How Do European Voters Make Up Their Mind?
Candidate Characteristics, Leadership Characteristics, and Candidate Communication in the Decision-Making Process of European Voters
Lotte Aalbers (10352791) University of Amsterdam
Graduate School of Communication
Research Master’s programme Communication Science
Master’s Thesis
Supervised by dr. Katjana Gattermann
1 February 2019 Word count: 7459
Abstract
This paper examined which candidate characteristics, leadership characteristics, and candidate communication European voters prefer in political candidates, specifically in
Spitzenkandidaten, candidates for the Presidency of the European Commission (EC). A
conjoint analysis survey was conducted in the run-up to the 2019 European Parliament (EP) campaigns among Dutch voters to see which candidate attributes matter for voters when it comes to their vote choice, and if certain voter characteristics (i.e., voters’ party affiliation, and their attitude towards the EU’s performance, utilitarianism, and strengthening) play a role in this decision-making process. Respondents were repeatedly presented with two
Spitzenkandidaten who varied among their gender, party, nationality, political experience,
leadership characteristics, and vision for the future of the EU, and they had to choose which candidate they would vote for as President of the European Commission, and they had to evaluate their support for both candidates separately. The results were analyzed using average marginal component effects (AMCEs) and confirm that voters preferred a Dutch
Spitzenkandidat compared to a non-Dutch Spitzenkandidat. Moreover, voters’ attitude
towards the EU’s strengthening (i.e., European integration) was found to explain vote choice
in candidate communication, it explained vote choice for Spitzenkandidaten with the vision nothing but the single market, and whose vision is doing much more together. The
contribution of this paper is twofold. It advances our general knowledge on the importance of candidate attributes, and especially the role of candidate communication in vote choice by showing that voters use candidate party, nationality, leadership characteristics, and vision to base their vote choice on. More specifically, by researching vote choice in the EU context, the results give insight into how European voters make up their mind during EP campaigns.
Keywords: candidate attributes, campaign communication, candidate evaluation,
How do European voters make up their mind?
The focus of political campaigns has shifted from political parties to political candidates over the years (van Aelst, Sheafer, & Stanyer, 2012). Traditional ties between voters and parties have weakened (Dalton, McAllistar, & Wattenberg, 2000), resulting in a more ‘candidate-centered politics’ where voters increasingly base their vote choice on leaders
and candidates (van Aelst et al., 2012). Subsequently, Caprara and Zimbardo (2004) propose a congruency model of political preference, which means that voters seek to match their own personalities, dispositions, social attitudes, motives, and values to that of political candidates: Voters increasingly rely on their personal judgments (i.e., their likes and dislikes).
Different candidate attributes could matter when voters try to match themselves to political candidates. Thus, a compelling question is which candidate attributes voters prefer in political candidates, and which attributes they use to evaluate these candidates. Empirical research has investigated this question with a research technique that is rather new in political science research, namely a conjoint analysis survey (Franchino & Zucchini, 2015, p. 224). Such a design makes it possible to include a large set of candidate attributes and to estimate the causal effects of these attributes simultaneously (Hainmueller, Hopkins, & Yamamoto, 2014, p. 2). Most conjoint analysis surveys included the candidate attributes age, gender, race/ethnicity, and political experience (Carlson, 2015; Carnes & Lupu, 2016; Franchino & Zucchini, 2015; Hainmueller et al., 2014; Kirkland & Coppock, 2018; Vivyan & Wagner, 2015, 2016). Regarding political communication research, Arnesen, Duell, and Johannesson (2018) were one of the firsts to include some form of candidate communication, namely candidates’ issue positions. Hence, more empirical research is necessary to investigate the
role of candidate communication in the decision-making process of voters, which is a first contribution of this paper.
Besides, most aforementioned studies were conducted in the context of national or local elections. A second contribution of this paper is to investigate the role of candidate attributes in a different context, namely among Dutch voters in the run-up to the 2019 European Parliament (EP) elections. These are second-order elections, which means most voters regard these elections as less important than the first-order elections (i.e., national parliamentary elections) (Reif & Schmitt, 1980). Consequently, voter turnout for the EP elections has never transcended 62%1 (Schmitt, Hobolt, & Popa, 2015, p. 352). During the 2014 EP campaigns, all major European Union (EU) political groups were asked to nominate their lead candidate in the race of becoming President of the European Commission (i.e., the key executive office of the EU), who were then called the Spitzenkandidaten. Gattermann, de Vreese and van der Brug (2016) showed that Dutch voters had a hard time making sense of these 2014 Spitzenkandidaten: Only the most knowledgeable voters were able to match their own party preferences or ideology to that of the Spitzenkandidaten. The 2019 EP elections are thus a perfect case to study the importance of candidate attributes because it is unknown how European voters make up their mind regarding their vote choice. The main research question (RQ1) of this paper is: Which candidate attributes (candidate characteristics, leadership characteristics, and candidate communication) are important in the decision-making process of European voters, and do certain voter characteristics play a role?
Most research investigating vote choice treat candidate characteristics and candidate communication separately, but increasing evidence can be found supporting an interaction between the two (e.g., Franchino & Zucchini, 2015).To explore this, it was investigated if vote choice for the leadership characteristics moderated vote choice for candidate
communication (here: a Spitzenkandidat’s vision for the future of the EU). Resulting in a second research question (RQ2): Is vote choice for a Spitzenkandidat’s vision moderated by vote choice for a Spitzenkandidat’s leadership characteristic?
To answer these research questions, a conjoint analysis survey was conducted in the run-up to the 2019 EP campaigns where Dutch participants repeatedly had to choose between two hypothetical Spitzenkandidaten who differed among their candidate attributes, and participants had to evaluate both candidates separately. The candidate attributes gender, party affiliation, nationality, political experience, leadership characteristics, and vision were
included to investigate which of these voters prefer in political candidates, and which of these voters use to decide their vote choice on. European political groups and politicians could benefit from the results of this study by anticipating on it with upcoming EP campaigns, which makes the study also socially relevant.
Theoretical background
A vote choice can be made by looking at the different attributes of the alternatives presented, their social characteristics, personalities, and policy positions (Caprara &
Zimbardo, 2004; Hayes, 2009; Popkin & Popkin, 1994), thus, many candidate attributes play a role in the decision-making process of voters. A conjoint analysis survey is able to include such a large set of attributes and to separately assess the effects of the attributes (Hainmueller et al., 2014, p. 2). Most conjoint analysis surveys covered the attributes gender, party
affiliation, and experience. Some studies included the attributes age (Arnesen et al., 2018; Hainmueller et al., 2014; Kirkland & Coppock, 2018), race or ethnicity (Carlson, 2015; Carnes & Lupu, 2016; Hainmueller et al., 2014; Kirkland & Coppock, 2018), and education (Arnesen et al., 2018; Franchino & Zucchini, 2015; Hainmueller et al., 2014). Arnesen and colleagues (2018) were one of the firsts to include a form of candidate communication by including candidates’ policy preferences on three issues. Here, I elaborated on that idea by examining a candidate’s vision for the future of the EU. Also, voter characteristics (i.e., voters’ party affiliation and attitude towards the EU) were analyzed as moderators to see if
Candidate Characteristics
The most common attribute expected to play a role in the decision-making process of voters is gender. The gender identity hypothesis confirms the idea that voters match
themselves to political candidates (Caprara & Zimbardo, 2004) because women are more likely to support female candidates and men are more likely to support male candidates
(Welch & Studlar, 1986). Moreover, Sanbonmatsu (2002) proposed that fifty percent of voters prefer one gender over another, women are more likely to have such a preference, and their preference would be for female candidates. However, research on aggregate vote totals in Britain, Canada, and the U.S. suggests that female candidates, in general, have an advantage compared to their male counterparts (Black & Erickson, 2003; Borisyuk, Rallings, &
Thrasher, 2007; Dolan, 1998; Smith & Fox, 2001). Additionally, Schwarz, Hunt, and Coppock (2018) reanalyzed data from 30 experiments looking into gender and vote choice, and found an overall preference for female candidates over male candidates.
