“Spitzenkandidaten” in the 2014 European Parliament Election
Does Campaign Personalization Increase the Propensity to Turn Out?
Politics, University of Manchester and MZES, University of Mannheim
Sara B. Hobolt
European Institute, London School of Economics
Sebastian Adrian Popa
MZES, University of Mannheim
Paper prepared for presentation at the ECPR General Conference
to be convened at the University of Glasgow, Scotland, 3-6 September 2014
(draft September 1, 2014: 13245 words overall)
The 2014 European Parliament (EP) elections are considered a turning point in the history of the
European Union. For the first time is there a direct link between the vote in these elections and
the nomination of the President of the European Commission. Consequently, the major political
groups each nominated a lead candidate, or “Spitzenkandidat”, for the post. We postulate that
these developments towards personalization increased the visibility and the mobilization
potential of the EP election campaign, hence calling for a gradual revision of the classical second
order election model customarily used to analyse these elections. Based on 28 nationally
representative post-electoral survey data collected EU wide by the 2014 European Election Study
in co-operation with the European Parliament, we analyse whether and how the presence of the
lead candidates influenced the individual propensity to participate in these elections. We show
that recognition of the candidates leads to a higher propensity to turn out, even when controlling
for a host of other individual-level factors explaining turnout and the context factors known to
facilitate participation. Furthermore their campaign efforts (both online and offline) also had an
impact on the propensity to turn out both directly and by reinforcing the effect of recognition.
It was the motto of the European Parliament ahead of the 2014 election of its members: “This
time it’s different.” And the 2014 election indeed was different on at least two accounts. The first
was that it was held in times of a lasting public debt crisis in parts of the Union. Austerity
measures were imposed on the debtor countries by the Eurozone government (European
Commission and the relevant part of the European Council of Ministers) together with the
International Monetary Fund (quotes). The economic consequences – rocketing youth
unemployment for example – of these measures were severe. Whether these economic
turbulences have set the scene for a “critical election” (Key 1955) by fostering a lasting
realignment between citizens and voters on the one hand and political parties on the other, is
The other important difference of the 2014 election was that the member parties of the
major political groups of the European Parliament rallied behind a common lead candidate (or
Spitzenkandidat as these people have commonly been called using the German term). For the
first time in 35 years of a directly elected European Parliament, the extra-parliamentary party
organisation of five major political groups of the European Parliament nominated a lead
candidate during their respective party conventions (or by way of primaries) in order to support
their local campaigns—and offered EP voters a choice regarding the next President of the
European Commission. It was a common understanding during the campaign that the
predominant political camp would also win the presidency of the European Commission.
However, it must be noted that this link between the electoral result and the selection of the next
EC President while in accordance with Art 17 (7) of the Lisbon Treaty on European Union was
not commanded by it. It rather was as a gamble that the European Parliament played in order to
further “democratize” the EP elections and at the same time increase its power vis à vis the
Council. As could have been expected, the Council did not immediately consent with this new
selection procedure. Even if it was finally accepted the decision was not unanimous as it used to
be in the past. A few pockets of resistance (viz. the governments of the United Kingdom and
Hungary) could not be persuaded to support the Council’s nomination as Commission president
of Jean Clause Juncker, whose EPP had won the election by a comfortable margin.
Five of seven (or eight if we count the non-affiliated as a group) groups nominated a lead
candidate : EPP (Jean Claude Juncker), PES (Matin Schulz), ALDE (Guy Verhofstadt), the
Greens (Ska Keller and José Bové), and the Left (Alexis Tsipras). The two groups which did not
come along belong to the Euro-sceptical camp in the EP: the European Conservatives and
Reformists (ECR) in which the British Conservatives played a leading role, and Europe for
Freedom and Democracy (EFD) in which the British UKIP was the strongest force. In what
follows we will analyse whether and how these lead candidates affected the voting behaviour of
EU citizens in the 2014 European Parliament election. Based on survey data of the 2014
European Election Study we concentrate our attention on three most visible of the five lead
candidates (Juncker, Schulz and Verhofstadt), and restrict our curiosity to electoral participation
(and ignoring for now the question of party choice). Did the lead candidates increase the
propensity of citizens to turning out (as compared to a hypothetical situation in which they would
not have run)? This is the leading research question of this article.
2 The Emergence of a Parliamentary System in the EU
At the heart of the argument in favour of “Spitzenkandidaten” is the expectation that it will
strengthen executive accountability in the European Union. It is well known that thee EU is a
hybrid system with a mixture of parliamentary and presidential features. Its legislative power are
shared between the Council and the European Parliament, and it has a ‘dual executive’ where
national governments in the Council and European Council possess long-term executive power
and set the overall political and legislative agenda, whereas the EU Commission has the sole
right of legislative initiative. Yet, the EU has until recently lacked mechanisms for citizens to
hold the EU executive to account, or “to throw the rascals out” of executive office, through the
process of competitive elections. Unlike a presidential system, there are no direct elections
where citizens can elect the president of the Commission (or the Council). Unlike a
parliamentary system there is not even a strong indirect link between the party choice in
parliamentary elections and the executive, at least not until recently. Prior to the Maastricht
Treaty, the Commission President was chosen unanimously by the national governments. The
public therefore had no real way of influencing the election of the EU’s executive or hold it to
account for its actions.
During the early decades of European integration this was not regarded as a problem,
since democratic legitimacy rested solely, in an intergovernmental manner, on the national
governments in the Council. However, the pooling of sovereignty at the European level, the
move away from unanimity in the Council (meaning that individual governments could be
outvoted) and the end of the “permissive consensus” in the early 1990s put pressure on the EU to
establish a ‘European’ electoral dimension, where voters could be directly represented at the
European level, rather than only indirectly through their national governments. Strengthening the
powers of the European Parliament was at the core of these reforms. First, the Amsterdam and
Maastricht Treaties strengthened the legislative powers of the EP, gradually making it a genuine
co-legislature with the Council. Second, the Maastricht Treaty (1993) introduced a new
Marsh 1998. despite the new powers of the European Parliament. or mobilize citizen interest in EP elections. Lodge. Due to their
perceived insignificance. Parliament also introduced hearings of Commissioners-designate in 1994. The European
Parliament had argued that in choosing the Commission President. where a majority of voters stayed at home. Hix and Marsh 2007. European elections would
really start to matter to citizens and this would bolster interest and turnout (e. The Amsterdam Treaty (1999) took matters further by requiring Parliament’s specific
approval for the appointment of the Commission President.
Turnout to European Parliament elections continued to decline in successive elections from 62
per cent in 1979 to only 43 per cent in 2009. 1995. since parties and election campaigns focused largely on domestic matters.
While these reforms clearly strengthened the powers of the European Parliament vis-àvis the executive institutions in the European Union.g.. and others cast a
vote in protest against national government or with their hearts without any regard to government
formation (e. Hix
second-order nature of European elections have been attributed to the fact that citizens generally
have little knowledge of policies implemented or promised at the European level by parties. Worryingly.“investiture procedure” where the Council must consult the European Parliament on their
nominee for the Commission president and Parliament’s approval was required before the
Member States could appoint the President and Members of the Commission as a collegiate
body. Schmitt 2005). they did not bring about
the genuine electoral connection between voters and EU policy-making that was hoped for. there was also evidence that the
elections failed in providing a strong democratic mandate for policy-making at the European
elections. the elections continued to be “second-order national elections” (Reif
and Schmitt 1980. and
. However. prior to that of the other
Commissioners. van der Eijk and Franklin 1996).g. they did little to strengthen the link between
voters and the EU executive.
2006.parties themselves often use these elections as opportunities to test their standing with the public
in terms of their domestic political agendas. While Euro-parties produce electoral manifestos.
