of Candidates’ Attractiveness on Candidate Evaluations and
Electoral success.
University of Amsterdam Graduate School of Communication Research Master of Communication Science
Thesis supervisor: Yphtach Lelkes Student: Moniek Irene Mewafi
Student number: 6051553 Date of completion: June 27th, 2014
I would like to thank the following amazing individuals:
My dear friend Evgenija Kravtsjenko,
for helping me making my video for recruiting respondents My fellow-student Violet Luif,
who sticked with me in the library in good times and in bad times My SPSS guru Ward van Zoonen
who answered all my simple question about statistics To lecturer Marcel van Egmond,
for helping me out with analysis I never heard of before and to my supervisor Yphtach Lelkes
who has encouraged my ambitions and stayed calm and friendly despite my impatience and frustrations throughout the process.
The Beauty of Politics?
A Critical Note to Contemporary Research into the Assumed
Effects of Candidates’ Attractiveness on Candidate Evaluations
and Electoral success.
Abstract
Candidates’ appearance is said to be one of the key determinants for electoral success. The idea that good-looking candidates are evaluated more positively and thus enjoy greater electoral success is a predominant assumption in recent literature. However, research about other tentative determinants remains scarce. As a response to that, this paper seeks to examine the influence of candidates’ attractiveness, alongside candidates’ competence, on candidate evaluations. The results of the experiment show that attractiveness does matter, leading to a more positive attitude towards the candidate, however the found effects are rather weak. On the contrary, candidates’ competence seems to be a more important determinant, and has a strong effect on perceived competence, perceived leadership and generates a more positive attitude towards the candidate. To further validate these findings into a real-world setting, a second study has been employed. In this second study, subjects rated candidates of three municipalities on their appearance. The rated attractiveness served as a predictor for electoral outcomes of municipal elections of Utrecht, Den Haag and Rotterdam. Candidates’ background information with personal and professional details were required by the means of a content analysis and served as control variables. The results show that there is no significant effect for candidate attractiveness, but candidates’ competence also reveals to be an important predictor for electoral success.
Introduction
A vivid example of the extent to which the appearance of a political candidate matters became painfully clear with the debate between John F Kennedy and Richard M Nixon in 1960. The good-looking and charming Kennedy gained the victory over the pale and unshaved Nixon. Research suggests that this television victory could be explained by Kennedy’s superior image (Druckman, 2003), since voters evaluate attractive political candidates more positively than their unattractive counterparts (Bailenson, Garland, Iyengar, & Yee, 2006; Chiao et al., 2008, Ottati & Deiger, 2002). Even though it is getting more and more commonly accepted that appearance is an important asset in politics (Olivola & Todorov, 2005; Todorov et al. 2008, Surawski & Osshoff; 2006), the variety of studies investigating the effect of attractiveness on a candidates’ success is rather limited and one-sided.
Can we really speak about a “beauty premium” in politics, such that good-looking political candidates have a higher chance of becoming elected? Some scholars argue that physical appearance of a political candidate indeed does matter and can even predict electoral outcome (Banducci et al., 2008, Chiao et al., 2008; Lawson et al., 2010; Olivola et al., 2012; Olivola & Todorov, 2010; Todorov et al., 2005). Even though these studies give us a great insight in the assumed interplay between attractiveness and electoral success, predicting election outcomes based solely on rated candidates’ pictures might not lead to a fully valid inference since voters are commonly not only exposed to candidates’ pictures alone.
Therefore, the first study aims to examine through the use of an experiment if candidate appearance still has an overpowering effect, even when substantive text about the political candidate is presented which is varying in the candidates’ suitability for the candidature. So far, there is no research devoted to the effect of attractiveness in situations
where both substantive and non-substantive elements are present. Based on the above, the following research question serves as a guideline in the first study:
RQ1: Does the appearance of a political candidate have an effect on a person’s evaluation, regardless of the actual competence of this politician?
The ultimate aim of this study is to elaborate knowledge on how voter evaluation of political candidates are developed and to what extend attractiveness matters if political competence (eg. competence for the job) is presented as well. This way, a better understanding of the development of candidate evaluations can be gained.
The outcomes from the first study have been required by means of an experiment, which could have implications for the generalizability of the results into a real life setting. Therefore, the second study aims to determine the possible effects of attractiveness, competence and gender in a multiparty system by performing a correlational study, while controlling for other important background factors. The following research-question will serve as a guideline in the second study:
RQ2: To what extent do the patterns found in study 1 apply to a real-life electoral study?
With the second study more insight will be in the extent the results of the experimental can be placed in perspective; the influence of candidates’ gender, attractiveness and competence on electoral success will be examined. The following chapter will deal with
the theoretical foundations of the situation mentioned above in detail, which will lead to the formulation of the hypotheses of both studies that will be tested further along this paper.
Theoretical Framework
Even though there is an increasing body of research devoted to the influence of physical appearance, attention for the influence of candidates’ physical appearance on their electoral success have only developed recently. This is quite surprising, since research of other disciplines have been devoted to this phenomenon for years. For example, in the discipline of persuasive communication, attention has been paid to the influence of the attractiveness of an endorser have a great impact on purchase intention and brand attitude (Amos et al. 2008; Till & Busler, 2000) and persuasion (Cacioppo, and Goldman 1981; Shafid et al., 1997). Also in the discipline of economics, several studies have been devoted to the benefits for good-looking employees in the labor market (Hamermesh & Biddle, 1994; Harper, 2000). Compared to these disciplines, research into the effects of candidates’ attractiveness on candidates’ evaluation and electoral success is rather new research scope.