H1a: In general, voters will be more likely to vote for a female Spitzenkandidat compared to a male Spitzenkandidat.
Voters also use gender as a cue to stereotype candidates’ political views and their
characteristics (McElroy & Marsh, 2010). Empirical research found that voters regard female candidates as more liberal and left-leaning than male candidates, even if both candidates belong to the same party (Huddy & Terkildsen, 1993; McDermott, 1997). The congruency model of political preference (Caprara & Zimbardo, 2004) provides reason to expect that more liberal, left-leaning voters prefer female candidates over male candidates.
H1b: More liberal, left-leaning voters will be more likely to vote for a female
Spitzenkandidat compared to a male Spitzenkandidat.
The candidate attribute nationality could be investigated due to the European context of this study. A potential reason for the absence of literature about this candidate attribute is
the national or local context of most studies investigating candidate attributes. Mostly, voters do not have the possibility to vote for a candidate with a different nationality, which is also the case with EP elections. However, when asked to decide voters could have a preference, and with voters matching their own personalities to that of political candidates (Caprara & Zimbardo, 2004), voters are expected to prefer a candidate with their own nationality over a candidate with a different nationality.
H2a: In general, voters will be more likely to vote for a Dutch Spitzenkandidat compared to a non-Dutch Spitzenkandidat.
By conducting large-sample surveys, Boomgaarden, Schuck, Elenbaas and de Vreese (2011) and de Vreese, Azrout and Boomgaarden (2018) distinguished five dimensions in the concept attitude towards the EU, namely negative affection, identity, performance,
utilitarianism, and strengthening. Specifically, strengthening relates to citizens’ attitude towards European integration; the deepening (i.e., integration into one country) and widening (i.e., integration of more member states) of the EU. It can be expected that the more positive voters are towards European integration, the more likely it would become for them to vote for a candidate with a different nationality than for a candidate with the same nationality.
H2b: The more positive voters’ attitude towards the EU’s strengthening, the more likely they are to vote for a non-Dutch Spitzenkandidat compared to a Dutch
Spitzenkandidat.
Candidate experience also influences vote choice: If all other factors were left out, a more experienced candidate wins it from a candidate with less or no experience in politics (Jewell & Breaux, 1988; Lee, 2008). Subsequently, it was found that journalists report more about politicians with more experience (Schoenbach, De Ridder, & Lauf, 2001; Sellers & Schaffner, 2007), but contradicting evidence was also found, more experience not necessarily means more media visibility for politicians (Gattermann & Vasilopoulou, 2015). Studies
including experience as a candidate attribute have used different conceptualizations and measures because experience heavily depends on context (e.g., country, political system). For instance, experience has been explained as (prior) political experience (Carnes & Lupu, 2016; Kirkland & Coppock, 2018), prior office (Carlson, 2015), or tenure in Parliament (Vivyan & Wagner, 2015, 2016). In the context of the EP elections, candidates could have experience in national politics, or in EU politics. Vote choice for one of these two is expected to be
moderated by voters’ attitude towards the EU’s performance because performance relates to how well the EU functions democratically and financially, and how well European institutions perform (Boomgaarden et al., 2011; de Vreese et al., 2018). It is expected that the more
positive voters are towards the EU’s functioning and the performing of the European institutions, the more likely it would be for them to vote for a candidate with experience in EU politics.
H3: The more positive voters’ attitude towards the EU’s performance, the more likely they are to vote for a Spitzenkandidat with experience in EU politics compared to a
Spitzenkandidat with experience in national politics.
Leadership Characteristics
Besides descriptive candidate characteristics, personal characteristics like leadership characteristics could play a role in the decision-making process of voters. Different
dimensions with various conceptualizations of leadership characteristics have been used in research over the years. Hence, Aaldering and Vliegenthart (2016) provided a valid overview of the leadership characteristics most often used in discussing Dutch leaders in the media, namely communicative performances (i.e., trying to inspire and mobilize the public, and knowing how to handle the media), consistency (i.e., reliability and stability of visions, views, and opinions), integrity (i.e., putting general interests before personal interests), political craftsmanship (i.e., having political skills: political intelligence and political experience), and
vigorousness (i.e., dominating decision-making by being a strong negotiator, and able to make difficult choices when necessary). Politics is mainly communicated to the public via the media, called mediatization (Kepplinger, 2002; Mazzoleni & Schulz, 1999). This makes media the dominant source of political news and information for voters (e.g., Gunther & Mughan, 2000; Strömbäck, 2008). Moreover, social media are increasingly used with respect to political communication (e.g., Stieglitz & Dang-Xuan, 2013). Consequently, voters are mainly, if not only, exposed to political candidates via the media. Since voters first contact with politicians is via the media, it is expected that voters regard candidates’ communicative performances as most important.
H4a: In general, voters will be more likely to vote for a Spitzenkandidat with the leadership characteristic communicative performances compared to a Spitzenkandidat with one of the other leadership characteristics.
As stated before, voters also use gender as a cue to stereotype candidates’
characteristics (McElroy & Marsh, 2010). Aaldering and Van Der Pas (2018) found that male and female party leaders were equally often discussed in terms of their consistency, and integrity. However, male party leaders were more often discussed in terms of their
communicative performances, political craftsmanship, or vigorousness in Dutch newspapers. H4b: Voters will be more likely to vote for a male Spitzenkandidat with the leadership characteristics communicative performances, political craftsmanship, or vigorousness compared to a female Spitzenkandidat with these leadership characteristics.
Candidate Communication
Voters also base their vote choice on social attitudes, motives and values: They vote for politicians that have the same values and beliefs as they do (Caprara & Zimbardo, 2004). Furthermore, theory on regulatory fit suggests that voters have to feel right about the message that is narrated because when the feeling is right, a more intense and a more deliberate
decision can be made (Aaker & Lee, 2006; Cesario, Grant, & Higgins, 2004).Thus, candidates need to communicate their values and beliefs to the public because from the moment they are nominated they have to get their message out there, thereby trying to mobilize support for themselves (Kim, Rao, & Lee, 2008).
Jean-Claude Juncker (2017), current President of the European Commission, wrote a White Paper on the Future of Europe in which he explains that the discussion about Europe’s
future most often focuses on the question: Should the future contain more or less Europe? Yet Juncker finds this view too simplistic and in the White Paper, he presents five different scenarios – visions – for the future of Europe, namely carrying on, doing less efficiently, doing much more together, nothing but the single market, and those who want more do more (see Appendix A for a policy overview of the scenarios). The 2019 Spitzenkandidaten should support one of these visions for the future of the EU, in order to mobilize support amongst voters that have the same beliefs as they do (Kim et al., 2008). Vote choice for these visions is expected to be moderated by three dimensions of voters’ attitude towards the EU, namely performance, utilitarianism, and strengthening2. Utilitarianism was not explained before, and relates to a country’s and someone’s personal benefits from the EU (Boomgaarden et al.,
2011; de Vreese et al., 2018). If voters are positive about how well the EU functions and how well European institutions perform, they are expected to vote for a candidate who does not want to change this, just carry on.