These problems led scholars and politicians alike to suggest constitutional innovations
that could remedy the growing democratic deficit in the European Union. This is not
least owing to the fact that. European
election campaigns have tended to focus on domestic political matters and be dominated by
national political actors. unlike national parliamentary systems. voters are generally unaware of this
and Euro-parties have traditionally played a limited role in EP election campaigns. This idea of Europarties nominating competing candidates was discussed already in the 1990s. and the candidate nominated by the
winning party group would in turn be nominated by the Council and elected by the European
. by scholars such as
Simon Hix (see Hix 1997. Peter and de Vreese 2004). the extent to which the national parties
use these manifestos in their own campaigning has traditionally been minimal. However. The legislative process in the European Parliament
operates very much like in any national legislature with members belonging to EU-level political
groups – such as the centre-right EuropeanPeople’s Party (EPP) and the centre-left Progressive
Alliance of Socialists and Democrats (S&D) – that structure debate over and support for
legislation and they decide vital political issues (Hix et al. these elections have not been
genuine contests between competing government alternatives and over incumbent performance
records. In between European elections. It is therefore unsurprising
that citizens have limited knowledge of and interest in the European Parliament. The core idea was to inject real political and personalized
choice into the EP election campaigns by having competing Commission President candidates
with alternative political agendas nominated by Euro-parties. Instead. despite the
presence of traditional party politics at the European level. the European Parliament is largely
ignored by national media (Norris 2000. 1998). 2007).
If he does not obtain the required majority. or
. As Hix noted optimistically in his 2008
book on What’s Wrong with the European Union and How to Fix It. acting by a qualified majority. (Article 17(7) TEU. in turn. such changes could lead to
“public identification of the policy options on the EU table and the winners and losers in the EU. In the Lisbon Treaty the
investiture procedure was revised to emphasize that the European Council should ‘take into
account the elections’ before nominating and that the European Parliament subsequently ‘elects’
the Council nominee:
Taking into account the elections to the European Parliament and after having held
the appropriate consultations. emphases added).
The wording of the treaty is ambiguous when it comes to the powers of the European
Parliament to impose its own candidate. But the European Parliament seized upon the treaty
change by deciding that the European political groups would nominate lead candidates.
shall propose to the European Parliament a candidate for President of the
Commission. the European Council. in European Parliament elections also played a central role in the debates leading to the
(failed) Constitutional Treaty and. This candidate shall be elected by the European Parliament by a
majority of its component members. the Lisbon Treaty (2009). shall within one month propose a
new candidate who shall be elected by the European Parliament following the same
In short. and enhance public
interest.Parliament to become the President of the Commission. there would be democratic politics in the EU for the first time” (Hix 2008: 164). the
European Council. acting by a qualified majority.
These discussions about how to strengthen electoral accountability.
the European Parliament presented its main argument:
[The Parliament] urges the European political parties to nominate candidates for the
Presidency of the Commission and expects those candidates to play a leading role in the
parliamentary electoral campaign. arguing that this “would make concrete and visible the link
between the individual vote of a citizen of the Union for a political party in the European
elections and the candidate for President of the Commission supported by that party”2and
thereby increase the legitimacy and accountability of the Commission. and more generally the
democratic legitimacy of EU policy-making. In a resolution agreed on 22
November 2012. for the post of European Commission president. stresses the importance of reinforcing the political
legitimacy of both Parliament and the Commission by connecting their respective
elections more directly to the choice of the voters. which fully supported the
move towards Spitzenkandidaten. The hope is that this would mobilize citizens to take greater
interest in and participate in the elections in greater numbers.Spitzenkandidaten. in particular by personally presenting their programme
in all Member States of the Union. Secondly.
Commission Recommendation of 12 March 2013 on enhancing the democratic and efficient conduct of the elections to the
European Parliament (2013/142/EU)
. the politicisation of
European issues should also allow voters to vote on the basis of issues that matter to EU policy1
European Parliament Resolution of 22 November 2012 on the elections to the European Parliament in 2014 (2012/2829(RSP)).
These institutional resolutions thus echo the message found in the academic literature
concerning the key objectives of the reformed process of nominating and electing the
Commission president. The first aim is to transform the nature of elections to the European
Parliament by creating a genuine contest for the top executive job and a choice between
alternative political platforms.1
This message was reinforced by the European Commission.
In addition to these lofty democratic aims.
While the Parliament’s slogan that “this time is different” held plenty of promise. where similar issues are being debated at the same time (see Koopmans
and Statham 2010. the objective is that
by increasing electoral accountability in EP elections. in time. Hobolt 2014). Kriesi and Grande 2014. and to subsequently
reward or punish them for the degree to the fulfilled this mandate.
However. strengthen electoral accountability in the EU: EP elections will
also voters to provide the executive with a genuine democratic mandate. and thus to attract more voters to the polls and
create a clearer democratic mandate for the European Commission. as it eventually happened (Schimmelfennig 2014. for an opposite view Fuchs 2000 and more often).
By introducing its own candidate with the democratic legitimacy conveyed by the vote of
Europe’s citizens the European Parliament put significant pressure on national governments to
nominate the elected candidate to accept informally. Finally. this paper will focus on the extent to which there is any evidence that the
Spitzenkandidaten had the desired impact on the campaign and the vote. personalize the electoral campaign.
this does not amount to a Europe-wide public debate on the elections akin to what we know from
. if not formally. there
were clearly significant challenges to overcome for the “Spitzenkandidaten” to have any real
impact on the campaign and the elections. Not least the fact that the European Union lacks a
common public sphere with a common media. this will also contribute to the legitimacy
(so-called input legitimacy) of the European Union.
there may also be more prosaic inter-institutional reasons for introducing the Spitzenkandidaten. by raise the stakes of
the vote. or even a common language. to discuss alternative
political visions.making rather than treating the elections as a mid-term ‘beauty contest’ for national
governments. the Parliament’s right to
appoint the EU’s executive. While recent studies have shown an increasing ‘parallelization’ of public
spheres across Europe. This may.
Juncker (former Luxembourgian Prime Minister and head
of the Euro Group) and Schulz (President of the European Parliament) rather than politicians
elected primarily for their broad electoral appeal.g. the
five candidates had a total budget of 4. While the lead candidates had held important
posts inside the EU and in their own member states. the
procedures adopted by the two major groups to nominate their candidates resulted in the
nomination of two Brussels insiders. they were largely unknown outside their
country of origin before the start of the campaign. the effectiveness of the Spitzenkandidaten depended largely on the campaign
itself. According to the EU-Observer. Finally. however. Although these numbers should be taken with a grain of salt. and
they may have strong incentives to fight on domestic issues (e. the Spitzenkandidaten did make efforts to run a
distinctly European campaign. Moreover. they can give us
some idea about the activities of the candidates. This is discussed in the next section.
it is still national parties that dominate the election campaigns.com/euelections/124152).5 million Euro (see http://euobserver.
Spitzenkandidaten in the 2014 campaign
Despite the challenges outlined above. while the candidates were officially nominated by Euro-parties. and to a lesser degree in Italy).
Hence.other federal systems. with both Schulz and Juncker disposing of a
. which quotes party sources. opposition parties in opposition
to the national government) and even to deliberately disassociate themselves from the
Spitzenkandidaten (as happened in the UK. European voters
would also need to take notice of the competing candidates. For the European Parliament’s argument to be convincing. The lead candidates’ impact on
national campaigns was therefore largely determined by the extent to which national party
leaders and the national media involved the European candidates in their national campaign.
049 non-voters).5 million Euro. Verhofstadt spending under 1 million.
