. 65).I argue that in order to understand the (limited) impact of European Spitzenkandidaten among European voters one needs to understand the personalization of EU politics in context.First, context is time dependent, which requires an assessment of the scope of personalization as a development over time.In other words, public attention paid to Spitzenkandidaten is likely to be contingent upon longitudinal changes in the personalization of EU politics more generally.In the absence of any such trend, it would hardly be surprising if the Spitzenkandidaten procedure failed to engage voters.Second, context also varies at the domestic level, in terms of political cultures as well as media and electoral systems.Put differently, some domestic contexts are more prone to personalized politics than others (see Gattermann, 2018;Holtz-Bacha et al., 2014;Kriesi, 2012;Šimunjak, 2017), which may explain potential cross-country variations in the attention paid to Spitzenkandidaten.
For these reasons, I analyse the personalization of EU politics in seven countries over the course of five European elections since 1999.I specifically examine the extent to which European broadsheets report on individual members of the European Parliament (MEPs) and Spitzenkandidaten.Media coverage is important for European voters to learn about candidates and issues at stake.If media do not report on candidates, and Spitzenkandidaten in particular, their impact on voter awareness and behaviour is likely to be limited.The findings indicate that EP election coverage generally does not become more personalized over time.This also holds for attention given to Spitzenkandidaten between 2014 and 2019, although they are often more visible than a typical MEP from the same party family.This leads me to conclude that the Spitzenkandidaten process is somewhat detached from mediated personalization developments at the national level.In the concluding section, I briefly discuss the implications of these findings.

I. European Elections in the Media
EP election campaigns have become more salient in European media outlets over recent years, although cross-country variation is still prevalent (Boomgaarden and de Vreese, 2016).Furthermore, scholars have observed that media coverage of EP elections has not become more European in nature, but continues to focus on domestic issues and political actors (Belluati, 2016;Boomgaarden and de Vreese, 2016;Schuck et al., 2011).Yet political actors at both levels have often been studied as a collective, that is, no distinction has been made between politicians, parties and institutions, which impedes any examination of personalization trends.
We know relatively little about the media visibility of individual candidates during EP elections, with the exception of their activities on social media (Daniel and Obholzer, 2020;Koc-Michalska et al., 2016).The Spitzenkandidaten have received slightly more scholarly attention (Belluati, 2016;Nulty et al., 2016;Braun and Schwarzbözl, 2019;Schulze, 2016), and studies indicate that there is considerable variation in terms of media or campaign attention paid to Spitzenkandidaten among individual candidates, political parties and country contexts.However, it is difficult to assess whether these findings pertain to the verypan-Europeannature of the Spitzenkandidaten procedure and individual candidates put forward, or whether they actually align with personalization trends at the domestic level (or absence thereof).Put differently, existing research has rarely offered insights into media personalization trends with respect to EP elections.