H5a: The more positive voters’ attitude towards the EU’s performance, the more likely they will vote for a Spitzenkandidat whose vision is carrying on compared to a
Spitzenkandidat with one of the other visions.
On the other hand, if voters do not think they and their country benefit from the EU, or if they are not in favor of European integration, they are expected to vote for a candidate who wants countries to only focus on the single market.
H5b: The more negative voters’ attitude towards the EU’s utilitarianism or the EU’s strengthening, the more likely they are to vote for a Spitzenkandidat whose vision is nothing but the single market compared to a Spitzenkandidat with one of the other visions.
Yet, if voters think they and their country benefit from being a member of the EU, they are expected to vote for a candidate who wants to motivate only those citizens and countries that recognize these benefits and want to do more.
H5c: The more positive voters’ attitude towards the EU’s utilitarianism, the more likely they will vote for a Spitzenkandidat whose vision is those who want more do more compared to a Spitzenkandidat with one of the other visions.
Last, if voters are in favor of European integration, they are expected to vote for a candidate who wants to focus on less issues, but tackle them more efficiently, or for a candidate who wants to do tackle issues together.
H5d: The more positive voters’ attitude towards the EU’s strengthening, the more likely they will vote for a Spitzenkandidat whose vision is doing less efficiently or doing much more together compared to a Spitzenkandidat with one of the other visions.
As stated before, most research analyzes candidate characteristics and candidate communication separately, but an argument can be made for an interaction between the two (Arnesen et al., 2018). Mondak and Huckfeldt (2006) conclude that candidate characteristics matter more when candidates take on an extreme and clear political position than when they take on a central and unclear political position. Yet, research does not provide a clear answer: Some researchers claim that they matter more for candidates with a similar ideology (Buttice & Stone, 2012; Green & Hobolt, 2008), and other researchers believe the exact opposite
(Clark & Leiter, 2014; Pardos-Prado, 2012). Due to these contrasting findings, a second exploring research question was investigated:
RQ2: Is vote choice for a Spitzenkandidat’s vision moderated by vote choice for a
Spitzenkandidat’s leadership characteristic?
Method
A conjoint analysis survey (see Appendix B for the questionnaire) was conducted. This method allows including a lot of attributes, measuring the effects of the attributes separately, and looking at interactions between them (Hainmueller et al., 2014, p. 2). Thus, it is the best method to test the proposed hypotheses and to answer the research questions of this study. Participants were six times presented with two Spitzenkandidaten who differed among their attributes. With every Spitzenkandidaten-pair they had to choose which they would vote for as President of the European Commission, and they had to evaluate their support for both candidates separately. The results give insight into how Dutch voters make up their mind during EP campaigns.
Sample
In total, 271 Dutch, eligible to vote (i.e., 18 years or older), citizens participated in the study. Forty-three responses were deleted3, leaving a total sample of 228 Dutch respondents. Convenience sampling was used to gather as many participants as possible and participants were recruited in two ways. One group (i.e., the student sample) was recruited online via the University of Amsterdam where undergraduate Communication Science and Psychology students could fill in the questionnaire for research credits. The second group (i.e., the
snowball sample) was recruited via social media (i.e., Twitter, Facebook, and LinkedIn) using snowball-technique, and these participants were not compensated for their participation. The student sample (n = 105, 46.1%) differed significantly from the snowball sample (n = 123, 53.9%) in terms of gender, age, education levels, and residence (see Appendix C for the
independent group t-test between the samples). Therefore, a dummy variable (student sample = 0; snowball sample = 1) was included in all analyzes models to control for the differences between the samples.
Data collection took place from 20 of November until the 31 December 2018. During this period, five Spitzenkandidaten were nominated, among which were two Dutch ones: Bas Eickhout was nominated as a co-candidate for the Greens-European Free Alliance (Greens-EFA) on the 24th of November, and Frans Timmermans was nominated on the 8th of December for the Progressive Alliance of Socialists and Democrats (S&D).
Conjoint analysis survey design
Table 1 shows the candidate attributes, the attribute levels, and the values of these levels presented in the questionnaire. The design of the study is a 2 (gender) x 4 (party affiliation) x 4 (nationality) x 4 (political experience) x 5 (leadership characteristics) x 5 (vision) factorial design, which means there were 3,200 possible Spitzenkandidaten-profiles.
Participants were randomly exposed to 6 x 2 (i.e., twelve) profiles with the use of the Conjoint Survey Design Tool designed by Strezhnev, Hainmueller, Hopkins and Yamamoto (2014). Except for nationality, all attribute levels had the same weight randomization (i.e., chance of being shown in the questionnaire). The value Dutch was assigned weight 1
2
⁄ and the other three values (i.e., Finish, Slovak, Spanish) were assigned weight 1
6
⁄. Both nationality and political experience were recoded into two attribute levels (see Table 1). There were no other restrictions or limitations because all possible profiles seemed plausible. Thus, the design for the analyzes is a 2 (gender) x 4 (party affiliation) x 2 (nationality) x 2 (political experience) x 5 (leadership characteristics) x 5 (vision) factorial design, leaving 800 possible profiles to which respondents could be exposed to. Hence, it is very likely that all possible profiles were shown and measured because the total sample was exposed to 2,736 Spitzenkandidaten-profiles (i.e., 288 respondents x 12 Spitzenkandidaten-profiles).
Independent variables
Gender. The candidate attribute gender was included to test H1 and could be either man or woman.
Nationality. To test H2, nationality was included as a candidate attribute. A
candidate’s nationality could be Dutch, Finish, Slovak, or Spanish. These nationalities were
chosen because these countries are geographically distinct, and they have different political (and media) systems (see e.g., Hallin & Mancini, 2004). This variable was recoded into Dutch or non-Dutch.
Political experience. Political experience was included as a candidate attribute to test H3. The candidate could have experience in national politics or experience in EU politics. Per level, two values could be presented in the questionnaire to reduce chances of respondents being exposed to two candidates with the exact same political experience (see Table 1 for the values).
Leadership characteristics. For H4, the candidate attribute leadership characteristics was included, with the attribute levels: communicative performances, consistency, integrity, political craftsmanship, and vigorousness. These levels were translated into values based on the study of Aaldering and Vliegenthart (2016), who conducted a computerized content analysis on leadership characteristics in Dutch newspapers. From the search strings they used, two words per leadership characteristic were chosen to describe the characteristics properly4 (see Table 1 for the values).
Vision. To test H5, the candidate’s vision for the future of the EU was included, using the five scenarios proposed in the White Paper on the Future of the European Parliament written by Juncker (2017). The five visions are carrying on, doing less more efficiently, doing much more together, nothing but the single market, and those who want more do more.
Propensity to vote. The moderator propensity to vote was measured to test H1b. Thus, to distinct liberal, left-leaning voters from other voters. For each party in the Dutch political system5, participants were asked how likely it would be on a scale of 0 (totally unlikely) to 10 (totally likely) that they would ever vote for that party (van der Brug, 2010). Dutch liberal political parties are D66, GL, and VVD, and Dutch left-leaning parties are DENK, GL, PvdA, SP6. The candidate attribute party was included to match voters’ propensity to vote for the Dutch political parties to voting for a candidate’s party (i.e., the European political groups7). Robustness checks show that voting for the European political groups (i.e., candidate attribute party) in all cases interacted with the propensity to vote for the affiliated Dutch parties (see Appendix J for a table of the interaction effects).