This scarcity of research into attractiveness can be partly explained from out the dominant classical democratic theory; one might argue that the possible influence of physical attractiveness on electoral success is fairly ‘unwelcome’ (Olivola et al., 2012; Rosar et al, 2008). The classical democratic theory expects voters to be fully informed when selecting a political candidate and bringing out their vote for a politician (Banducci et al, 2008, Olivola et al., 2012). A democratic system requires careful and informed choices about the governing politicians in order to function properly, instead of choices being led by the appearance of political leaders. In short, the ideal perception of ‘well-informed voters’ which is still dominant in social and political science today, could be one of the reasons why the research
of the effects of non-political traits have been beyond the scope of the mainstream of social and political research (Banducci et al, 2008; Rosar et al., 2008).
However, the scope of the mainstream research is slowly changing, more and more studies are devoted to the effects of appearance on a candidates’ political success. The reason why this focus has shifted could probably be explained from the changes is the practice of both the political field and the developments in mass media.
Shifts within politics and mass media: personalization.
Recent research suggests that politics in liberal democracies has become increasingly personalized (McAllister, 2007; Kriesi, 2012). Even though the empirical evidence concerning the personalization of politics is scattered, there have most definitely been shifts in the practice of the political field (Dalton et al, 2000; Kriesi, 2012, Swanson & Manchini, 1996). Nowadays, it is not exceptional for a political party to put focus on candidates rather than on political platforms (Mughan, 2000, Berggren et al., 2006). There is a decline of the internal distinctiveness, diversity and extremity between political parties, whereby parties increasingly seek the political center to direct an extensive group with a range of diverse interests and perceptions (Brants & Praag, 2006). Additionally, the process of non-alignment in the party system, weakens the traditional adhesion of citizens to political parties (Dalton et al., 2000), and has changed the engagement with political parties and the overall political participation of today’s citizens (de Vreese, 2007; Henn, Weinstein, Wring, 2002, Livingstone, Bober, and Helsper 2005; Scheufele and Nisbet 2002). However, this changing focus cannot be accounted to the political field alone, since there is an increased focus of the mass media on party leaders’ and politicians’ personal evaluations as well (Kriesi, 2012).
The ongoing trend of personalized politics is re-enforced by the media, which have developed their own ‘media logic’ over time in the way they cover politics, and political campaigning and election periods in particular (Swanson and Mancini, 1996). Mass media have the power to select and shape the presentation of messages, and thereby influence the framing of public issues (Entman, 1993; Habermas, 2006). The higher degree of personalization of politics, just as values like negativity and drama, is an important news value pursued by the mass media in order to defeat their competition for a mass audience (Harcup & O’neill, 2001; McGregor, 2002). Especially the development of television and the rise of televised debates led to an increase of personal attention for the candidates, and these debates also drastically changed the relevance of the dimensions by which the candidates were judged (Reinemann and Wilke, 2007).
The so-called ‘personalization of politics’ brings two related phenomena forward. First, because of personalization there is a higher media attention for political candidates instead of a focus on their parties, institutions, or issues. Secondly, a shift in the criteria for the evaluation of political candidates seem to have shifted from political features such as a candidates’ professional competence and overall performance, to non-political features concerning personality traits and politicians’ physical attractiveness (Kriesi, 2012; Reinemann and Wilke, 2007).
Increased acknowledgement of influence non-political traits.
Several controlled experiments showed that attractive political candidates were evaluated more favorably than unattractive political candidates (Olivola & Todorov, 2010; Ottati & Deiger, 2002). Political candidates who “look the part”, overall received higher percentage of votes and therefore have a higher chance of becoming elected (Chiao et al.,
2008; Lawson et al., 2010; Todorov et al., 2005). Notwithstanding, these findings are in line with psychological studies that shows that individuals tend to judge unfamiliar people based on their appearance, and based on this impression inferring several personality traits such as intelligence, honesty, friendliness and trustworthiness (Hassin and Trope 2000; Zebrowitz et al. 2002). Political candidates who are ‘blessed’ with certain facial features, are believed to be more competent, sociable and dominant then their opponents (Antonakis & Dalgas, 2009; Ballew & Todorov, 2007, Olivola & Todorov, 2010). In addition, attractiveness is important asset for determining the degree to which a person is judged to be assertive and a leader (Mazur & Mueller, 1996). Individuals with a high perceived attractiveness prevail over those who do not possess this trait, and therefore seem more submissive person (Mazur & Mueller, 1996; Chiao et al., 2008; Little et al., 2007). Based on these findings, the following hypotheses are formulated:
H attractiveness 1: A higher level of candidates’ physical attractiveness will lead to a higher degree of candidates’ perceived competence.
H attractiveness 2: A higher level of candidates’ attractiveness will lead to a higher degree of perceived leadership.
H attractiveness 3: A higher level of candidates’ attractiveness will lead to a more positive attitude towards the candidate.