II. Why Mediated Personalization of European Elections?
Proponents of the personalization of politics thesis have argued that mediated personalization is linked to three main factors (see also Gattermann, 2018;Gattermann and Marquart, 2020).First, scholars have observed a trend towards partisan dealignment as citizens have become more detached from political parties (Dalton and Wattenberg, 2000).
As a consequence, voting behaviour has become more volatile in recent decades, which impacts on party competition in the electoral arena and offers opportunities for personalized electoral behaviour (Garzia, 2014;Lobo and Curtice, 2015).Second, and most prominently, scholars have identified mediatization processes as being responsible for the changing relationship between media and politics (Mazzoleni and Schulz, 1999;Strömbäck, 2008).Journalists have become more independent over time and have gradually gained the upper hand in setting the campaign agenda, and political communication has been increasingly influenced by media logic as opposed to party or political logic, which entails a greater focus on individuals and personalities, among other things (see Mazzoleni, 1987;Strömbäck, 2008).Third, developments in political institutions, such as electoral system reforms, may trigger changes in the way the news media report on politics (Rahat and Sheafer, 2007).
Despite these developments, there does not appear to be any universal trend towards the increased personalization of media coverage because findings differ across countries (Langer, 2007;Šimunjak, 2017;Vliegenthart et al., 2011), also with respect to election campaign coverage (Holtz-Bacha et al., 2014;Kriesi, 2012;Rahat and Sheafer, 2007).Regarding the mediated personalization of EU politics, I could not confirm such a trend with respect to individual commissioners or the Commission president at the expense of the institution itself amid considerable variation across several EU member states (Gattermann, 2018).Although I have argued that a focus on individual politicians instead of institutions and political parties could potentially make EU politics less abstract and hence more tangible and accessible to its citizens (Gattermann, 2018, p. 347), the reverse appears to be the case with the EC.Commissioners are not directly elected by EU citizens and their party affiliation plays only a subordinate role to their work in the EC, which undermines the provision of additional heuristics that help the audience understand who they are.These circumstances, alongside the rising complexity of EU politics that entail an abundance of individual responsibilities, may explain why journalists opt to report on the institution at the expense of individual commissioners if their aim is to make EU politics accessible to their audiences.However, this may be different for the EP because '[i]ts members are directly elected and, hence, incentivized to gain media attention, and journalists have greater responsibilities to hold them accountable' (Gattermann, 2018, p. 362).
So, why would we expect mediated personalization of EP elections?As second-order national elections, EP elections are prone to low voter turnout and high voter volatility because parties and voters prioritize domestic issues (Reif and Schmitt, 1980).This may be a favourable condition for personalized voting behaviour, although a preliminary empirical assessment showed that it mainly occurs among sophisticated voters and that party preferences still play a more decisive role compared to candidate evaluations (Gattermann and de Vreese, 2017).Likewise, few voters actually recognize Spitzenkandidaten (Hobolt, 2014) and thus their impact on voter turnout and vote intention is somewhat limited (Schmitt et al., 2015;Gattermann and Marquart, 2020).This may also indicate that media coverage of EP election campaigns as main source of information for voters is not very personalized (see Belluati, 2016: Schulze, 2016).
Nonetheless, high volatility also provides incentives for personalized campaigning.Generally, EP election campaigns tend to be less professionalized compared with national ones (Tenscher and Mykkänen, 2014).However, over the course of eight EP elections, campaign posters of national political parties in the Netherlands and Italy have placed greater emphasis on individual candidates, suggesting that campaigning has become more professionalized in that respect (Gattermann and Vliegenthart, 2019).As professionalization and mediatization processes are interlinked (Strömbäck, 2008, p. 240) and notwithstanding the potentially limited scope of such developments, this may also indicate that mediated personalization of EP elections has taken place over time.
Lastly, the introduction of the Spitzenkandidaten procedure has clearly been a structural innovation in EP elections.As electoral reforms are considered to impact on mediated personalization (Rahat and Sheafer, 2007), they may also trigger more personalized media coverage of EP electionsnot only with respect to Spitzenkandidaten themselves, but also regarding MEP candidates.However, such structural innovations have generally not led to more personalized broadsheet coverage of the EC, as new treaties have not had any impact on longitudinal changes (Gattermann, 2018).And while more contestation during EP election campaigns triggers more comprehensive media coverage (Schuck et al., 2011), long-term trends of the politicization of EU integration (Hooghe and Marks, 2009;Hutter et al., 2016) do not appear to play a crucial role for the mediated personalization of the EC (Gattermann, 2018).
Regardless of these considerations and given the mixed evidence from national politics, we are likely to find considerable variation across domestic contexts.While this may not necessarily hinder mediated personalization trends from emerging over time, as they could evolve at different levels of personalization, we should not ignore differences in media and electoral systems that potentially condition the scope of personalization trends (Holtz-Bacha et al., 2014;Vliegenthart et al., 2011).EP elections have been based on proportional representation since the coming into force of the uniform electoral procedures in 2002, but electoral institutions differ in terms of how much emphasis is put on individual candidates versus parties.For example, the Irish single transferable vote system allows voters to rank individual candidates across parties, whereas French voters choose between parties on the ballot.One assumption is that the more personalized the electoral system, the more personalized the campaigning of election contenders (Bowler and Farrell, 2011), which the media is likely to pick upon.
Likewise, traditional media systems differ in terms of levels of political parallelism, professionalization of journalists as well as commercialization and market competition, among other things (Hallin and Mancini, 2004).This has consequences for the degree to which media coverage is personalized.Studies, for example, find more pronounced personalization trends in the UK than in Germany (Holtz-Bacha et al., 2014) and the Netherlands (Vliegenthart et al., 2011), suggesting that the highly competitive British newspaper market plays a role in this.Differences in media systems are also reflected in the extent to which broadsheets personalize their news about the EC (Gattermann, 2018).Taken together, system-level differences at the domestic level may also explain why studies have found cross-country differences in the attention paid to Spitzenkandidaten (Belluati, 2016;Schulze, 2016), and it remains to be seen whether such patterns prevail with respect to the mediated personalization of EP elections over time.