Attitude towards the EU. Three dimensions of attitude towards the EU were measured, namely performance (H3, H5a), utilitarianism (H5b, H5c), strengthening (H2b, H5b, H5d). Performance (M = 4.21, SD = .93), and utilitarianism (M = 5.10, SD = .85) were both measured using four items, and strengthening (M = 3.33, SD = 1.07) was measured using
three items, based on the work of Boomgaarden et al. (2011) and de Vreese et al. (2018). All items were measured on a 7-point Likert scale from 1 (totally disagree) to 7 (totally agree). A principal component analysis (PCA) was conducted on all items and showed three
dimensions: Three factors had an eigenvalue above 1, and a clear point of inflexion after three components in the scree plot (see Appendix D for a table of the PCA). This matches the literature about the dimensions of attitude towards the EU: performance, utilitarianism, and strengthening (Boomgaarden et al., 2011; de Vreese et al., 2018). Reliability (i.e., Cronbach’s alpha) for the dimensions (i.e., performance: α = .79; utilitarianism: α = .72; strengthening: α = .678) can be interpreted as reasonable reliable. Thus, the three scales were made by
averaging the scores on their corresponding items. A higher score means a more positive attitude towards the EU’s performance, utilitarianism, or strengthening.
The demographical variables measured are gender, age, level of education, and residence (i.e., province) (see Appendix C for means and standard deviations).
Dependent variables
To be able to answer which attributes voters prefer, both the forced vote choice and the evaluation of support were measured as dependent variables, as is common in existing conjoint analysis surveys (e.g., see Hainmueller et al., 2014; Vivyan & Wagner, 2015, 2016).
Forced vote choice. Respondents had to choose which Spitzenkandidat they would vote for as President of the European Commission. The question asked: “Based on this information, which one of these two Spitzenkandidaten would you vote for to become President of the European Commission?”
Evaluation of support. Respondents also had to evaluate their support for both
Spitzenkandidaten separately, on a 1 (very unlikely) to 7 (very likely) Likert scale: “How
likely would it be for you to support Spitzenkandidat 1 (or 2) as President of the European Commission?”
Analysis
Data from conjoint analysis surveys can be analyzed by estimating the average marginal component effects (AMCEs), which “represents the marginal effect of one attribute averaged over the joint distribution of the other attributes” (Knudsen & Johannesson, 2018, p. 6): The average change in the probability that a candidate will win votes when it includes a particular attribute level instead of the reference category (Hainmueller et al., 2014, p. 19). Per candidate attribute, one attribute level was chosen as a reference category, and the other levels were compared to this reference category. To estimate the AMCEs, linear regression models were designed in Stata where the dependent variables forced vote choice (model 1) and evaluation of support (model 2) were regressed on all levels of each attribute (except on the reference category). Linear regression models could be used because three assumptions proposed by Hainmueller et al. (2014, p. 8,9) were met: stability of the effects and no carryover effects, no profile-order effects, and randomization of the profiles (Knudsen & Johannesson, 2018, p. 7). In the tables, the results of both models are shown, but due to space constraints, just the results of model 1 are discussed in text. Whereas, when model 2 gives different results, it is indicated in a footnote. In addition, not all full models are presented in the tables but note that all models control for all measured candidate attributes, and sample.
Results
RQ 1, the importance of candidate attributes and voter characteristics, will be answered by testing the hypotheses. Figures 1 and 2, and table 2 show the main linear
regression models used to test H1a, H2a and H4a. The models are significant; therefore, they can be used to predict vote choice, but the strength of the prediction is low: Only 3% of the variation in vote choice can be predicted based on the measured candidate attributes. The remaining hypotheses were tested by adding interaction effects to the main model. These results are reported further below.
H1a assumed voters, in general, to be more likely to vote for a female than for a male
Spitzenkandidat. The results show that voters are not significantly more likely to vote for a
female than for a male Spitzenkandidat (see Table 2). Thus, H1a was not supported. H1b proposed more liberal, left-leaning voters to be more likely to vote for a female than for a male Spitzenkandidat, but the results do not show a significant moderation of propensity to vote for the liberal, left-leaning parties on voting for a female Spitzenkandidat9 (see Table 3, and see Appendix E for robustness checks for all Dutch parties). Therefore, H1b was rejected.
H2a expected voters to be more likely to vote for a Dutch than for a non-Dutch
Spitzenkandidat, and the results show that voters were significantly more likely to vote for a
Dutch than for a non-Dutch Spitzenkandidat (see Table 2). Thus, H2a was supported. H2b assumed more positive voters towards the EU’s strengthening to be more likely to vote for a non-Dutch than for a Dutch Spitzenkandidat. The results do not show a significant moderation of attitude towards the EU’s strengthening on voting for a nonDutch Spitzenkandidat, b =
-.01, t = -.65, p = .517, 95% CI [-.06, .03] (see Appendix F for a table of the linear regression model). Therefore, H2b was not supported.
According to H3, the more positive voters’ attitude towards the EU’s performance, the
more likely they are to vote for a Spitzenkandidat with experience in EU politics. There is no significant moderation effect of attitude towards the EU’s performance on voting for a
Spitzenkandidat with experience in EU politics, b = .00, t = .02, p = .980, 95% CI [-.04, .04]
(see Appendix G for a table of the linear regression model). Which means H3 was rejected. H4a expected voters, in general, to be more likely to vote for a Spitzenkandidat with the leadership characteristic communicative performances, than for a Spitzenkandidat with one of the other leadership characteristics. Compared to a Spitzenkandidat with
communicative performances, voters were significantly more likely to vote for a
Spitzenkandidat with the leadership characteristic integrity10 (see Table 2), rejecting H4a. H4b proposed voters to be more likely to vote for a male Spitzenkandidat with the leadership characteristics communicative performances, political craftsmanship, or vigorousness than for a female Spitzenkandidat with these leadership characteristics. Voters were not significantly more likely to vote for a male Spitzenkandidat with those leadership characteristics than for a female Spitzenkandidat with those leadership characteristics (see Table 4 on the next page). Therefore, H4b was not supported.
Table 5 shows the results for hypotheses 511. H5a assumed voters with a more positive attitude towards the EU’s performance to be more likely to vote for a Spitzenkandidat whose vision is carrying on. Attitude towards the EU’s performance did not significantly moderate
voting for a Spitzenkandidat whose vision was carrying on. Thus, H5a was not supported. H5b expected voters with a more negative attitude towards the EU’s utilitarianism or the EU’s
strengthening to be more likely to vote for a Spitzenkandidat whose vision is nothing but the single market. The results show no significant moderation effect of attitude towards the EU’s utilitarianism on voting for a Spitzenkandidat whose vision was nothing but the single market. However, the results show a significant negative moderation effect of attitude towards the EU’s strengthening on voting for a Spitzenkandidat whose vision is nothing but the single
market12, meaning that the less voters’ were in favor of European integration, the more likely they were to vote for a candidate who wanted to focus on nothing but the single market. Thus,
H5b was partially supported. Regarding H5c, voters with a more positive attitude towards the EU’s utilitarianism were expected to be more likely to vote for a Spitzenkandidat whose
vision is those who want more do more. No significant moderation effect of voters’ attitude towards the EU’s utilitarianism on voting for a Spitzenkandidat whose vision was those who
want more do more was found. Hence, H5c was not supported. H5d expected voters with a more positive attitude towards the EU’s strengthening to be more likely to vote for a
Spitzenkandidat whose vision is doing less efficiently or doing much more together. Opposite
to what was expected, a negative significant moderation effect of attitude towards the EU’s strengthening on voting for a Spitzenkandidat whose vision was doing less more efficiently was found, meaning that the less voters’ were in favor of European integration, the more likely they were to vote for a candidate who wanted to focus on less issues, but do it more efficiently. As expected, a positive significant moderation of attitude towards the EU’s strengthening on voting for a Spitzenkandidat whose vision was doing much more together, meaning that the more in favor voters were towards European integration, the more likely they were to vote for a candidate who wanted to do much more together. Thus, H5d was partially supported.