The survey was conducted by AMR GmbH Dusseldorf on behalf of the AECR.083 voters and 6. and broadcast on the internet. It gained quite significant traction being watch by 330000
viewers in Austria (i. English and German.". on Euronews and on
selected national channels.8 million people in German (i. we don't know where our bed is. The poll was in the field on 25 and 26 of May on a
sample base of 12.4
Not surprisingly these debates generated the most interest in the “home countries” of the lead
candidates: in Luxembourg (Juncker) and in Greece (Tsipras) where 36 and 26 per cent of
respondents respectively reported to had watched one of the debates whereas only 6 per cent of
Dutch and British citizens had seen any of the debates. 6% of the
market share). An important role was also played by the
language in which the debate was carried out. A further example of the debates is the of the 8th
of May. much more than other political talk shows in the two countries (cite the news
article). 13% market share) and 1.e.132 respondents across 15 EU countries (6.
This is probably best summarised by a quote of the campaign chief of the Juncker campaign.
The most eye-catching initiatives were the nine televised debates between the
‘Presidential candidates’ that took place between 9 April and 20 May 2014.budget of approximately 1. They were
conducted in French. the debate was carried out in German and
broadcasted in Austria and Germany. And
this seems to be true given the intensity of the campaign activity of the three most visible
This “funding hierarchy” also confirms that the three candidates we included in our surveys are the most visible and important
of the five. A post-election survey of citizens in 15 EU countries reveals that 15
per cent of European citizens claim to have seen at least one of the TV debates (AECR 2014).
In addition to the debate the candidates also had a substantial presence on the ground. Schulz and Juncker opposed one another.
Martin Selmayr: "Our wives don't know us anymore. while the other
two candidates has substantially smaller amounts of money available3.
Most of these were classical campaigning events such as meeting party
activists and party supporters.ebu. and probably more important. these numbers do not take into account that in some cases they visited several cities
or attended several campaign events in the same day (see Appendix 2 for a complete description
of the campaign events).
Furthermore. Germany and Malta. having approximately 110k twitter
gained special transaction during around the time of the TV debate. Still the campaign of both Schulz and Juncker each had its
specificity.html). ”#TellEUROPE” was trending in Austria. Germany. Second. Juncker had
several meetings with the heads of national government and other important national and
European political figures. For example during the 15
May debate. or being present by the
launching of national candidates.candidates. Schulz had several events in which he directly addressed trade union members
or factory workers.
When we take into account that he was the favourite to ensure the nomination as President of the
European Commission. days spent in the
country). but this is not surprising considering that the before mentioned groups are the
traditional base of the European Socialists. participating at large campaign gatherings. Netherlands and the UK and was mentioned in 110k tweets
(http://www3.ch/contents/news/2014/05/ebu-makes-history-with-the-eurov. these meeting were most likely an attempt to secure the nomination. For example he had private meeting with the German.
Juncker covered 17 countries and participated in 34 campaign visits (i. Finish and Latvian. In the two months prior to Election Day Schulz had 38 visits in 20 countries.
Portuguese.e. France. Prime ministers and the ex PMs of France. while Verhofstadt had a more “modest” presence only had 29 visits in 12 countries.
especially bearing in mind the somewhat ambiguous text of the Lisbon treaty. Greek. All in all
Schulz was the most active in the online environment. Polish. Greece. Belgium.
Last but not least the on-line campaign of the Spitzenkandidaten was not negligible. First.
6 Aggregate statistics speak a very clear language: the
proportion of respondent in representative surveys who claim to have participated in an election
while they did not is between 10 and 20 percent higher than the official participation rate as
reported by the national statistical offices (and further down the line by Eurostat).
. true and false. for a question mark Bernstein et al.5
Individual level studies of electoral participation in European Parliament elections
Individual level analyses of electoral participation as a dependent variable is a difficult task to
address. but also some of the Scandinavian
countries and the US. A number of studies has revealed that indeed the true and the false voters
are pretty much of the same kind so that at least for the study of the determinants of electoral
participation the phenomenon of over-reporting does not constitute a major problem (e.
and turned even lower over time. i. 2001).
Participation levels in European Parliament elections started out at a low level in 1979.g. in countries where
electoral registers are held. Britain is a good example here. was the last active twitter user as he is
not even among the top ten most “popular” European leaders. As a consequence. http://www.
We can differentiate between these two classes of voters.
If we want to understand the determinants of abstention and voting. individual-level analyses of EP electoral
The source of these number is the TNS leader watch available at. Probably
the most surprising is the fact the favourite.e. Verhofstadt also
had a remarkable presence with 26k followers and 105k mentions in the same period.
Rosenstone and Hansen 1993. The problem is the notorious over-reporting of survey respondents who claim to have
participated while in fact they did not. Cassel 2003.followers and almost 250k mentions during the two months before the elections.com/what-we-do/european-leaderwatch
The most frequently cited is social desirability. Juncker.tnsglobal. this over-reporting is
only a problem if those who falsely claim to have voted are indeed different from the true voters.
However. and the second is that Euro-skepticism is not a convincing explanation since in membercountries where a significant Euro-skeptical voter segment existed in the electorate. the most popular view expressed in the media is that Euroskepticism is a major driving force behind Euro-election abstentions (a recent example is The
Guardian of 19 May 2014). the authors here expect
deficient mobilization to be the main factor explaining non-voting. the tenor of the analyses of individual level participation seems to point in a
different direction. The “new East” of the European Union (from 2004 on) has a number of additional
examples on offer. Evans & Ivaldi 2012). for at least two reasons: the
first one is that campaign efforts of the competing parties in the past notoriously turned out to be
shallow. In line with the second-order elections model. Steinbrecher & Rattinger 2012. What is cause and what
is effect is the critical question here or in other words: which variable is causally prior to which
other variable in the model. 1998. But there are also a number of scholarly pieces of research pointing
in this direction (Blondel et al. or the British
UKIP.participation have concentrated on the meaning of non-voting. and more in particular whether
electoral abstention is indicative of critical or even hostile attitudes on the side of non-voters
about European integration in general. the Swedish June list. and/or the institutions and policies of the European
Union in particular. political
entrepreneurs in general did not fail to compete for these votes and represent them in the
European Parliament (as members of one or the other Euro-skeptical group of the house).
Results of individual level participation analyses obviously depend essentially upon the
causal structure that the analyst imposes on the data he or she analyses. Indeed.
Obvious examples here are the Danish Folkebevaegelsen. and can additional (control) variables therefore be ignored?7
This will be an important issue in the next section when we discuss our extensive set of control variables that we employ in
order to strengthen the validity of our central regression results.
The first is personalization: the addition of European faces and voices to the
campaign is expected to increase mobilization and contribute to an increase in turnout (cites). Upon that background.2. the campaign activities and the subsequent media
coverage they receive are likely to reaffirm the importance of the Spitzenkandidaten for the
individuals who are able to recognise them and thus reinforce the effect of personalization. Furthermore. This is not to say that Euro-skepticism nowhere
did play a role for shaping turnout in the past. Schmitt & van der Eijk
2003. we expect that the Spitzenkandidaten contributed to electoral
mobilization and help to raise turnout (H1). we anticipate that the
mechanism one and two have an additive interactive effect: personalization and campaigning
should reinforce one another. 2007. or never will become an important co-determinant
of non-voting. On the one hand.
There are several mechanisms through which the Spitzenkandidaten can increase
Therefore we expect that those who are able to recognize the Spitzenkandidaten benefited from
their mobilization potential and hence have a higher propensity to turnout (H1. 2002. Joslyn
.). 2008. But so far.Considering those concerns.
The second mechanism consists of the actual campaigning efforts of the lead candidates.1). a number of European Election Study based analyses of the
determinants of individual electoral participation have been elaborated in the past which all tend
to support the mobilization hypothesis (Schmitt & Mannheimer 1991. the main story goes in the opposite direction: the main determinant of
electoral participation has been shown to be mobilization rather than citizens’ attitudes about the
the other hand.
We expect that in countries where the campaign efforts of the Spitzenkandidaten are stronger
individuals are more like to report casting a vote (H1. van der Eijk & Schmitt 2009). previous research has shown that candidate’s visits are more effective for
individuals that have at least some basic previous knowledge of them (Fowler et al.