Prior research also indicated that the effect of attractiveness differed between male and female candidates: Features like compassion, people-oriented, gentle, kind, caring, approachable, sensitivity are mostly characteristics ascribed to female candidates, whereas male candidates are considered to be more dominant, belligerent, self-assured and assertive
(Chiao et al., 2008; Eagly and Karau, 2002; Lawless, 2004). Since ‘leader-like’ characteristics as described are mostly ascribed to male candidates, we expect to find differences across conditions with regard to candidates’ gender. The following hypotheses address these expectations:
Hgender1: Male candidates will have a higher score on perceived competence than female candidates.
Hgender2: Male candidates will score higher on perceived leadership than female candidates.
Although previous studies have attempted to examine the influence of candidates’ physical appearance on the perceptions of voters, there are most definitely some limitations to the studies performed so far. Previous studies either focused on candidate evaluations made in the ‘blink of an eye’ and their correlations with electoral outcomes (Ballew & Todorov, 2007; Todorov et al., 2005), on the effect of the placement of a candidates’ photograph on a ballot (Banducci et al., 2008) and the effects of candidate evaluations and the effect on hypothetical voting likelihood (Chiao et al., 2008). In these studies, the researchers show their participants two pictures of candidates, without any other information, and ask for whom they would hypothetically vote. With the lack of any other information, it seems quite logical that participants choose for the most attractive candidate. However, the tendency of selecting attractive candidates over unattractive candidates might be embedded in a different phenomenon which will be further explained in the next section.
The European case: call for alternative view?
Recent research of Curran et al. (2009) indicates that knowledge of public affairs is for a great part determined by the information provided by the mass media, and differs extensively per media system. Public service models of broadcasting overall devote a great amount of attention to public affairs, national and international news; and thereby provides its’ citizens of a higher level of knowledge of these particular areas (Curran et al., 2009). Commercial broadcasting systems on the other hand, mostly offer programs containing entertainment and do not sufficiently stimulate its’ audience to acquire knowledge about public and societal affairs (Bennett, 2003). By means of a content analysis of broadcasting content, together with measurements of public knowledge Curran et al. (2009) show that there are significant differences in public knowledge of inhabitants of USA, which has a commercial media-system, and four European countries with public broadcasting systems (eg. Denmark, Finland and Great Britain). The market driven television system of United States devotes only a low level of attention to national hard news and public affairs, and even less attention to news events occurring outside of America. This low level of hard and international news facts results in a high level of public ignorance among American citizens, especially compared to the citizens living in the European countries where public service broadcasting models apply (Curran et al., 2009). In addition, research of Lenz and Lawson (2011) shows that US citizens with a high level of television exposure tend to favor physically attractive candidates over less attractive candidates. Good-looking politicians therefore benefit disproportionately from television exposure, primarily among less knowledgeable individuals (Lenz & Lawson, 2011).
Since most research on the effect of attractiveness on candidate evaluation and electoral success comes from the United States, we expect that the assumption that candidates’
attractiveness has a great effect on electoral outcome most likely applies to the United States, but does not necessarily hold for the European context. More specifically, we expect that European voters, and in this particular case Dutch voters, do value the actual competence of a politician when evaluation his or her suitability as a political candidate or future leader, instead of basing their decisions on the physical appearance of a candidate. In this line of reasoning we formulated the following hypotheses about the effect of competence of a candidate:
Hcompetence1: A higher level of competence will lead to a higher degree of candidates’ perceived competence.
Hcompetence2: A higher level of candidates’ competence will lead to a higher degree of perceived leadership.
Hcompetence3: H1j: A higher level of competence will lead to a more positive attitude towards the candidate.
Although we expect that competence is an important determinant when it comes to perceived competence, perceived leadership and attitude towards the candidate, we do not expect that competence is the main predictor in this case. We cannot deny that attractiveness would most likely be a beneficial for a political candidate, and for this reason we do expect positive effects for the combination of competence and attractiveness on perceived competence, perceived leadership and attitude towards the candidate. The final expected interaction effects are hypothesized as following:
Hinteraction1: A higher level of attractiveness and competence will lead to a higher degree of perceived competence.
Hinteraction2: A higher level of attractiveness and competence will lead to a higher degree of perceived leadership.
H interaction3: A higher level of attractiveness and competence will lead to a more positive attitude toward the candidate.
In this study, attractiveness of a candidate will be depicted through a picture of a candidate, whereas the competence will be expressed by means of a political advertisement text alongside the candidates’ picture. Surprisingly, the combination of textual and visual messages and the combined effect on candidate evaluations have not been researched before. Nevertheless, it’s of great importance to further explore this relationship between visual, non-substantive information such as a candidates’ picture and non-substantive, textual information such as the vision and work-related background of the politician, since most information reaching the electorate is a combination of both elements.
Method Study 1
The first study of is aimed at investigating the relationship between text and visual representation of political candidates. A within-subject 2 (male/female) x 2 (attractive/unattractive) x 2 competence (low competence/ high competence) experimental design was performed. 209 university students participated in this study, in which female students were highly overrepresented (71,3%). All participants received Research credits when the experiment was completed. Moreover, the study was approved by the ethics committee of the Amsterdam School of Communication Research.
Independent variables.
Political trust. Political trust refers to the amount of trust the participants has in public authorities such as the government, local authorities, the court, the police and political parties. The trust was assessed by these 5 items (e.g. “To what extent do you think the local authorities are to be trusted?) Answer categories ranged from 1 very untrustworthy to 7 very trustworthy. The items measuring political trust were found to be very reliable (a=0.861), and were hence combined in a variable Political trust.