Turning to RQ2, it was investigated if voting for a Spitzenkandidat’s vision was moderated by voting for a Spitzenkandidat’s leadership characteristic. The results show no significant moderation effects of vote choice for the leadership characteristics on vote choice for the visions (see Appendix I for a table of the linear regression models).
Discussion
The aim of this paper was to examine which candidate attributes (i.e., candidate characteristics, leadership characteristics, and candidate communication) European voters use in their decision-making process, and to see if certain voter characteristics influence this process. Furthermore, by including candidate communication it was investigated if vote
choice for a candidate’s vision was moderated by vote choice for a candidate’s leadership characteristics. The results of the conducted conjoint analysis survey among Dutch voters in the run-up to the 2019 EP campaigns will be discussed below.
First, the importance of the candidate attributes in the decision-making process of voters will be discussed. The importance of gender and political experience cannot be confirmed by the results of this study, which is similar to the conclusions of Vivyan and Wagner (2015, 2016). Regarding gender, no general preference for female candidates was found (e.g., Schwarz et al., 2018). Evidently, as McElroy and Marsh (2010) state, the role of candidate gender differs significantly per context, and per election. Also, voters did not use political experience to base their vote choice on. This refutes results of Kirkland and Coppock (2018), who found political experience to have the strongest effect of all measured candidate attributes, and other studies where political experience was found to play a role (Carlson, 2015; Carnes & Lupu, 2016). The candidate attributes party, nationality, leadership characteristics, and vision do play an important role in the decision-making process of European voters contributing to the candidate attribute literature in general, and to the EU literature specifically. The candidate attribute party was included to perform robustness checks (see Appendix J), nonetheless, the results indicate that voters use a candidate’s party affiliation to base their vote choice on13. One of the contributions of this paper was to investigate vote choice in the EU context, and thereby being able to include a candidate attribute that has been hardly researched before (i.e., nationality). The results suggest an important role of candidate nationality, substantiating the congruency model of political preference by Caprara and Zimbardo (2004): Voters matched themselves to political
candidates with regard to their nationality (i.e., being Dutch). Opposite to what was expected regarding the leadership characteristics, voters preferred an integer (forced vote choice as dependent variable), or a politically crafted (evaluation of support as dependent variable)
candidate over a candidate with good communicative performances, meaning that voters prefer leaders who put general interests before their own, or leaders who are politically skilled. A potential reason could be that voters already expect politicians knowing how to handle the media (i.e., communicative performances) because they are more than used to seeing or hearing about them via the media (e.g., Strömbäck, 2008). Testing if voters use candidate gender to stereotype leadership characteristics (McElroy & Marsh, 2010) did not confirm expectations which were based on the study of Aaldering and Van Der Pas (2018) who found that found that Dutch party leaders were more often discussed in terms of those leadership characteristics in the media than Dutch female party leaders. Thus, suggesting that voters do not stereotype candidates the same way as the media does. Last, candidate vision, and thus candidate communication was important for voters14. Therefore, future studies investigating candidate attributes should include different measures of candidate
communication to explore candidate communication in vote choice even further.
Second, the role of voter characteristics will be discussed. Voters’ propensity to vote for liberal, left-leaning parties was not able to explain voting for a female candidate, which contradicts claims made by Huddy and Terkildsen (1993), and McDermott (1997). Possibly these claims are outdated. Different ideological distinctions could be researched in the future, but robustness checks for all political parties in the Dutch system do not point into such a direction (see Appendix E for robustness checks). Future research could focus on different voter characteristics (e.g., political knowledge, political efficacy) to see if there are other voter characteristics moderating vote choice for specific candidate attributes. Regarding the
dimensions of attitude towards the EU, performance could not explain vote choice for a candidate’s political experience, and vote choice for the candidate vision carrying on, thus, it
did not matter how positive voters were about how well the EU functions, and how well European institutions perform. Likewise, utilitarianism was not able to explain vote choice for
the candidate visions nothing but the single market, and those who want more do more. Specifically, voters’ thoughts about if they and their country benefit from being a member of the EU does not moderate voting for these visions. The results regarding attitude towards the EU’s strengthening are mixed. On the one hand, strengthening (i.e., how positive voters are
towards European integration), did not explain vote choice for a candidate’s nationality. On the other hand, it did explain vote choice for the candidate vision nothing but the single market negatively, and vote choice for the candidate vision doing much more together
positively. Robustness checks show that attitude towards the EU’s strengthening explains vote choice for all candidate visions, except for carrying on. Furthermore, these checks show that performance and utilitarianism were able to explain voting for some of the candidate visions (see Appendix H for robustness checks). In the future, the other dimensions of attitude towards the EU, negative affection and identity (Boomgaarden et al., 2011; de Vreese et al., 2018), should be researched to see if these are able to explain vote choice for a candidate’s vision as well.
Turning to the second research question, no moderation effects of voting for a candidate’s leadership characteristics on voting for a candidate’s vision were found. Thus,
claims about extreme, clear versus neutral, unclear political positions cannot be validated (Mondak & Huckfeldt, 2006). An explanation could be that this study only investigated the moderating effect of leadership characteristics on candidate communication. Therefore, future research could focus on different candidate characteristics (e.g., party, nationality) to see if these will be able to explain vote choice for candidate communication.
A first limitation of this study concerns the method of this study. Respondents were presented with two clear hypothetical profiles of candidates, and they had to choose between the two. During EP campaigns, voters are not presented with such clear overviews of the candidates and their candidate attributes, and voters do not have to make such a definite
choice because there are more than two candidates on the list (Arnesen et al., 2018).
Subsequently, the Spitzenkandidaten are not on the ballot. This concerns the external validity of the study because it not fully represents the real-life voting decision voters have to make. However, as Hainmueller et al. (2014, p. 19) argue, the forced vote choice respondents were asked to make can be compared in some sense to the trade-offs voters have to make in the ballot booth. Also, the method allows to measure AMCEs for six different candidate
attributes, and thereby this study is able to offer great insight about which roles the different candidate attributes play in vote choice. Another strength of this study is the internal validity. The power of this study is high because the 2,736 observations far exceed the number of possible profiles (i.e., 800), meaning all possible candidate profiles were measured. Also, two significantly different samples in terms of gender, age, education, and residence were studied: a student sample and a snowball sample (see Appendix C). The snowball sample being much more representative of Dutch voters. With most results holding across the samples, the findings of this study are robust (see Appendix K for robustness checks for the samples).
The results of this study hold across two different samples; therefore, the findings are robust, but the question remains if these findings are representative for all European voters. Notwithstanding, the results of this study have important implications for understanding vote choice in non-national elections, which can also be used by European political groups and politicians. In conclusion, Dutch voters have a preference regarding four of the six candidate attributes studied here: Voters use the candidate attributes party, nationality, leadership characteristics, and vision to base their vote choice on.