The EES part of the study was funded by a consortium of private
foundations8 and benefited in addition from the generous support of TNS Opinion. The data
collection was carried out by TNS Opinion in collaboration with its local partners between 30
May and 27 June 2014 (it started five days after the European Parliament elections and lasted for
four weeks). and the Portuguese Gulbenkian Foundation. Any campaign activity benefits from
public attention. All in all.
Last but not least the mobilization potential of the lead candidates can be also dependent
on “external” factors. King and Morehouse.
Data and methods
The present paper is the first that uses the European Election Study (EES) 2014 Voter Study. the Swedish Rijksbank
This is a nationally representative post-election survey that was realised in each of the 28
member countries of the EU.100
Lead by the Volkswagen Foundation and supported in addition by the Mercator Foundation.
this study was commissioned in collaboration with the Public Opinion Monitoring Unit of the
European Parliament. the more
attention there is.
. This study continues the EES tradition of post European Parliament
election surveys which started in 1989 (and actually in 1979 with an addition to the
Eurobarometer at the time).3). The most obvious one that comes to mind is the strength of the local
member party of his EP party group in a given country. we expect that the effect of candidate
recognition is stronger in countries where the candidates campaigned (H1. We therefore expect that the effect of campaigning on turnout is stronger the
larger the local party of a candidate is (H2). The stronger the national party is that supports the lead candidate. The sample is representative at the country level and it consists of roughly 1.and Ceccoli 1996. It is worth mentioning that for the first time in the history of EES.
including traditional items such as left-right and pro-anti EU self.and partyplacements.
Furthermore in Germany the sample was 1648 (consisting of two representative samples for West and East Germany) and the
United Kingdom where the sample was 1442. the
respondents where offered four response options and thus not only the three that applied to one
The exceptions are: Malta. some people in [OUR COUNTRY] did not vote in these elections”).net/voter-study-2014/
The master questionnaire in both English and French is available at the following link: eeshomepage.respondents in each EU member country. the survey consists of approximately 60 question units. Martin
Schulz and Guy Verhofstadt. All the interviews
were carried out face to face (by way of Computer Assisted Personal Interviews.
More details regarding the study can be found at http://eeshomepage. Cyprus and Luxembourg where only approximately 550 respondents were interviewed. One of
the main innovations of the 2014 study consists in a battery inquiring about respondents’
recognition of the Spitzenkandidaten. The core of the questionnaire is similar to the EES 2009
Voter Study. a PTV battery.
summing up to a total of 220 items. or CAPI)10. the total sample size being 300649.e. media use items. Although the question is not in an ideal open-end format. the personalization of the campaign) is a “name-party” recognition
battery.net/wpcontent/uploads/2014/05/Master-Questionnaire. This requires respondents to identify which EP party group or which national party
supports the nomination of the three most important candidates: Jean-Claude Juncker.
The main instrument which we use to measure the mobilizing potential of the
Spitzenkandidaten (i. A second innovation is a focus on the effects of the
economic crisis. a “most important issue” battery. The third and not the last is that this study uses issue questions that will also be
used in the Chapel Hill Expert Survey 2014. therefore allowing for a direct comparison between
expert placements of political parties and self-placements of voters11.
Excluding demographics. of which 338 interviews were conducted in Northern Ireland. and so on.pdf
The dependent variable of this paper is measured by a standard self-reported turnout
variable that also includes a memory cue (the date of the elections) and a “face saving” statement
(“For one reason or another.
the battery does not only measure the familiarity with
the Spitzenkandidaten but it also tests the ability of respondents to associate them with a specific
party.of the candidates but also a fourth and false one: “Socialists & Democrat (S&D)” (identified e. exposure to the campaign. internal political efficacy. In countries
where there was no party supporting one of the four EP groups.
measured as the share of the votes which the respective party gained in the 2014 EP election. in order to discourage guessing. Furthermore. We expect this effect to be moderated by the strength of the
candidates’ local party (i. level of political
discussion. “Liberals and Allies Group (ALDE)” (identified in German by the FDP) and
finally “The Greens” (identified in Germany by Die Grünen)12. the “Don’t
know” option was also offered.g. thus appropriately capturing the mobilisation potential of the three candidates. the biggest party was mentioned. These variables are
generally considered as proxies for political mobilisation (Gerber and Green 2000. campaign involvement and
contact by a party) and general political engagement (interest in politics.
In addition to candidate recognition we also employ a host of control variables as they are
customarily used to explain the propensity of turnout.e.
in Germany by mentioning the SPD). and news consumption).
In addition to measuring the mobilisation through personalization we also take into
account possible macro level effects. In operational terms we rely on the offline campaign
activity of the candidates indicated by the number of campaign visits of each candidate per
member country. “European People's Party (EPP)” (identified in Germany
by the CDU/CSU).
. All in all. The correct answer was therefore
neither a trivial one nor easy to guess.e. the local member party in the EP group) in the respective country. Rosenstone
In countries where two or more parties were expected to join an EP group. partisanship. we chose to use a dummy variable that takes the value 1 if the candidate visited
the country and 0 otherwise. The first group assembles variables
measuring campaign engagement (i. Given the distribution of the variable (see Appendix 1) and the limited
campaign time. only the name of the EP group was provided.
Verba et al.
A second group of factor is represented by social background variables that are indicative
for social integration and individual resources. The second is the legitimacy of European integration measured by trust in
EU institutions and evaluation of EU membership. Rosenstone and Hansen 1993.
. Thus we can safely assume that any effect
that candidate recognition might have on the propensity to vote is a result of the “mobilizing
effect” of the Spitzenkandidaten. lead to a higher
propensity to recognise the Spitzenkandidaten (i. 2007) and individual resources (Burns et
al. To be more specific one might claim that
once the decision to vote is taken. individuals start looking for all the relevant information that
would help them in making the best choice (Downs 1957) and this could. gender. reverse causality).e. marital status. church attendance. as soon as we
control for the level of political engagement of respondents we reduce the possibility that the
relation between recognition and propensity to vote is a result of previous knowledge or of
information acquired during the electoral campaign. 1995. Verba et al. The third and last factor here is the perceived
performance of the economy both at the socio-tropic and pocket-book level. However. 2001. rural vs. They include union membership.and Hansen 1993. in turn. operationalised by trust in the
national parliament. employment status. Historically these were among the first factors
used to explain individual turnout (Tingsten 1937. Zuckerman et al. and internet use. but also
age. immigrant status.
The final group of individual level factors for which we control are three attitudinal
constructs. The first is the legitimacy of the electoral process. Verba et al. both of which are known to be strong predictors of turnout. However these indicators also allow to control for the possible endogeneity
between candidate recognition and propensity to vote. urban residence. education. 1995).
1995). Verba and Nie 1972.
It is important to mention that our unit of analysis at the second level is party
systems rather than countries. and the British and Northern Irish one
in the UK case. whether other elections took place
at the same time as the EP elections. Wessels and Franklin 2009).
In order to test our hypothesis we proceed in two steps. and post-communist past of the country13.
It is worth mentioning that post-communism. Second we make use of a series of multilevel logistic regression models to
present the mobilisation effects of the Spitzenkandidaten on the propensity to vote in the 2014
EP elections. Franklin and Hobolt 2011. all being factors
that were shown in previous studies to have a strong influence on turnout in EP elections (e. In Belgium and the UK. and turnout in the previous national elections are
highly correlated. the level of GDP per capita. We use random intercepts
and random slops for the variables measuring candidate recognition and grand mean centering
for aggregate level variables (Enders & Tofighi 2007).At the macro level we control for compulsory voting. We would like to point
out that all independent variables where rescaled to have values between a theoretical minimum
of 0 and a theoretical maximum of 1. The analysis is conducted in R. there are effectively two party systems in
operation: the Wallon and the Flemish in the Belgian case.e. thus allowing for a straightforward comparison of their
effects (see Appendix 1 for a complete description of all variables). Controlling for any of the three yielded a very similar pattern of results (i. using the
lme4 package version 1.1-7.