Dependent variables.
Attractiveness. Attractiveness was measured by a single item question, in which participants had to answer to what extend they agreed with the statement (eg. “The political candidate is attractive”). Attractiveness was measured on a 7 points Likert scale ranging from 1 strongly disagree to 7 strongly agree.
Perceived Competence. Perceived competence was a single item statement as well, which participants had to answer to what extent they agreed with the statement (eg. “The political candidate is competent”). Perceived competence was measured on a 7 points Likert scale ranging from 1 strongly disagree to 7 strongly agree.
Perceived leadership. Perceived leadership was measured just the same as the previous items, with the statement: (eg. “The political candidate is a good leader”).
Attitude towards candidate. Participants were asked to give their general impression of the candidate. The attitude towards the candidate was measured by one single item (eg. ‘What is your general impression of the political candidate?’), to which they could answer on a 5 points Likert scale ranging from 1 very negative to 5 very positive.
Stimuli.
The stimuli used in this experiment were designed for the purpose of this study. The study consisted of eight conditions1, all conditions were similar in their set-up; a political website was imitated. The party name and logo were unrecognizable. On the website a small text about the party vision was described, which was equal in all conditions. In a small paragraph underneath the general party vision some personal information about the political candidate was present. This personal information was manipulated across conditions: either the text presented the politicians as being highly competent by means of education or prior working experience in the political field, or the text described a low competent candidate by means of a low, not related educational background and insufficient working experience. For example, candidates in the low competency condition enjoyed their education in a not related discipline and had jobs on a high school or in engineering. Candidates in the high competency condition enjoyed university education in fields such as international relations or political science and had previous jobs at one of the public authorities.
Next to this personal text, a picture of a candidate was displayed, either a male or female, attractive or unattractive. The pictures were retrieved through the dataset of the real-life study, which is described in detail in ‘Method study 2’. Candidates that scored higher on the attractiveness scale (M=5.20, sd.=1.022) in the real-life study, where selected for the use in attractive conditions of the experiment. Of the candidates that scored low on the attractiveness scale (M=2.84, sd.= 1.394), 4 unattractive candidates were selected for the ‘unattractive conditions’. More details about the candidate pictures and the Attractiveness
1 The set-up of the study contained eight conditions. Due to a problem with the data-gathering in Qualtrics, the
sixth condition containing an unattractive, not competent, female candidate was not shown to participants and is therefore not included in the analysis of this experiment.
Consensus Method are described in detail in ‘Method section study two’. For an example of the stimulus material please consult Appendix 1.
Task procedure.
All candidates were asked to answer question about their social demographic background. Next, the participants were randomly assigned to one of the eight conditions. After the display of the manipulated political website of one of the candidates, respondents were asked to answer to questions described in the measurement section. After finishing the first questions, participants were randomly assigned to another condition. The possibility of getting the same text or picture was ruled out, because of the randomizer tool in Qualtrics.
Manipulation check.
Attractiveness. Participants in the condition with an attractive candidate rated the politicians significantly higher (M=4,72, sd.=1.26) than the participants whom did not receive an advertisement with an attractive candidate (M=2,84, sd.=1,39), t(204) = 9.985, p<0.01, and therefore we can conclude that the manipulation on the base of attractiveness succeeded.
Competence. All participants whom received the stimulus with a competent candidate answered significantly higher to the competence statement (M=5.10, sd.=1.12) , than the participants who received an advertisement with an incompetent candidate (M=3,49, sd.=1.34), t(205)=9.242, p<0.01. The manipulations that were designed to depict a competent candidate were perceived as being more competent, and therefore this manipulation has been successful as well.
Testing the Hypotheses
Since the experiment was a 2 x 2 x 2 design, in which all participants saw 2 of the 8 conditions, the dataset is hierarchically structured. Therefore, a multi-level model regression analysis is performed (Field, 2010; Hox, 2002). Since the focus in this experiment lies on the effect of attractiveness, gender and competence on perceived competence, perceived leadership and attitude towards the candidate, the results will be reported per dependent variable.
Effects of gender, attractiveness and competence on perceived competence.
The first hypotheses of gender, attractiveness and competence were aimed at testing the relationship between attractiveness and competence. Hattractiveness1 postulated that a higher degree of physical attractiveness has a positive effect on perceived competence. The results show different: Attractiveness does not yield a significant main effect on perceived competence (b=0.137, rob sd.= 0.114, n.s.). Attractive candidates were not perceived as being more competent than unattractive candidates. Therefore, Hattractiveness1 is rejected.
The second hypothesis, Hgender1, was directed at the effect of gender on perceived competence. The multilevel regression analysis showed that gender has a not-significant negative effect on competence (b=-0.086, sd.=0.109, p<0.01); which indicates that male candidates overall scored lower on the perceived competence statement, however not significantly lower. Therefore, Hgender1 is rejected.
The third hypothesis, Hcompetence1 was directed at the effect of competence on perceived competence; in which competence logically yield a significant, strong effect on perceived competence (b=1.646, sd.=0.119, p<0.01). Hcompetence1 is therefore confirmed.