References
Aaker, J. L., & Lee, A. Y. (2006). Understanding regulatory fit. Journal of Marketing
Research, 43(1), 15–19. https://doi.org/10.2307/3150973
Aaldering, L., & van der Pas, D. J. (2018). Political leadership in the media: Gender bias in leader stereotypes during campaign and routine times. British Journal of Political
Science, (2017), 1–21. https://doi.org/10.1017/S0007123417000795
Aaldering, L., & Vliegenthart, R. (2016). Political leaders and the media. Can we measure political leadership images in newspapers using computer-assisted content analysis?
Quality and Quantity, 50(5), 1871–1905. https://doi.org/10.1007/s11135-015-0242-9
Arnesen, S., Duell, D., & Johannesson, M. P. (2018). Do citizens make inferences from political candidate characteristics when aiming for substantive representation? Electoral
Studies. https://doi.org/10.1016/J.ELECTSTUD.2018.10.005
Black, J. H., & Erickson, L. (2003). Women candidates and voter bias: Do women politicians need to be better? Electoral Studies, 22(1), 81–100. https://doi.org/10.1016/S0261-3794(01)00028-2
Boomgaarden, H. G., Schuck, A. R. T., Elenbaas, M., & de Vreese, C. H. (2011). Mapping EU attitudes: Conceptual and empirical dimensions of Euroscepticism and EU support.
European Union Politics, 12(2), 241–266. https://doi.org/10.1177/1465116510395411
Borisyuk, G., Rallings, C., & Thrasher, M. (2007). Women in English local government, 1973–2003: Getting selected, getting elected. Contemporary Politics, 13(2), 181–199. https://doi.org/10.1080/13569770701562674
Buttice, M. K., & Stone, W. J. (2012). Candidates matter: Policy and quality differences in congressional elections. Journal of Politics, 74(3), 870–887.
https://doi.org/10.1017/S0022381612000394
political preference. American Psychologist, 59(7), 581–594. https://doi.org/10.1037/0003-066X.59.7.581
Carlson, E. (2015). Ethnic voting and accountability in Africa: A choice experiment in Uganda. World Politics, 67(2), 353–385. https://doi.org/10.1017/S0043887115000015 Carnes, N., & Lupu, N. (2016). Do voters dislike working-class candidates? Voter biases and
the descriptive underrepresentation of the working class. American Political Science
Review, 110(4), 832–844. https://doi.org/10.1017/S0003055416000551
Cesario, J., Grant, H., & Higgins, E. T. (2004). Regulatory fit and persuasion: Transfer from “feeling right.” Journal of Personality and Social Psychology, 86(3), 388–404.
https://doi.org/10.1037/0022-3514.86.3.388
Clark, M., & Leiter, D. (2014). Does the ideological dispersion of parties mediate the electoral impact of Valence? A cross-national study of party support in nine western European democracies. Comparative Political Studies, 47(2), 171–202.
https://doi.org/10.1177/0010414013488537
Dalton, R. J., McAllistar, I., & Wattenberg, M. P. (2000). The consequences of partisan dealignment. In Parties without partisans: Political change in advanced industrial
democracies (pp. 37–63). Oxford: Oxford University Press.
de Vreese, C. H., Azrout, R., & Boomgaarden, H. G. (2018). One size fits all? Testing the dimensional structure of EU attitudes in 21 countries. Institutional Journal of Public
Opinion Research, 1–25. https://doi.org/10.1093/ijpor/edy003/4902938
Dolan, K. (1998). Voting for women in the “year of the woman.” Midwest Political Science
Association, 42(1), 272–293.
Franchino, F., & Zucchini, F. (2015). Voting in a multi-dimensional space: A conjoint analysis employing valence and ideology attributes of candidates. Political Science
Gattermann, K., de Vreese, C., & van der Brug, W. (2016). Evaluations of the
Spitzenkandidaten: The role of information and news exposure in citizens’ preference formation. Politics and Governance, 4(1), 37. https://doi.org/10.17645/pag.v4i1.460 Gattermann, K., & Vasilopoulou, S. (2015). Absent yet popular? Explaining news visibility of
Members of the European Parliament. European Journal of Political Research, 54(1), 121–140. https://doi.org/10.1111/1475-6765.12071
Green, J., & Hobolt, S. B. (2008). Owning the issue agenda: Party strategies and vote choices in British elections. Electoral Studies, 27(3), 460–476.
https://doi.org/10.1016/j.electstud.2008.02.003
Gunther, R., & Mughan, A. (Eds. . (2000). Democracy and the media: a comparative
perspective. Cambridge University Press.
Hainmueller, J., Hopkins, D. J., & Yamamoto, T. (2014). Causal inference in conjoint analysis: Understanding multidimensional choices via stated preference experiments.
Oxford Journals, 22(1), 1–30.
Hallin, D., & Mancini, P. (2004). Comparing Media Systems. Three Models of Media and Politics.
Hayes, D. (2009). Has television personalized voting behavior? Political Behavior, 31(2), 231–260. https://doi.org/10.1007/s11109-008-9070-0
Huddy, L., & Terkildsen, N. (1993). The consequences of gender stereotypes for women candidates at different levels and types of office. Political Research Quarterly, 46(3), 503–525. https://doi.org/10.2307/448945
Jewell, M. E., & Breaux, D. (1988). The effect of incumbency on State Legislative Elections.
Legislative Studies Quarterly, 13(4), 495–514.
Juncker, J.-C. (2017). White Paper on the future of Europe. European Commission. https://doi.org/10.2775/32364
Kepplinger, H. M. (2002). Mediatization of politics: Theory and data. Journal of
Communication, 52(4), 972–986.
Kiesraad. (n.d.). Verkiezingen. Retrieved from
https://www.verkiezingsuitslagen.nl/verkiezingen/index/1
Kim, H., Rao, A. R., & Lee, A. Y. (2008). It’s time to vote: The effect of matching message
orientation and temporal frame on political persuasion. Journal of Consumer Research,
35(6), 877–889. https://doi.org/10.1086/593700
Kirkland, P. A., & Coppock, A. (2018). Candidate choice without party labels: New insights from conjoint survey experiments. Political Behavior, 40(3), 571–591.
https://doi.org/10.1007/s11109-017-9414-8
Knudsen, E., & Johannesson, M. P. (2018). Beyond the limits of survey experiments: How conjoint designs advance causal inference in political communication research. Political
Communication, 00, 1–13. https://doi.org/10.1080/10584609.2018.1493009
Lee, D. S. (2008). Randomized experiments from non-random selection in U.S. House elections. Journal of Econometrics, 142, 675–697.
https://doi.org/10.1016/j.jeconom.2007.05.004
Mazzoleni, G., & Schulz, W. (1999). “Mediatization” of politics: A challenge for democracy?
Political Communication, 16(3), 247–261. https://doi.org/10.1080/105846099198613
McDermott, M. L. (1997). Voting cues in low-information elections: Candidate gender as a social information variable in contemporary United States Elections. American Journal
of Political Science, 41(1), 270–283.
McElroy, G., & Marsh, M. (2010). Candidate gender and voter choice: Analysis from a multimember preferential voting system. Political Research Quarterly, 63(4), 822–833. https://doi.org/10.1177/1065912909336270
electoral decision making. Electoral Studies, 25(1), 20–34. https://doi.org/10.1016/j.electstud.2005.02.006
Pardos-Prado, S. (2012). Valence beyond consensus: Party competence and policy dispersion from a comparative perspective. Electoral Studies, 31(2), 342–352.
https://doi.org/10.1016/j.electstud.2012.01.004
Popkin, S. L., & Popkin, S. L. (1994). The reasoning voter: Communication and persuasion
in presidential campaigns. University of Chicago Press.