. the significance levels for the
effects of interest were the same).g. Therefore our N at the second lever is 30 and not 28.
Schmitt 2005. First we present country level
descriptives and illustrate the aggregate relation between turnout and the mobilisation efforts of
we notice something
like a stand-still in 2014. However.6%
from 2004 to 2009 and 4% from 1999 to 2004. which is still extremely low
when compared to the turnout registered in first order national elections. The notorious turnout decline from 1979 seems to have come to an end. the Liberal Guy Verhofstadt who was
only recognised by 9 % of all respondents. Of course there are significant country differences as
the candidates are better known in their countries of origin and the neighbouring ones. In far
away countries such as the Czech Republic or the UK. More remarkable though are their campaigning efforts as between the
. we realise that the proportion of
citizens who recognised the candidates is not too impressive. the overall EU turnout only dropped by only 0. These numbers are even lower for the candidate of
the weakest of the three political groups that we consider. only around 5% of respondents were able
to correctly identify them. at a
first glance this was not reflected by the 42. if we
compare this number elections with the turnout at previous the EP election.
[Table 1 around here]
Although the motto for the current European elections was “This time it’s different”. Furthermore.4% compared to 2.
Turning to the campaign effects of the candidates.We start by presenting some descriptive statistics showing turnout levels in the 2014 EP
elections and the country-specific campaign of the Spitzenkandidaten (recognition and campaign
activity). 19 % of our respondents recognised
Juncker and 17% of them recognised Schulz.
Between 2009 and 2014.5% overall turnout level. in ten of the 28 countries we even
notice an increase in turnout.
At the aggregate level.
[Table 3 around here]
In Table 3 we present a series of multilevel models that test the potential mobilizing
effects of the Spitzenkandidaten. Hungary. Needless to say. Lithuania and Great Britain.
we are looking forward to the findings of the individual level analysis. A quick inspection
The only countries not visited by either of the candidates were: Estonia. We find that the
turnout difference goes in a positive direction in member countries where the candidates have
visited and where the proportion of respondents recognizing them was higher. Model 1 serves mostly as a reference model because it includes
all the relevant variables except for the recognition and the number of visits.
. and therefore in the turnout difference between 2009
these aggregate level associations are only suggestive of possible campaign effects on the
individual-level propensity to vote.three of them they managed to cover 24 of the 28 EU countries14 in a period of just two months
before Election Day.
[Table 2 around here]
A first glimpse into how campaign mobilisation (measured as recognition and campaign
activities) links to turnout is presented in Table 2. However. Given the general direction of these associations. We are interested in whether the lead
candidates made a difference “this time”. we notice that the campaign
characteristics of the three lead candidates all point in the same direction. there is not a single statistically significant relation between
turnout change and the above mentioned variables. however.
but given the multicoliniarity
between candidate recognition (the correlation between the recognition of Schulz and Juncker is
0. had a
substantial effect on the individual’s propensity to vote. And we are confident that this is not an
All predicted probabilities were computed using simulations based on the normal distribution of coefficients. What needs to be
noted is that all subsequent models have a better fit than Model 1.
In the case of Verhofstadt the size of the effect is much smaller and did not reach the
conventional levels of statistical significance.
All in all.of this model shows that there are no effects that go against previous findings. Not only do they go in the expected directions.
. the effects are similar for those who recognised Juncker.
First we looked into the effect of recognition and noted that these effects only reach statistical
significance in the case of Schulz and Juncker. Although
slightly smaller. the mobilizing effect of the Spitzenkandidaten.4% for those who did not recognise him. Recognizing him therefore did very little to boost the interest in the EP
elections and thus to mobilise individuals to vote. we chose to investigate these effects separately for each candidate.
these effects are also quite substantive (and comparable to the effects of most sociodemographics variables as well as the effects of some of the variables measuring political
engagement such as political knowledge and contact by party during the campaign).
Everything else being equal15 this corresponds to an increase of 7% (from 32% to 39%) in the
predicted probability that respondents who recognized him went on and cast a vote. measured as recognition. In the case
of Schulz we note that recognizing him increases the likelihood of casting a vote by 35%.
Model 2 examines the mobilizing effect of all candidates.61) and campaign visits. while keeping
all continuous variables at their mean and all categorical variables at zero. The predicted
probability that they cast a vote is 0. A possible explanation is that he is the least
relevant of the three candidates (regarding the race for EC presidency) and had practically no
chance to be nominated.5% compared to 0.
i. Schulz (Model 4) and
Verhofstadt (Model 5). and to mobilize turnout. the effect is slightly smaller. As we control for political engagement (both general and campaign specific)
it is much more likely that recognition actually measures mobilisation and not a facet of political
engagement that was not covered by one of the several indicators we use to measure this
concept16. In both cases the effects are substantial ones.endogenous effect.e. whether they campaigned or not in a country) had an impact on turnout.
If in the previous section we focused on a rather indirect measure of campaign
mobilisation. the predicted probability to vote for respondents who live in a country in which Schulz
campaigned is 45% compared to a baseline predicted probability of 32% for those who live in
another country.e. Everything else being
equal. The strongest effect can be noticed in the case of Martin Schulz who had the
highest number of campaign events (see Table 1) and covered by far the largest number of
Among other indicators our models take into account political knowledge and political interest that are the most likely
. These results reflect to a certain degree the campaign intensity of
the candidates. Finally. we only record a
statistically significant effect in the case of two of our three candidates.
Our basic expectation is that through their campaign visits (an event we can safely assume was
covered by national media) candidates managed to raise interest in and awareness of the
forthcoming European Parliament election. given the rather small proportion of respondents who actually recognised
Schulz and Juncker. In the case of Verhofstadt (Model 5). In what follows we look at how the offline campaign efforts
of the candidates (i. Again. the predicted
probability of a respondent living in a country that he campaigned is 47% compared to 39% for
those living in other countries. we need to acknowledge that the impact of their visits on the overall turnout
is bound to be rather small. name recognition.
Models 6 to 8.
[Figure 1 and 2 around here]18
As expected the highest propensity to vote is recorded in the case of those citizens who
recognised the candidates and resided in a country in which they campaigned.
A test of this hypothesis is presented in Table 3. were only about 10% more likely to vote than all other groups.
For the sake of simplicity we chose to only show the combined models. we expected that they would reinforce one another (H1. Confidence intervals that do not overlap are only a sufficient but not necessary condition for statistical
One should note that overlapping confidence intervals when plotting interaction effects are not necessarily suggesting a lack
of statistical significance. More specifically. the separate models are documented in Appendix 3. In the case of
Schulz (Figure 1) such a person on average is 10% more likely to vote compared to people who
reside in a country that Schulz visited but who were unable to recognise him as the S&D
nominee. Figures 1 and 2 support our understanding of these
Models excluding the support of the EP party group and its interaction with campaign visits yielded almost identical results.3).countries/political regions (21 in comparison Juncker who covered 18 and Verhofstadt who
covered only 14). The magnitude of the effects is lower for Verhofstadt (Figure 2) as
individuals residing in countries that he visited and were able to recognise him as the ALDE
In addition to their main effects we expected these two facets of mobilisation to have a
The lack of an interactive effect in the case of Juncker is not surprising given that we did not find
any effect of his campaign visits on turnout17. and 20% more likely to vote when compared to people residing in countries where
Schulz did not campaign. We note that the interaction
reaches statistical significance only in the case of Schulz (Model 7) and Verhofstadt (Model 8).
(2003) ‘Overreportung and Electoral Participation Research. N. Sinnott. In such countries campaigning increased turnout as it had a chance to activate the large
number of S&D supporters.L.
Burns. We only found such an effect in the case of Schulz (Model 7)..