The fourth hypothesis, Hinteraction1, was aimed at the interaction effect of gender and attractiveness, which postulated that male candidates with a higher level of physical attractiveness will have a higher score on perceived competence than female candidates. A positive, significant interaction effect between gender and attractiveness was found (more details see Model 2 in table 1), which entails attractiveness does have a significant effect on perceived competence, but that this effect is stronger for male candidates (b=0.501, sd. =0.253, p<0.05).
[TABLE 1 ABOUT HERE]
Although there were only four hypotheses aimed at testing the relationship between attractiveness and perceived competence, more significant effects were found in the multi-level regression analysis. When controlling for the age and gender of the participants (for more details please consult Model 3 in table 1), we find that the older respondents were, the higher they rated the candidates to be competent (b=0.024, sd. =0.008, p<0.05). Even though the regression is significant, the effect is quite small. We also found that female respondents tend to perceive candidates as more competent (b=0.361, sd. =160, p<0.05) than the male participants did. In addition, a significant, negative main effect of candidates’ gender on perceived competence; which entails that male candidates are scored -0.804 lower on the perceived competence statement (b=0.804 sd.=0.160, p<0.05). In addition, there is a marginally significant effect of attractiveness, which means that attractive candidates scored slightly higher on the perceived competence statement.
Effects of gender, attractiveness and competence on perceived leadership.
The following four hypotheses are aimed at investigating the relationship between attractiveness, competence and perceived leadership (for more details please consult Table 2). Attractiveness yield a significant, positive main effect on perceived leadership (b=0.240, rob sd. =0.122, p<0.05), for this reason Hattractiveness2 is confirmed. The second hypothesis, Hgender2, was aimed at examining the relationship between gender and perceived competence, however there is no significant effect found (b=-0.001, sd.=0.121, n.s.).
Competence also has a strong, positive effect on perceived leadership (b=0.848, rob sd.=0.128, p<0.01). Therefore, Hcompetence2 is confirmed:. Hinteraction2 was aimed at the interaction effect of attractiveness and competence on perceived leadership, this hypothesis was rejected; the interaction between attractiveness and competence does not yield a significant effect on perceived leadership (b=-0.435, b=0.288, ns).
[ TABLE 2 ABOUT HERE ]
The interaction effect between gender and competence does yield a significant effect on perceived competence (b=0.231, sd.=0.269), which entails that competent male candidates score 0.231 point higher on perceived competence than competent female candidates do (for more details consult Model 2 of Table 2). In short, attractiveness does have a significant effect on perceived leadership, but that this effect is stronger for male candidates (b=0.501, sd. =0253, p<0.05).
However, when adding the control conditions to the multilevel model (Model 3, Table 2), the interaction effect between gender and competence on perceived leadership does not hold any longer. We do find a marginally significant effect of participants gender on
perceived leadership, so female participants tend to rate candidates 0.282 higher than male participants do (b=0282, sd.=0.159, p<0.10). Also, the participants with a higher trust in political institutions tend to rate the candidates higher on the perceived leadership statement (b=0182, sd.=0.072, p<0.05).
Effects of gender, attractiveness and competence on attitude towards the candidate.
The final hypotheses were aimed at the influence of attractiveness and competence on overall attitude towards the candidate (For table, see Table 3). Hattractiveness3 formulated that a higher level of physical attractiveness has a positive effect on candidates’ evaluation. Attractiveness indeed yields a significant, positive effect on overall candidate evaluation (b=0.492, rob sd.= 0.066, p<0.01). Therefore, Hattractiveness3 is confirmed.
[ TABLE 3 ABOUT HERE ]
Hcompetence3 was aimed at the main effect of competence on candidate evaluation. Competence has a significant positive effect on candidate evaluation (b=0.323, rob sd.=0.073, p<0.01). The candidates presented in a high competence condition were evaluated more positively than the candidates in a low competence condition; Hcompetence3 is confirmed. Lastly, we expected that attractive and competent candidates would be evaluated more positively then unattractive candidates, however there is no significant interaction effect for attractiveness and competence; Hinteraction3 is therefore rejected.
When controlling for age, gender and political trust of the participants, we find significant effects for participants gender (b=0.286, rob sd.=0.101, p<0.05), for participants political trust (b=0.149, rob sd.=0.046, p<0.05) and for the interaction between gender of
candidate and attractiveness (b=0.318, rob sd.=0.145, p<0.05). Female participants tend to evaluate candidates more positive then male participants. The participants with a higher degree of trust in political institutions tend to be more positive towards the candidates as well. Attractive, male politicians seem to benefit most in this case, and get rated more positive than unattractive candidates.
[ TABLE 4 ABOUT HERE]
Discussion Study 1
The aim of this experiment was to gain more insight in the working of non-substantive, visual message (attractiveness) and a substantive, textual message (competence) on the evaluation of political candidates. The results show that attractive leaders score overall higher on perceived competence, which indicated that attractiveness correlates with the extent to subjects judge a politician to have leading skills. Attractiveness also has a positive effect on the general attitude towards the candidate; attractive candidates get evaluated more positively than unattractive candidates.
In addition, we found one interaction effect; attractive male candidate are seen as more competent than attractive, competent female candidates. Even though attractiveness yields significant effects on the dependent variables, the significant effects of attractiveness can be put in perspective quite easily when looking at the overall means of each condition on the dependent variables; overall the differences in the scores on perceived competence, perceived leadership and attitude towards the candidate did not differentiate extremely.