Reif, K., & Schmitt, H. (1980). Nine second-order National Elections - A conceptual
framework for the analysis of European Election results. European Journal of Political
Research, 8, 3–44.
Sanbonmatsu, K. (2002). Gender stereotypes and vote choice. Midwest Political Science
Association, 46(1), 20–34.
Sapiro, V., & Conover, P. J. (1997). The variable gender basis of electoral politics: Gender and context in the 1992 US election. British Journal of Political Science, 27(4), 497– 523. https://doi.org/10.1017/S0007123497000240
Schmitt, H., Hobolt, S., & Popa, S. A. (2015). Does personalization increase turnout?
Spitzenkandidaten in the 2014 European Parliament elections. European Union Politics,
16(3), 347–368. https://doi.org/10.1177/1465116515584626
Schoenbach, K., De Ridder, J., & Lauf, E. (2001). Politicians on TV news: Getting attention in Dutch and German election campaigns. European Journal of Political Research,
39(4), 519–531. Retrieved from
https://onlinelibrary.wiley.com/doi/abs/10.1111/1475-6765.00586
Schwarz, S., Hunt, W., & Coppock, A. (2018). What have we learned about gender from candidate choice experiments? A meta-analysis of 30 factorial survey experiments, 1–21. Retrieved from http://alexandercoppock.com/papers/SHC_gender.pdf
Sellers, P. J., & Schaffner, B. F. (2007). Winning coverage in the U.S. Senate. Political
Communication, 24(4), 377–391. https://doi.org/10.1021/acs.nanolett.6b01595
Smith, E. R., & Fox, R. L. (2001). The electoral fortunes of women candidates for Congress.
Political Research Quarterly, 54(1), 205–221.
https://doi.org/10.1177/106591290105400111
Stieglitz, S., & Dang-Xuan, L. (2013). Social media and political communication: A social media analytics framework. Social Network Analysis and Mining, 3(4), 1277–1291. https://doi.org/10.1007/s13278-012-0079-3
Strezhnev, A., Hainmueller, J., Hopkins, D. J., & Yamamoto, T. (2014). Conjoint Survey
Design Tool: Software manual.
Strömbäck, J. (2008). Four phases of mediatization: An analysis of the mediatization of politics. The International Journal of Press/Politics, 13(3), 228–246.
https://doi.org/10.1177/1940161208319097
van Aelst, P., Sheafer, T., & Stanyer, J. (2012). The personalization of mediated political communication: A review of concepts, operationalizations and key findings. Journalism,
13(2), 203–220. https://doi.org/10.1177/1464884911427802
van der Brug, W. (2010). Structural and ideological voting in age cohorts. West European
Politics, 33(3), 586–607. https://doi.org/10.1080/01402381003654593
Vivyan, N., & Wagner, M. (2015). What do voters want from their local MP? Political
Quarterly, 86(1), 33–40.
Vivyan, N., & Wagner, M. (2016). House or home? Constituent preferences over legislator effort allocation. European Journal of Political Research, 55(1), 81–99.
Welch, S., & Studlar, D. T. (1986). British public opinion toward women in politics: A comparative perspective. Western Political Science Association, 39(1), 138–154.
Notes
1
Voter turnout for the Dutch general election was 81.57% in 2017 and 74.57% in 2012, and since 1917 the turnout has not been below 73.35% (1998) (Kiesraad, n.d.).
2
The other two dimensions of attitude towards the EU, negative affection and identity, are not expected to be related to the candidate attributes measured in this study.
3
Forty-two responses had to be deleted because they did not finish the questionnaire, and one response was deleted because the response time exceeded 22 hours.
4 Personal communication with one of the authors (Loes Aaldering) took place, who advised
to use the following words: inspiring and enthusiastic for communicative performances, predictable and consistent for consistency, honest and sincere for integrity, expert and smart for political craftsmanship, and decisive and confident for vigorousness.
5
Dutch parties in the political system are: CDA, CU, D66, DENK, FVD, GL, PvdA, PvdD, PVV, SGP, SP, VVD, and 50+.
6
DENK (M = 1.49, SD = 2.01), D66 (M = 5.94, SD = 2.76), GroenLinks (M = 5.91, SD = 3.22), PvdA (M = 4.56, SD = 2.63), SP (M = 3.14, SD = 2.60), VVD (M = 4.63, SD = 3.02), measured on a scale of 0 (totally unlikely) to 10 (totally likely) how likely it would be that participants would ever vote for that party.
7
The four biggest and most established European political groups were chosen, namely the Alliance of Liberals and Democrats for Europe (ALDE: D66 and VVD), the European People’s Party (EPP: CDA), the Progressive Alliance of Socialists and Democrats (S&D:
PvdA), and the Greens/European Free Alliance (Greens-EFA: GroenLinks). Other European political groups (e.g., more extreme groups) were not included due to time and space
constraints, and due to the fact that this would limit the design because candidates from extreme European political groups will not propagate certain visions for the future of the EU. The link between the European political groups and the national political parties was
presented because research has found that most Dutch voters do not have a lot of knowledge about the European political groups (Gattermann et al., 2016).
8 Item 6, from the dimension utilitarianism, has a factor loading of only .39. Also, if item 9
would be deleted from the dimension strengthening, Cronbach’s alpha would increase to .71, but item 9 is also the item that loads best (.82) on the dimension strengthening (see Appendix D). Therefore, it was chosen to keep both items, and to follow the work of Boomgaarden et al. (2011) and de Vreese et al. (2018).
9
Model 2 (evaluation of support as dependent variable) shows a different result, a negative significant effect of propensity to vote for D66 on voting for a female Spitzenkandidat: Voters with a higher propensity to vote for D66 are significantly less likely to vote for a female
Spitzenkandidat compared to a male Spitzenkandidat (see Appendix E). Still, H1b was not
supported. 10
Model 2 (evaluation of support as dependent variable) shows a different result, a significant positive effect of the leadership characteristic political craftsmanship: Voters are significantly more likely to vote for a Spitzenkandidat with the leadership characteristic political
craftsmanship compared to a Spitzenkandidat with the leadership characteristic communicative performances (see Table 2).
11
All moderation effects of voters’ attitude towards the EU’s performance, utilitarianism, and strengthening on vote choice for a candidate’s vision were checked (see Table H1 in
Appendix H). 12
Model 2 (evaluation of support as dependent variable) shows a different result, no significant negative effect of attitude towards the EU’s strengthening on voting for a
Spitzenkandidat whose vision is nothing but the single market: Voters with a more negative
attitude towards the EU’s strengthening are not significantly more likely to vote for a
13
The results show that Dutch voters were significantly less likely to vote for EPP and S&D compared to voting for ALDE (see Table 1). Only the result for the EPP was confirmed by model 2 (evaluation of support).
14
Voters preferred candidates with the vision nothing but the single market less, and preferred candidates with the vision those who want more do more than candidates with the vision carrying on (see Table 2). Model 2 (evaluation of support as dependent variable) shows a different result, no significant positive effect of the vision those who want more do more: Voters were not significantly more likely to vote for a Spitzenkandidat whose vision is those who want more do more (see Table 2). Thus, conclusions about this particular vision should be interpreted with care.