III. Case Selection and Analysis
The study considers five EP elections  in seven countries, namely Ireland, the UK (England only), France, the Netherlands, Poland, Austria, and Italy.The countries are diverse in terms of the length of time they have been members of the EU; size and therewith number of elected MEPs; traditional media systems (Hallin andMancini, 2004, 2012); and electoral systems, although proportional representation has been employed in all countries under study and at all points in time.
For the content analysis, one left-leaning broadsheet per country was chosen: The Irish Times, The Guardian (UK), the French Le Monde, the Dutch De Volkskrant, the Polish Gazeta Wyborcza, the Austrian Der Standard and the Italian La Stampa (note: La Stampa is considered rather centrist).As Poland joined the EU in 2004, only four elections are considered for Gazeta Wyborcza; the study period of the Austrian broadsheet comprises three elections since 2009 due to lack of data availability.I am aware that the newspaper selection does not include other media outlets with different political leanings or different types of media.However, newspapers are well suited for a study of personalization as a longitudinal process (Holtz-Bacha et al., 2014;Langer, 2007;Vliegenthart et al., 2011).Moreover, personalization trends are, for example, comparable between newspapers and television (Kriesi, 2012, p. 831).
For each election period, six weeks are considered, namely four weeks prior to the final election day and two weeks thereafter, which resulted in a total number of 21,423 articles collected (see Table A1). 2 The data collection and automated content analysis procedures are described in Gattermann (2018, p. 352) and the coressponding supplementary material.The data were retrieved from Nexis Uni by applying several keywords pertaining to EU institutions, MEPs in general and the elections themselves.These data were then analysed by searching the content for the names of all elected MEPs (and all Spitzenkandidaten) in each election, bar MEPs from Scotland, Northern Ireland, Wales, and French overseas territories .Importantly, this means that we cannot generalize the findings at the individual level, because we do not take into account any news coverage of unsuccessful contenders.However, we are able to compare the results between periods before and after the elections (additional results are reported in the Appendix); the selection problem is no longer present after election day.A range of 291 to 313 elected MEPs is considered in each election across the respective countries (see Table A2).In 2014, there were six Spitzenkandidaten, while a total of 14 lead candidates contested the 2019 EP elections, because the Liberals presented a team of seven lead candidates.However, I consider only those nine Spitzenkandidaten from 2019 who took part in at least one of the pan-European televised debates (see Table A3).
To measure personalization trends, some studies have applied relative measures of individuals versus political parties (Kriesi, 2012;Vliegenthart et al., 2011).With respect to EP elections, the party political reference points for MEP candidates and Spitzenkandidaten are more difficult to identify using uniform measures for several reasons: national parties could be referenced with respect to domestic politics instead of EU politics; many national parties have joined electoral lists under one umbrella organization (e.g., Olive Tree in Italy); Spitzenkandidaten compete for transnational party families and sometimes there is no successful national party affiliated to any of them (such as Liberals in Poland in recent elections).Moreover, there are some independent MEPs, especially in Ireland; in other countries, independents often join electoral coalitions.Thus, to allow for comparability across countries and over time as well as between MEPs and Spitzenkandidaten, I analyse the absolute visibility of these politicians (for similar approaches see Langer, 2007;Šimunjak, 2017).Generally, all references to an MEP or Spitzenkandidat's last name were counted on the condition that each of them is mentioned at least once by their full name in an article.Variations and nicknames were included.References in the headlines count twice compared with references in the article text (see also Vliegenthart et al., 2011, p. 99).The overall number of mentions were averaged per country and election across all respective MEPs and Spitzenkandidaten included in this study (or any subsets of MEPs; see below).After that, all scores were averaged per 100 articles in order to compare the results across elections and newspapers.