So even though some results indicate that candidate attractiveness is an influential asset, a candidates’ competence seems even to play a bigger role; candidates’ competence has
strong significant effect on perceived competence, perceived leadership and candidate evaluation. The effects of candidates’ competence on perceived competence and perceived leadership were overall substantially stronger than candidates’ attractiveness.
Placing the results in context.
We gained more insight in the relative importance of attractiveness and competence with regard to perceived competence, perceived leadership and participants’ attitude towards political candidates. However, one might argue that a positive score for perceived competency, perceived leadership and attitude towards the candidate in experimental settings does not necessarily hold for real-world circumstances. Therefore, the determinants used in the experiment will be tested in a real life setting as well, to further comprehend the effects of these factors outside of an experimental condition and compare these results to earlier published research, which mostly comes from the United States, that both has a different political and media system.
Method Study 1
On March 19th, there were local elections in all 403 Dutch municipal districts. In this study the elections of the six biggest political parties of the municipalities of Utrecht, Rotterdam and The Hague were analyzed. Only candidates of the six big political parties were examined, because the websites of these parties are better developed then the local parties’ websites; the pictures used on these websites were overall of high quality and suitable for standardizing and rescaling for the purpose of this study. Of the three municipalities, a total of 657 candidates of CDA, PvdA, D66, VVD, CU/SGP, and GroenLinks were registered on the electoral roll. Of those 657 candidates 173 candidates were actually analyzed. The decision to
only analyze 173 candidates was mainly because political parties only presented pictures on their websites of the first 15-20 political candidates of the list, and it was hard to require suitable pictures of the remaining candidates on the list.
Sample.
The online sample was drawn by means of the snowball method; raters were retrieved by sending invitations through social media network sites like Facebook and Twitter. A total of 369 respondents were recruited. 61,2% of the raters were female, and so men were slightly underrepresented in this study. Most of the raters were between 18-28 years old (76,6%) and were highly educated (68,9%).
Data.
Just as in the research of Rosar et al. (2008; 2012), the Truth of Consensus Method of Patzer (1985) was used, which is a method that entails the idea that attractiveness can be determined by a group of raters. The average of all individual ratings brings forward an ‘attractiveness score’ for each political candidate; this method is based on the idea that attractiveness can be seen as a characteristic that is perceived in a very similar way by a number of observers. In this study, every respondent rated nine or ten candidates, which is in accordance with the procedures carried out in previous studies (Berggren et al. 2010; Lautsten, 2013; Rosar et al., 2008, Rosar et al., 2012).
The candidates’ photographs were randomly grouped into 17 candidate-groups, and each respondent was randomly assigned to either one of these groups. Each candidate was ranked by 18 raters on average, which equals the number of subjects in similar studies (Banducci et al., 2008; Lautsten, 2013, Rosar et al., 2008; Rosar et al., 2012). After showing
a candidates’ picture, subjects answered to what degree they agreed upon statements about the candidates’ attractiveness and perceived competence (Chiao et al., 2008).
In the second step, the personal and political background of all candidates was researched on the internet, mainly on the candidates’ election campaign homepages or LinkedIn accounts. All information about the candidates was coded in a joint data set, in order to be used as control variables later on in the analysis. This procedure has been carried out before (Rosar et al., 2008; Rosar et al., 2012).
Dependent variables .
Electoral Success. The number of votes for the different parties and different candidates were gathered from the official websites of the three municipalities. The numbers of votes are used to determine a candidates’ relative success, as defined by earlier studies examining in proportional representation election systems (Berggren et al., 2010; Lautsten, 2013; Poutvaara et al., 2009). Relative Electoral Success is estimated as (pi/vj), in which pi represents the number of preference votes on the candidate, and where vj corresponds with the total number of votes divided by hundred in order to calculate the percentage relative success for each candidate. Since only the first one or two candidates receive a disproportionate number of personal votes, the variable Relative Electoral Success is highly skewed. Therefore, a logarithm of Relative Electoral Success is calculated (Lautsten, 2013). At final, the variable is recoded to a 0-1 scale where zero represents the minimal observed Electoral Success and 1 reflects the maximum Electoral Success.
Independent variables.
Attractiveness. Attractiveness was measured by a single item question, in which participants had to answer to what extend they agreed with the statement “The political candidate is attractive”. Attractiveness was measured on a 7 points Likert scale ranging from 1 strongly disagree to 7 strongly agree.
Competence. Since we used competence as an independent variables in the experiment, by manipulating both educational background and prior experience in the political field, we used two items representing these elements of competence in this real life study as well by means of two coded items: whether the candidate hold and academic degree and (has been) incumbency. The two items were entered as dummy variables in the linear regression model.
Results Study 2
All following empirical analysis are performed by means of linear regression models. The logarithmic Relative Electoral Success of the individual candidates represents the dependent variable in the regression models. In first stage of analysis, a model including candidates’ gender, attractiveness and competence was calculated (For more details see Model 1 in Table 4). In the first model we found significant effects for candidates’ gender (b=0.069, sd.=0.029, p<0.05), which entails that females have a slight disadvantage; female candidates receive 0.069% less of votes than their male opponents do. In addition, prior incumbency yields a positive effect on electoral outcome, if a candidate is or has been incumbent before the percentage of votes will increase with 0.126 points (b=0.126, sd.=0.028, p<0.01). Incumbency has been proven to be a strong predictor for electoral success before,
both in studies into two-party systems (Lee, 2001; Olivola et al. 2012) and multiparty systems (Berggren et al., 2010; Rosar et al., 2008).