Discussion (to be added)
AECR (2014) ‘Post EU Election polling project’. The picture is totally different in countries where the local S&D member party
is strong. K. J. The Private Roots of Public Action: gender. Schlozman.
Cassel. S. (1998) People and Parliament in the European
Union. 25–26 May.[Figure 3 around here]
Last but not least we expected the effect of campaign visits of our Spitzenkandidaten to
be stronger in countries where the local member party of the EP group that supported their
nomination is itself stronger. Fieldwork conducted by AMR GmbH
Dusseldorf. MA: Harvard University Press. C.
Figure 3 confirms that the Schulz campaign practically did not have any effect in countries with
low S&D support. on file with author. Svenson P. 2001.’ American
Politics Research 31(1): 81-92. & Verba. Oxford: Oxford University Press. and political participationo Title... Cambridge.A.
. H. (2006) ‘The news
coverage of the 2004 European Parliamentary election campaign in 25 countries’.
595–610. "The Legacy of Lethargy: How Elections for the European
Parliament Depress Turnout" Electoral Studies (30) 67-76. 2012.
“EBU Makes History with the EUROVISION DEBATE.
Franklin.A. pp. and Hix. 2002. and Roland. An Economic Theory of Democracy.html. S. Cambridge: Polity
Press. Fowler. “The Effects of Retail
Politics in the New Hampshire Primary. What’s Wrong With the European Union and How to Fix It.ch/contents/news/2014/05/ebu-makes-history-with-the-eurov.A. and Hobolt S. 2008. B. Constantine J. Noury. (2006) ‘Dimensions of politics in the European Parliament’. and Ivaldi..ebu. A. S. and Boomgaarden. Psychological methods.
Political Research Quarterly 65 (4): 855-867
Føllesdal. 2007. Lynn. G. New York: Harper Collins. & Tofighi.” 2014. Banducci.De Vreese.
http://www3. C.121–38. J.
Downs. H. 1957.
European Union Politics 7(4): 477–504. Spiliotes. Cambridge: Polity Press. 12(2). ‘Deriving a forcast model for European election turnout.. Journal of Common Market Studies 44(3): 533–62. (2011). A. M. G. 2012. S.. S.
Enders. (2006) ‘Why there is a democratic deficit in the EU: a response to
Majone and Moravcsik’. The Crisis of the European Union: A Response.
American Journal of Political Science 50(2): 494–511.
. J. & Linda L..” American Journal of Political Science 46(3). Centering predictor variables in cross-sectional multilevel
models: a new look at an old issue.
Vavreck. A. Semetko.K.H. C. EBU.
(1998) ‘Testing the second-order election model after four European elections’. and Morehouse D.
Hix. G. H. Democratic Politics in a European Union under Stress. Hobolt (eds). and Tilley. (2014) ‘Political debate in a polarizing Union’. C. P. 1955. in Lights. (2000) A Virtuous Circle: Political Communications in Postindustrial Societies. Electoral Studies 30(1): 4–15.
Norris. & Steve Ceccoli. Cramme.
Cambridge: Cambridge University Press. and Marsh.
British Journal of Political Science 28: 591–608.
. E. Mark R.
Hix. “A Theory of Critical Elections. D.141–170
Oxford University Press. and
S. (2005) ‘ Moving voters in the 2000 Presidential Campaign:
Local Visits. S.
King. Peter Lang Publishing Incorporated. (2007) Democratic Politics in the European Parliament. VO.. M. Politics and Political
Joslyn. Oxford: Oxford University Press. M.B. and Grande. by David Schultz.
Marsh. S. in O.
Hobolt. Campaign! Media. and Roland. and Marsh. Noury.. M. S.
Cambridge: Cambridge University Press. S. (2014) Blaming Europe? Responsibility without Accountability in the
European Union. Camera.Hix. 1996. Local Media’.” Political Behavior 18 (2).” The Journal of Politics 17(1): 145–55. (2011) ‘Second-order effects plus pan-European political swings: an
analysis of European Parliament elections across time’. The Journal of Politics 69(2): 495–510. (2007) ‘Punishment or protest? Understanding European Parliament
elections’. “Attentiveness to Television News and Opinion Change
in the Fall 1992 Presidential Campaign. J. A..
Posters: EU Campaign in Full Swing. In Opposing Europe.W.
Rittberger. European Journal of Political Research
8(1): 3–44. 2008. „Non-voting in European Parliament elections
and support for European integration.
Schmitt. and Schmitt. H. 2014. B. 1993.
http://euobserver. (2005) Building Europe’s Parliament: Democratic Representation beyond the
Nation State. (1999) Governing in Europe: Effective and Democratic? Oxford: Oxford
University Press. Selfies.
Schmitt. 208-237. Valentina. H. J. pp. and John Mark Hansen. pp. K. New York: Mcmillan.. Steven J. Vol. “Buses.
Schmitt. Participation. Wouter van der Brug und Cees
van der Eijk.Peter. eds. Southbend: University of Notre Dame Press. Oxford: Oxford University Press. Aleks Sczcerbiak and
Paul Taggart. Harvard Journal of Press/ Politics 9(4):
3–24. C. „There is not much Euro-sceptic non-voting in
European Parliament elections“. Mobilization. F.
Pop. (2005) “The European Parliament elections of 2004: Still second-order?”
West European Politics 28: 650-679. and de Vreese.” EUobserver. Hermann and Cees van der Eijk. (2004) ‘In search of Europe – a crossnational comparative study of
the European Union in national television news’. and
Democracy in America.
Scharpf. eds. 2007.
Reif. (1980) ‘Nine second-order national elections: a conceptual framework
for the analysis of European election results’. Hermann and Cees van der Eijk.com/eu-elections/124152.H. Oxford: Oxford University Press. 145-167.“ In European Elections and Domestic Politics –
Lessons from the Past and Scenarios for the Future.
Hermann and Renato Mannheimer. N. ed. 1991. D.S. with Ackaert.
Brettschneider. (2012) ‚‘Explaining Turnout in European Parliament
elections’ In The Europeanization of National Polities? Eds. van Deth and Edeltraud Roller. M. MI: University of
Schimmelfennig. H. “Legitimacy and electoral abstention in
European Parliament elections. Eine empirische Studie über die Motive der
Nichtwahl.. In: Political Democracy and Social
Maghalaes." European Journal of Political Research 19: 31-54. G. pp. Nie.H. pp. C. 2009.
Schmitt.“ In Europäische Integration in der öffentlichen Meinung. 2003.
. Toka. King and
Van der Eijk. available at http://europedebate. and Rattinger.
279-302. London: P. pp. P. H. 208-224.. Oxford: Oxford University Press. (2014) ‘The Spitzenkandidaten plot: the European Parliament as a strategic
Steinbrecher. and Franklin. (1937).
Van der Eijk. New York. (1996) Choosing Europe? The European
Electorate and National Politics in the Face of Union. Ann Arbor. S. J. Cees and Hermann Schmitt. „Die politische Bedeutung niedriger
Beteiligungsraten bei Europawahlen.” In The Legitimacy of the European Union after
Enlargement. eds. Harper & Row. Jacques Thomassen. Oxford: Oxford University Press. Political Behaviour: Studies in Election Statistics. M. Hermann and Cees van der Eijk.Schmitt. Opladen: Leske und Budrich. 1972.. F.ie/spitzenkandidatenploteuropean-parliament-strategic-competence-maximizer/ (accessed 20 August2014). "About Voting and Non-Voting in the
European Elections of June 1989.. Sanders. Jan W. Participation in America.
Wessels. Brady. Schlozman. S..Verba. Cambridge.. MA. H.
.. and Franklin M. B. In: Civic Voluntarism in
American Politics. (2009).E. ‚Turning out or turning off?’ European Integration 31 (5)
609-626. F. Harvard University Press. Voice and Equality.