IV. Findings
Figure 1 shows the average number of references to MEPs (disregarding Spitzenkandidaten) in each newspaper and election.It distinguishes between pre and post-election periods.Generally, Irish MEPs are the most visible: a typical Irish MEP receives between 2.9 (2019, pre-election) and 8.5 mentions (2009, post-election) in the Irish Times.Der Standard, La Stampa, De Volkskrant and Gazeta Wyborcza follow, despite considerable variation over time.The average values for English and French MEPs in their respective newspapers are lowest of all.This may be an indicator of electoral system differences as these MEPs, as well as Polish MEPs, are elected via closed party lists, while media system differences may not play a crucial role compared with the personalization of the EC (Gattermann, 2018).
However, we are mainly interested in variations over time to examine possible personalization trends.For most newspapers (Irish Times, Le Monde, Gazeta Wyborcza and Der Standard) there are no clear patterns.English MEPs receive continuously more attention in The Guardian over time (increasing from 0.2 to 2 references per MEP between 1999 and 2019 in the post-election period), as do Dutch MEPs in De Volkskrant (in the pre-election period values rise from 0.7 to 4.6 mentions).However, there is a reverse trend for La Stampa: over time, Italian MEPs receive less attention during pre and post-election periods (for the latter period, average mentions of MEPs drop from 4.9 in 1999 to 0.5 in 2019).In the Dutch case, the 2019 scores could be driven by the fact that two Spitzenkandidaten were Dutch, but mediated personalization had already taken place beforehand.Moreover, the depersonalization trend in La Stampa occurs despite the candidacy of Italian Spitzenkandidat Emma Bonino in 2019.
To examine whether these trends are driven by particular groups of MEPs, Figure 2 distinguishes between newly elected and returning MEPs.Newcomers could receive more attention than returning members.Alternatively, the latter have an incumbency bonus which may show up in newspaper coverage (Kriesi, 2012).Figure 2 shows that patterns are generally similar to the findings reported above.With few exceptions, returning MEPs often receive more attention on aggregate than newly elected MEPs in all newspapers bar the Note: Returning MEPs were in office at the end of the previous legislative term.

Media Personalization during European Elections
Irish Times.They also appear to be the main contributor to personalization trends in the British and Dutch newspapers, although newly elected MEPs sometimes receive more attention than returning MEPs in post-election periods (Figure A1).
As populist parties have gained more support in EP elections over time, Figure 3 distinguishes between MEPs from populist far-right domestic parties (based on the classification by Rooduijn et al., 2019) and other MEPs.Personalization in The Guardian is more applicable for MEPs from the UK Independence Party and the Brexit Party compared with other English MEPs, while the previously reported (de-)personalization patterns are similar for Dutch and Italian MEPs from populist far-right and other parties, albeit at slightly different levels.On aggregate, the former group of MEPs often receives more attention than other MEPs, but this varies over time.Prominent individual MEPs, such as Nigel Farage, Marine Le Pen or Matteo Salvini, may be driving this pattern, which we cannot examine using the current study design.The patterns are similar if we only assess post-election periods (Figure A2), with the exception of De Volkskrant in 2019 and Der Standard.
Having received an overview of the rather limited personalization trends for MEPs in European newspapers, we now turn to the Spitzenkandidaten.Figure 4 shows the average number of Spitzenkandidaten references in 2014 and 2019, distinguishing between pre and post-election periods.Though similar to Figure 1, the scale is larger in Figure 4, meaning that on aggregate, Spitzenkandidaten receive more mentions in European newspapers than MEPs.The British, Dutch and Austrian newspapers pay most attention to them, followed by the French and Italian newspapers.Remarkably, with the exception of De Volkskrant, which probably reported on the two Dutch Spitzenkandidaten, the pan-European lead candidates received less attention in 2019 than before.Likewise, in many newspapers the attention paid to Spitzenkandidaten is higher after the elections than before, but this is not the case in Gazeta Wyborcza in both years, The Guardian and Der Standard in 2019 and Le Monde and La Stampa in 2014.In the latter case this may relate to the fact that a Frenchman, José Bové, stood for the European Greens and the Greek Alexis Tsipras headed an electoral list in Italy; none of them had been considered as Commission president after the elections.
The final analysis constitutes a comparison between references to MEPs of the European People's Party (EPP) and their Spitzenkandidaten as well as between MEPs and Spitzenkandidaten mentions among the Social Democrats (S&D group in the EP, Party of European Socialists (PES) for Spitzenkandidaten) in Figures 5 and 6, respectively.
Individual Spitzenkandidaten are often mentioned more frequently than a typical MEP from the same party family, which also holds for the post-election periods (Figures A3 and  A4).The differences are most pronounced in the French, Dutch and Italian newspapers for both party families, as well as for the EPP in Der Standard.In the remaining newspapers, the differences are more marginal and reversed for the EPP in two instances: Manfred Weber received less attention in 2019 (6.3 mentions) than EPP MEPs in 2009 (seven mentions on average; but not post-election) in the Irish Times, and less attention compared to Polish MEPs in Gazeta Wyborcza in 2019 (0.9 versus 1.5 mentions).Nonetheless, there