Appendix A
The five scenarios: policy overview by Juncker (2017)
Carrying on Nothing but the
single market
Those who want more do more
Doing less more efficiently Doing much more together Single market & trade Single market is strengthened, including in the energy and digital sectors; the EU27 pursues progressive trade agreements Single market for goods and capital strengthened; standards continue to differ; free movement of people and services not fully guaranteed As in “Carrying on”, single market is strengthened and the EU27 pursues progressive trade agreements Common standards set to a minimum but enforcement is strengthened in areas regulated at EU level; trade exclusively dealt with at EU level Single market strengthened through harmonization of stronger enforcement; trade exclusively dealt with at EU level Economic & Monetary Union Incremental progress on improving the functioning of the euro area
Cooperation in the euro area is limited
As in “Carrying on” except for a group of countries who deepen cooperation in areas such as taxation and social standards
Several steps are taken to
consolidate the euro area and ensure its stability; the EU27 does less in some parts of employment and social policy Economic, financial and fiscal Union is achieved as envisioned in the report of the Five Presidents of June 2015 Schengen, migration & security Cooperation in the management of external borders stepped up gradually; progress towards a common asylum system; improved coordination on security matters No single migration or asylum policy; further coordination on security dealt with bilaterally; internal border controls are more systematic As in “Carrying on” except for a group of countries who deepen cooperation on security and justice matters Cooperation on border management, asylum policies and counter-terrorism matters are systematic As in “Doing less more efficiently”, cooperation on border management, asylum policies and counter-terrorism matters are systematic Foreign policy & defense Progress is made on speaking with one voice on foreign affairs; closer defense cooperation Some foreign policy issues are increasingly dealt with bilaterally; defense cooperation remains as it is today As in “Carrying on” except for a group of countries who deepen cooperation on defense, focusing on military coordination and joint equipment The EU speaks with one voice on all foreign policy issues; a European Defense Union is created As in “Doing less more efficiently”, the EU speaks with one voice on all foreign policy issues; a European Defense Union is created EU budget Partly modernized to reflect the reform agenda agreed at 27 Refocused to finance essential functions needed for the single market As in “Carrying on”; additional budgets are made available by some Significantly redesigned to fit the new priorities agreed at the level of the EU27
Significantly modernized and increased, backed up by own resources; a
Member States for the areas where they decide to do more
euro area fiscal stabilization function is operational Capacity to deliver Positive agenda for action yields concrete results; decision-making remains complex to grasp; capacity to deliver does not always match expectations Decision-making may be easier to understand but capacity to act collectively is limited; issue of common concern often need to be solved bilaterally As in “Carrying on”, a positive agenda for action at 27 yields results; some groups achieve more together in certain domains; decision-making becomes more complex Initial agreement on tasks to prioritize or give up is challenging; one in place, decision-making may be easier to understand; the EU acts quicker and more decisively where it has a greater role Decision-making is faster and enforcement is stronger across the board; questions of accountability arise for some who feel that the EU has taken too much power away from the Member States
Appendix B
Questionnaire T1: Factsheet and informed consent/Informatieblad Beste deelnemer,
Hierbij bent u uitgenodigd om deel te nemen aan een onderzoek dat wordt uitgevoerd onder de verantwoordelijkheid van onderzoeksinstituut ASCoR, onderdeel van de Universiteit van Amsterdam.
Voordat het onderzoek begint, is het belangrijk dat u op de hoogte bent van de procedure die in dit onderzoek wordt gevolgd. Leest u daarom onderstaande tekst zorgvuldig door.
Het doel van het onderzoek
Het doel van dit onderzoek is om te kijken naar hoe mensen politieke kandidaten evalueren. We zullen u een paar algemene vragen stellen en vragen over uw politieke voorkeur. Hierna krijgt u telkens twee hypothetische politieke kandidaten te zien en wordt u gevraagd hen te beoordelen.
De onderzoeksprocedure
Eerst zult u een paar algemene vragen beantwoorden. Daarna zult u een reeks stellingen te zien met betrekking tot uw attitude ten opzichte van de Europese Unie. Hierna zult u telkens twee hypothetische politieke kandidaten te zien krijgen en wordt u gevraagd tussen beide kandidaten te kiezen en beide kandidaten apart te beoordelen.
Duur
Het volledige onderzoek zou niet langer dan 10 minuten moeten duren. Vergoeding
U zult 0.5 research credit ontvangen voor uw deelname aan dit onderzoek. OR Er is geen vergoeding. Deelname aan dit onderzoek is geheel vrijwillig. Dat betekent dat er geen sanctie is als de vragenlijst niet wordt ingevuld.
Ongemak en risico’s
Deelname aan dit onderzoek brengt geen noemenswaardige risico’s of ongemakken met zich mee. Ook zult u niet met expliciet aanstootgevend materiaal worden geconfronteerd.
Anonimiteit van deelnemers
Uw antwoorden zijn vertrouwelijk en we verzamelen geen identificerende informatie zoals uw naam, ID-nummer, e-mailadres of telefoonnummer. We zullen uw persoonlijke gegevens ook niet delen met derden. Om uw gegevens te beschermen, zal de opgeslagen informatie geen informatie bevatten die u zou kunnen identificeren, en zullen alle gegevens collectief worden geanalyseerd.
Nadere inlichtingen
De projectleider van dit onderzoek is: Katjana Gattermann, telefoonnummer: 0615684525; e-mail: K.Gattermann@uva.nl. Voor meer informatie over het onderzoek en de uitnodiging om deel te nemen kunt u contact opnemen met Lotte Aalbers, e-mail:
lotte.aalbers@student.uva.nl.
Mochten er naar aanleiding van uw deelname aan dit onderzoek bij u toch klachten of opmerkingen zijn over het verloop van het onderzoek en de daarbij gevolgde procedure, dan kunt u contact opnemen met het lid van de Commissie Ethiek namens ASCoR, per adres: ASCoR secretariaat, Commissie Ethiek, Universiteit van Amsterdam, Nieuwe Achtergracht 166, 1018 WV te Amsterdam; 020-525 36980; ascor-secr-fmg@uva.nl. Een vertrouwelijke behandeling van uw klacht of opmerking is daarbij gewaarborgd.
Vrijwilligheid van deelname
Deelname aan dit onderzoek is vrijwillig. U kunt op ieder moment besluiten om af te zien van deelname aan dit onderzoek. U kunt binnen 7 dagen na het onderzoek verzoeken om uw onderzoeksgegevens te laten verwijderen.
Wij hopen u hiermee voldoende te hebben geïnformeerd en danken u bij voorbaat hartelijk voor uw deelname aan dit onderzoek dat voor ons van grote waarde is.
Let op: Als u op “ja” klikt hieronder, gaat u akkoord met het volgende: - Ik wil meedoen aan het onderzoek;
- Ik heb bovenstaande tekst gelezen en begrepen;
- Ik begrijp dat mijn antwoorden vertrouwelijk zijn en dat mijn anonimiteit gewaarborgd is;
- Ik begrijp dat mijn deelname volledig vrijwillig is en dat ik op elk moment van het onderzoek mag stoppen, zonder een reden op te geven;
- Ik ga ermee akkoord dat mijn gegevens geanonimiseerd worden gebruikt (u kunt tot 7 dagen na het onderzoek verzoeken om uw onderzoeksgegevens te laten verwijderen);
- Ik ben tenminste 18 jaar oud.
Q1: Wilt u deelnemen aan het onderzoek? (forced answer) Ja
Nee skip to end of survey
Hartelijk dank voor het meedoen aan dit onderzoek. Hieronder volgen eerst een aantal algemene vragen.
Wat is uw geslacht? (request response) 0 = Anders
1 = Man 2 = Vrouw
Wat is uw leeftijd? (request response) 1 = 18-24 jaar