20. country level descriptives.66%
* turnout level in the 2013 EP elections.73%
Candidate recognition **
Number of campaign visits
27.Table 1: Election turnout and candidate recognition.12%
92. ** unweighted figures
Note: first line in each cell is the Pearson correlation coefficient.87
. second line is the significance (one tailed).00
# visits Schulz
# visits Verhofstadt (VV)
# visits all
-. n=28 in
.Table 2: Country level correlations.00
0.121 (0.567 (0.366** (0.035)
-0.955*** (0.015 (0.(0.082** (0.600***
contact by party
0.113* (0.565*** (0.039)
0.Table 3: Effect of candidate recognition and campaigning on turnout.025
-0.040 (0.064*** (0.039)
interest in politics
-0.580*** (0.084** (0.126* (0.043
-0.548** (0.133** (0.448***
-0.209*** (0.(0.124*** (0.232)
(0.088*** (0.911*** (0.299)
0.085** (0.296*** (0.129** (0.211*** (0.181)
(0.090** (0.235 (0.885*** (0.052)
(0.143*** (0.132* (0.807*** (0.747***
0.085** (0.087** (0.572*** (0.279 (0.055)
0.296*** (0.234*** (0.044
0.083** (0.885*** (0.061)
trust EU institutions
0.156 (0.119 (0.036)
(0.209 (0.136** (0.036)
0.088** (0.109*** (0.052)
1.090 (0.288*** (0.039)
0.137** (0.295*** (0.294*** (0.318***
0. main effects
Model 1: Without
Model 2: All
Model 3: Juncker
Model 4: Schulz
(0.242*** (0.121*** (0.483*** (0.008
0. denotes p<0.302
denotes p<0.Juncker campaign
3.486*** (0.852*** (0.986*** (0.694** (0.060)
(0.568*** (0.265** (0.249
(0.134 (0.076* (0.116)
denotes p<0.081** (0.021
candidate campaign visits
rec candidate X visits
visits X EP group support
(0.924 (1.869*** (0.447***
23. **denotes p<0.040)
(0.137** (0. ***denotes p<0.084** (0.110)
(1.Table 4: Effect of candidate recognition and campaigning on turnout.207)
Model 8: Verhofstadt
interest in politics
exposure to campaign
trust EU institutions
-0.248 (1.108*** (0.064
0.574*** (0.322*** (0.091)
-0.108*** (0. interactions effects
Mode 6: Juncker
Model 7: Schulz
-0.081 (0.006 (0.274)
-0.1 .218*** (0.753***
contact by politician
trust national parliament
concurrent ntl election
Figure 1: Conditional effect of recognizing Schulz depending on countries where he
Figure 2: Conditional effect of recognizing Verhofstadt depending on countries where he
Figure 3: Conditional effect of Schulz campaigning depending on strength of PES
Did you yourself vote in the recent European Parliament elections?” recoded to 1 voted 0 did not vote. Original statements:
QPP23. original question available at the following link:
eshomepage. 2008. not don’t remember ”
and 1 reflecting “yes. “Don’t Know” answers
were coded as incorrect answers as we consider that they reflect a degree of ignorance similar to the one
reflected by incorrect answers (see Luskin and Bullock 2006. True/False
QPP23.9 Answers order was reversed and rescaled in the analysis.APPENDIX 1: Variable description
Turnout: question wording “European Parliament elections were held on the (INSERT CORRECT DATE ACCORDING
TO COUNTRY). the final
variables takes values for ” 0 reflecting “No. 2008. Sturgis et al. Sturgis et al. True/False
Interest in politics: original wording QP6. Sturgis et al. Hansen 2009a. 2008. Switzerland is a member of the EU.
Political discussion: a mean of three items (Cronbach alpha= 0. Hansen 2009a. individual component (level 1). “Don’t Know” answers
were coded as incorrect answers as we consider that they reflect a degree of ignorance similar to the one
reflected by incorrect answers (see Luskin and Bullock 2006. “Don’t Know” answers were
coded as incorrect answers as we consider that they reflect a degree of ignorance similar to the one
reflected by incorrect answers (see Luskin and Bullock 2006. There are [150% of real number] members of the [COUNTRY Parliament]. recoded 1 for those who correctly identify the Liberals and Allies
Group/(NATIONAL PARTY) as supporting Junker’s nomination and 0 otherwise. final variables recoded to take values between 0 reflecting a low frequencies and 1 high
frequency of discussion. 2008. Sturgis et al.
Political Knowlege: measure of political knowledge that ranges from 0 to 5.3.
. reflecting the correct True/False
answers given by each respondent to.2 Each Member State elects the same number of representatives to the European Parliament.1. Hansen 2009a. For one reason or another.87): d71_1 (discussion about national politics
matters) d71_2 (discussion about European politics matters) d71_3 (discussion about local politics
matters). remember”. True/False
QPP23. recoded 1 for those who correctly identify the Socialist & Democrats
/(NATIONAL PARTY) as supporting Junker’s nomination and 0 otherwise. some people in (OUR COUNTRY) did not vote in these elections. recoded 1 for those who correctly identify the European People's
Party /(NATIONAL PARTY) as supporting Junker’s nomination and 0 otherwise.
Schulz recognition: original question QPP24. responses was recoded to 0 reflecting “No.
Exposure to campaign: original wording QP8.net/wp-content/uploads/2014/05/Master-Questionnaire.
Verhofstad recognition: original question QPP24.pdf
Junker recognition: original question QPP24.4 NAME OF THE HEAD OF GOVERNMENT) belongs to (NAME OF CORRECT PARTY).
Explanatory variables. not at all” to 1 reflecting “yes totally”. “Don’t Know” answers were coded as incorrect answers as we
consider that they reflect a degree of ignorance similar to the one reflected by incorrect answers (see
Luskin and Bullock 2006. True/False
QPP23. Hansen 2009a.
Married: original question D7c.2. QP6.
Tertiary education: original question VD8.
Female: original question D10.
Partisanship: wording of question QPP21 Recoded in 1 yes if R is feeling close to any party and 0 if the response is
News consumption: variable computed as the maximum of three items QP9.
Trust EU institutions: original question wording QP6.2 (read in newspapers about the European election).
Retrospective economic evaluation: original question wording QPP15.3 (newspaper news): recoded to take values from 0 “never following the news” to 1 reflecting
“following the news every day/almost everyday”
Campaign involvement: mean of five items (Cronbach alpha= 0.
Economic situation: original question wording D60.3.Political efficacy: a mean of seven items (Cronbach alpha= 0.3 (talk to friends of
family about the European election).1 (watched a programme about the
European election). recoded to take values from 0 “had difficulties to pay bills
most of the time” to 1” Almost never/never had difficulties with paying the bills”.8. QPP11.1 (TV news).4 (attended a meeting or a rally about the European election)
Secondary edducatiom: original question VD11. QPP11.5 (read online about the European election). not contacted”
Trust national parliament: original question wording QPP1. recode to 1 “rural residence” and 0 “otherwise”.
Age: : original question VD11. recoded to 1 married and 0 otherwise. recoded to take 1”EU membership is a good thing” and 0
otherwise. recode to 1 “unemployed” and 0 “otherwise”. recoded to 1 “yes.7): QP11.2. recoded to take values form 0 reflecting no trust in
the national parliament to 1 reflecting high trust in the national parliament.4. recode to 1 “female” and 0 “men”. recode 1 for those who ended their education between the age of
16 and 19 and 0 otherwise. QP6. recode 1 for those who ended their education after the age of 20 and 0
otherwise. QPP9.7. QP9.1.
Unemployed: original question C14. final variables
recoded to take values between 0 reflecting a low sense of efficacy and 1 a high sense of efficacy. QPP9.
EU membership: original question wording QP7.76): QP6. ). recoded to take values form 0 reflecting no trust in the EU
institutions to 1 reflecting high trust in the EU institutions.