Conclusion
The aim of the Spitzenkandidaten procedure was to raise voter awareness and participation.However, evidence suggests that their impact on electoral behaviour has been rather mixed (Gattermann and Marquart, 2020;Schmitt et al., 2015).The purpose of this contribution was to study this phenomenon against the backdrop of mediated personalization processes during EP elections.As such, this contribution provides an important account of the media context in the 2019 EP elections, as voters learn about their prospective representatives through media coverage, which can ultimately inform their voting decisions.I argued that it is important to understand the Spitzenkandidaten procedure in the context of longitudinal developments and variations of personalized politics at the domestic level.Indeed, the findings show, first, that there are no universal personalization trends in the EP election coverage by European newspapers.Only De Volkskrant and The Guardian have paid increasingly more attention to elected MEPs over the course of five EP elections, while the reverse is the case for La Stampa.The patterns for the remaining newspapers are mixed.There is generally considerable variation across country, while returning MEPs and those representing populist far-right parties tend to receive slightly more attention than newly elected and other MEPs, respectively.However, these categories are not mutually exclusive, that is, patterns could be driven by returning MEPs from populist far-right parties.As the analysis excludes unsuccessful MEP candidates, additional research is needed to examine individual-level variation.Second, Spitzenkandidaten altogether received less attention in 2019 compared with previously, except in the Dutch press, which suggests that EP election coverage has also not become more personalized at the pan-European level.However, and third, Spitzenkandidaten of the EPP and PES often stand out compared with the newspaper attention that a typical MEP from the same party family receives.In other words, the Spitzenkandidaten are somewhat detached from mediated personalization processes during EP elections at the national level.
This has both positive and unfavourable implications.On the one hand, Spitzenkandidaten have received special attention, most notably when the procedure was first introduced, which is a positive precondition for European citizens to learn about them (see Gattermann and de Vreese, 2020).On the other hand, this particularity could also be a disadvantage because European voters, except perhaps Irish voters, are not necessarily used to personalized media reporting about EP election candidates.For example, as attention to individual candidates decreases in the Italian press, voters may have become less sensitized to the Spitzenkandidaten procedure.However, as mediated personalization in the Dutch press has already occurred prior to the Spitzenkandidaten procedure, voters here may have developed an apprehension for pan-European candidates, especially given that there were two Dutch Spitzenkandidaten in 2019.
If the procedure were to be revived, potentially enduring depersonalization trends would not enhance socialization with the procedure among European voters.One possible pre-emption would be to put forward candidates who already served in the EC.As the analysis showed, returning MEPs often tend to receive more media attention in European newspapers.Nominating Spitzenkandidaten with experience of serving in the EC would allow the media, and ultimately citizens, to hold them accountable for their past performance in future European elections, which would eventually mitigate the EU's accountability deficit (Hobolt and Tilley, 2014).

Figure 1 :
Figure 1: Average number of references to members of the European Parliament (MEPs), pre and post-elections.

Figure 2 :
Figure 2: Average number of references to new versus returning members of the European Parliament (MEPs)

Figure 3 :
Figure 3: Average number of references to populist far-right versus other members of the European Parliament (MEPs)

Figure 4 :
Figure 4: Average number of references to Spitzenkandidaten, pre and post-elections

Figure 5 :
Figure 5: Average number of references to EPP members of the European Parliament (MEPs) and Spitzenkandidaten (SKs)

Figure 6 :
Figure 6: Average number of references to S&D/PES members of the European Parliament (MEPs) and Spitzenkandidaten (SKs)