In a following step, interaction effects were included. We found a small interaction effect between gender and incumbency (b=0.243, sd.0.172, p<0.05), which entails that previous or current incumbency has a positive effect for male candidates, but not so much for female candidates. No other significant interaction effects were found in this model.
Following the research of Rosar et al. (2008), candidates’ background variables were entered in the regression to determine their (partial) influence on the candidates’ share of votes. This model did only yields a small significant effect of candidates’ incumbency on Relative Electoral Success (See Model 3, Table 5), which indicates that (previous) incumbent candidates have a slightly higher chance of having electoral success, however the effect is relatively low, which indicates that this is only a weak effect of incumbency on electoral success. Attractiveness still does not yield a significant effect on electoral success.
[TABLE 5 ABOUT HERE]
In the third model, the candidates’ place on the list, the logarithmic number of opponents and their parties were included as well. This model also shows that the candidates’ place on the list has a significant, strong effect on electoral success (b=-0.550, sd.=0.150, p<0.01) in a negative direction. This indicates that the higher a political candidate is on the list, that person receives fewer votes from the electorate. When looking closely to the electoral outcomes percentages, the confirmation of this hypothesis is not surprising: roughly all party leaders receive between 60-80% of all party votes, where after the second and third person get a share of 10%, and all other candidates on the electoral role less than 1% of the
preference votes. This indicates that the electorate votes for the first person on the list in order to get their party into the council, rather than actually voting for the party leader. The more votes the first person and the party gets, the more candidates get into the council from that party. This finding is in accordance with prior literature (Lautsten, 2013, Rosar, 2008).
Discussion Study 2
The aim of this experiment was to test to what extent the patterns found in the experiment were applicable to a real-world situation. Contradictory to previous research, attractiveness does not have a significant effect on electoral success. However, we did find a significant main effect of gender of the candidate on electoral success, which is a finding in accordance with previous research (Berggren et al., 2010). In addition, just as in the results of the experiment, competence has a great influence on electoral success. The fact that a candidate is or has been incumbent, and so has prior experience in the field, seems an important predictor when it comes to electoral outcome, which is in line with previous research (Lee, 2001; Rosar et al. 2012). Whether a candidate holds an academic degree has a marginal effect on electoral outcome as well, and so seems to be partially influencing ones decision to vote for a certain candidate.
Still, the place on the electoral role seems to be the most important determinant for one’s electoral success (Lee, 2001; Rosar et al. 2012). Therefore, one might assume that the choices made by the voters, are most probably lead by political ideology and the presented party program; and express the preference for this party by voting on the first person on the list. Studies into the effect of party affiliation, list place and candidate order have identified this phenomenon before (Miller & Krosnick, 1998).
General Discussion.
The aim of these studies was to increase general knowledge about 1) whether the appearance of a political candidate has more influence than the actual competence on candidates’ perceived competence, perceived leadership and attitude towards the candidate 2) to what extend the found patterns of the first study apply to the ‘real world’; and more specifically what the effect of attractiveness is on real life electoral outcome when controlling for background variables, in a multi-party electoral system. This paper responds to the research gap of knowledge about the influence of physical attractiveness of a political candidate on candidate evaluations and electoral success.
Despite the fact that several researchers already have devoted their studies to determine the effect of facial attractiveness on candidate evaluation and electoral success, both of the studies performed contribute in unique ways to prior studies. First of all, none of the previous studies focused on the workings between visual and textual representation of a candidate, in order to determine the effect of both of the elements, instead of only one (eg attractiveness only). By examining the effect of both of the elements, a deeper insight the workings of candidate evaluation, and maybe eventually about electoral choices can be achieved.
Secondly, the majority of the studies have been performed in the United States, possibly influences the generalizability of the outcomes. Since research of Curran et al. (2009) showed that individuals are overall less politically informed, and so possibly make decisions based on a candidates’ appearance rather than on a candidates’ competency, there is a need for research performed outside of the United States to determine to what extent the findings hold across different political and cultural contexts. The somewhat pessimistic view that voters base their decisions on superficial cues such as candidates’ attractiveness is hereby challenged. It seems that voter certainly take more grounded determinants such as
competence in consideration when evaluating a candidate. Comparative research in which identical studies are performed in several countries at once; could give further insights into the differences between candidate evaluation and voting behaviour of individuals of different countries.
Attractiveness mostly seems to function as a beneficial asset, which causes a slightly more positive attitude of voters, but does not really do more than that. A more important factor for voters to evaluate a candidate is competence; the previous education and experience of a candidate seem important factors for both perceived competence, perceived leadership and affects a candidates’ electoral outcome.
When looking closely to some of the prior literature a possible explanation for the differences in the results can be found: Previous studies mostly focused on candidate evaluations made in the ‘blink of an eye’ and their correlations with electoral outcomes (Ballew & Todorov, 2007; Todorov et al., 2005) or on the effects of candidate evaluations and the effect on hypothetical voting likelihood (Chiao et al., 2008). These studies examine the influence of attractiveness by the means of exclusively showing candidate photographs, with no other information, after which respondents are ought to make a choice between candidates. These results were compared with actual election outcomes afterwards. When researching attractiveness and thereby ignoring the possible influence of other determinants, it could be that attractiveness has a bigger effect when there is no control for other variables. This would explain the different outcomes between previous studies and the studies described in this article.