Rural: original question D25.2 (online news) and
QP9. recoded to take values from 0 “is a lot
worse” to 1 “is a lot better”.1
D72.2. reflecting internal and external political efficacy at both national and EU level. recoded to take values from 0 reflecting no
involvement to 1 reflecting strong involvement
Contact by party: original question wording QP12. contacted” and 0 “no. D72. QP11.
recode 1 if he campaigned in the country and 0 otherwise. recoded 1 if respondent and/or somebody else in the household is union
member and 0 otherwise
Immigrant: original question D2.3
(somewhere else internet usage). recode 1 if he campaigned in the country and 0 otherwise.1 (home internet usage).idea.cfm).2 (work internet usage) and D63. recode 1 if respondent of citizen of the country and 0 otherwise
Internet use: maximum of three items: D61.
Schluz campaign visits: number of campaigning days Schluz spent in a given country in the two month before the
Juncker campaign visits: number of campaigning days Junker spent in a given country in the two month before the
Verhofstadt campaign visits: number of campaigning days Verhofstadt spent in a given country in the two month
before the EP elections. D62.
. recode 1 if he campaigned in the country and 0 otherwise.Religious: original question D75.
Candidate nationality: code 1 if any of the candidates is a citizen of the given country and otherwise. recode to take values between 0 “never attends religious services” to 1 “attends
religious services more than once a week ”
Union member: original question D76. maro component (level 2)
Compulsory voting: coded 1 for countries that have compulsory voting and 0 otherwise (source:
Post-communism: coded 1 for countries with a communist/socialist regime before 1989 and 0 otherwise. recode to take values from 0 “never use internet” to 1 “use internet
every day”. regional or local elections took place in the same day as the
EP elections and 0 otherwise (source).
Explanatory variables. Election: coded 1 if any other national.int/vt/compulsory_voting.
46. Distribution macro variables
Turnout in previous
Appendix 2: Campaign Calendar
Table A2.1 :The Juncker campaign schedule
Type of Event
Keynote speech at IHK Akademie
Press Conference + Keynote Speech at CDU Congress
Press Conference (Nivelles) + Meeting EPP politicans (Nivelles/Antwerp) + Speech for
the press (Antwerp)
Campaign Speech at the CDA party conference
Visiting the farm of EPP President Joseph Daul + Speech on his candidacy and the
importance of the Common Agriculture Policy + Press Point
Campaign Event with National Coalition Party (EPP) + meeting with prime-minister +
Meeting with former Prime Minister + Live TV Debate with former Prime Minister
and MEP Ivars Godanis
Meeting with Prime Minister Laimdota Straujuma + press conference
Summit of EPP regional and local political leaders and Prime Minister Donald Tusk
Campaign with CDU (EPP) and German lead candidate David Mc Allister
Event launching the European campaign of Citizens for the European Development of
Meeting with Prime Minister and Deputy Prime Minister
Campaign Event with Alliance for Portugal (Lisbon)/Press Briefing (Athens)
Meeting with Prime Minister Antonis Samaras and other ministers
Receiving of the Youth-EPP campaign Vans + Press Conference
Campaign Event of the CSV
Campaign Event of the CDU with prime minister Angela Merkel
Campaign of the CDU in Hesse
. President of the Democratic Rally and
attendance of a ceremony celebrating the 10th anniversary of Cyprus's accession to
Keynote speech at the electoral Congress of DISY
Gala Dinner celebrating 10 years of Slovakia's EU Membership
Meeting with ÖVP-politicians (EPP). Campaign Events of ÖVP and press conferences
Event of Junge Union Berlin
Meeting with young EPP party activists from Italy
Rotenburg an der
Campaign Event with Partido Popular
Meeting with former Prime Minister Alain Juppé and participation in round table
Talk with Bundeskanzler a.4/30/2014
Meeting with CDU-politicians and press conference
Campaign of the Nationalist Party + Meeting with (former) Prime Minister + meeting
Meeting with President of the Republic.D.
Campaign Event of the SPD with Martin Schulz
Campaign Event of the PSOE (PES)
Meeting with NGO's
Speech at a three-day conference of the PES in
1.2:The Schulz campaign schedule
Meeting with Danish Social Democrats
Meeting with Finish Social Democrats
1.Table A2. Conference of the SPD for the EP Elections 2014
Round table discussion with European PES politicians
Campaign Event of the Luxembourg Socialist Party (LSAP)
Campaign Event of the SPD
Speech at the launch of PS France's European Election Campaign + on
economic governance to an audience of trade unionists entrepreneurs and
Round table discussion (Weimar)/ Campaign Event of the SPD (Erfurt)
Speech at a conference of work councils (Cottbus)/Campaign of the SPD with
local SPD candidates for the EP
Speech at the launch of the BSP European Election Campaign
Speech at the party Congress for the EP Elections of the PSD +
Campaign Event of the Irish Labour Party
Campaign Event of the Social Democratic and Labour Party (SDLP)
International Labour Day and Campaign with SLD
Visiting EU-financed project (Essen)/Campaign Event of the SPD + with Martin
Schulz (Dortmund) + SPD Campaign Event (Bremen)
SPD Campagn Events with Martin Schulz
Campaign Event of the SPD with Martin Schulz (Saarbrücken) + Meeting
working councils and trade union members
Visiting factories and two social projects (Lisbon) + press conference with
leader of Portuguese Socialist Party (Seguro) + Keynote Speech
Press Conference to present the policy programme
Speech in front of hundreds of trade union representatives in a steel fabric +
Meeting of Martin Schulz with Italian Prime-Minister and Party Leader of
Partito Democratico Matteo Renzi
Meeting with the Maltese Prime Minister Joseph Muscat + Campaign Event of
the Maltese Labour Party with Joseph Muscat
Campaign Event of the Spanish Socialists (PSOE)
Campaign Event of the PS France
Visiting agriculture and maritime industries + speech at the University of Brest
about youth unemployment
Round table discussion + Campaign Event Partito Democratico (PES) +
Campaign Event (Ljubljana)
Meeting local entrepreneurs. and public sector workers +
round table discussion + meeting local PS France politicans
Martin Schulz joint Swedisch Social Democrats (SAP) for door to door
canvassing + meeting local SAP politicans
Campaing Events SPD with Martin Schulz
Campaign Event SPD with Martin Schulz
Campaign Event PSOE (PES) and Catalan Socialist Party (PSE)
a. trade unionists.M./Aachen
Campaign Event SPÖ
Martin Schulz + Croation Social Democratic Party informing about the Balkan
Flood Situation/ Campaign Event of the PS France (Lyon)
Campaign Event SPD with Martin Schulz
3: The Verhofstadt campaign schedule
Type of Event
Campaign event of VLD
Participation on a debate about the future of the EU
Campaign Event of D66
Meeting with Ban-Ki Moon (Brussels)/Campaign Event with Croatian Liberals
IDS-DDI and HNS
Campaign Event with Italian liberals Scelta Europea
Campaign Event with Romanian liberals PNL
Participation in a book presentation about Arab spring
Congress of Italian liberals Scelta European
ALDE Press Conference
Campaign Event of German liberals FDP
Campaign Event with polish liberals Twoj Ruch
Press Conference to launch election campaign
Campaign Event with French Liberals Les Europeens
Campaign Event with Austrian liberals NEOS
Campaign Event with Italian liberals Scelta Europea
Discussion about digitalisation in Europe + Visiting google + campaigning
Campaign Event with Scelta Europea after TV-Discussion with other
Campaign Event with Open VLD
Campaign Event with French liberals
Meeting with liberal politicians + representatives from civil society and
entrepreneurs + campaign event in Barcelona
Presenting plan for Europe at the European Business Summit
Campaigning with Czech liberals ANO2011
Campaigning with Italian liberals Scelta Europea
Campaign Event with French liberals
Final Open VLD Campaign Event
Campaign Event with Drench liberals