In addition, one could also argue that the suggested relevance of attractiveness sometimes gets exaggerated. Even though attractiveness ought to have a great influence on electoral outcome, the actual effects of attractiveness that are found are overall quite small
(Barrett & Barrington, 2005; Beggren et al., 2010; Chiao et al., 2008; Lawson et al., 2010; Olivola & Todorov, 2010; Rosar et al, 2008; 2012). The great majority of previous studies report on the effect of attractiveness, while the report on the regression coefficients are most of the time relatively low and lack any kind of details about the Pearson’s product correlation. Therefore, more studies need to be done in order to develop a more valid notion on the influence of attractiveness on electoral success.
Limitations and suggestions for future research.
There are several limitations in the performed studies described in this article. First of all, because of practical concerns, only political website advertisements were included. A web advertisement, in this case a non-interactive, is a very static way of advertising, since only the candidates’ picture and a small text have been used to convey a message. In all probability, dynamic communication tools, in which a candidate is filmed or audiotaped, would lead to different results. Therefore, future research should include different types of media in order to determine the effect of candidates’ attractiveness on candidates’ evaluations.
The manipulations used in this experiment could be seen as a limitation of the experimental study as well. Despite that fact that the manipulation checks were significant and I therefore assumed that the manipulations succeeded, the political website advertisements were manually designed by the researcher and not by a professional advertising agency or marketing team. Possibly, participants of the experiment were conscious to the fictional nature of the political advertisements, which could have influenced their responses, and eventually even the results. In addition, participants were not exposed to the sixth condition containing an unattractive, not-competent female candidate due to
problems with the survey-tool Qualtrics. The results of the experiment may be influenced by this error.
Another caveat is that in the current studies variables such as perceived competence, perceived leadership and attitude towards the candidate were only measured by one item. For the reliability of the answers, a scale with multiple items measuring the concepts would be highly beneficial for the outcomes.
At final, the second study which is aimed at examining the influence of attractiveness on electoral outcome in a ‘real-world’ case consisted of a relative small sample. Only 173 political candidates were analysed. A suggestion for future research is to examine the influence of attractiveness on electoral outcome with the use of a bigger database in order to enhance the validity of the findings.
Literature
Amos, C., Holmes, G., & Strutton, D. (2008). Exploring the relationship between celebrity endorser effects and advertising effectiveness. International Journal of Advertising, 27(2), 209-234.
Antonakis, J., & Dalgas, O. (2009). Predicting elections: Child’s play! Science, 323, 1183. Artz, L., & Kamalipour, Y. R. (Eds.). (2005). Bring ’em on. Media and politics in the Iraq war. New York: Rowman & Littlefield.
Bailenson, J. N., Garland, P., Iyengar, S., & Yee, N. (2006). Transformed facial similarity as a political cue: A preliminary investigation. Political Psychology,27(3), 373-385.
Ballew, C. C., & Todorov, A. (2007). Predicting political elections from rapid and unreflective face judgments. Proceedings of the National Academy of Sciences,104(46), 17948-17953.
Banducci, S. A., Karp, J. A., Trasher, M. & Rallings, C. Ballot Photographs as Cues in Low-Information Elections. Political Psychology, 29 (6), 903-917.
Barrett, A. W., & Barrington, L. W. (2005). Is a picture worth a thousand words? Newspaper photographs and voter evaluations of political candidates. The Harvard International Journal of Press/Politics, 10(4), 98-113.
Benjamin, D. J. & Shapiro, J. M. (2009). Thin-Slice Forecasts of Gubernatorial Elections. Review of Economics and Statistics, 91 (3), 523–536.
Bennett, W. L. (2009). News: A political illusion. Pearson/Longman.
Bennett, W. L., & Iyengar, S. (2010). The shifting foundations of political communication: Responding to a defense of the media effects paradigm. Journal of Communication, 60(1), 35-39.
Bennett, W. L., & Segerberg, A. (2012). The logic of connective action: Digital media and the personalization of contentious politics. Information, Communication & Society, 15(5), 739-768.
Berggren, N., Jordahl, H., & Poutvaara, P. (2010). The looks of a winner: Beauty and electoral success. Journal of Public Economics, 94(1–2), 8–15.
Brants, K., & Van Praag, P. (2006). Signs of media logic: half a century of political communication in the Netherlands. Javnost-the public, 13(1).
Caprara, G. V., & Zimbardo, P. (2004). Personalizing politics: A congruency model of political preference. American Psychologist, 59, 581–594.
Caprara, G. V., Schwartz, S., Capanna, C., Vecchione, M., & Barbaranelli, C. (2006). Personality and politics: Values, traits, and political choice. Political Psychology, 27(1), 1-28. Chiao, J. Y., Bowman, N. E., & Gill, H. (2008). The political gender gap: Gender bias in facial inferences that predict voting behavior. Plos One, 3(10), 1-7.
Conger, J. A. (1989). The charismatic leader: Behind the mystique of exceptional leadership. Jossey-Bass.
Curran, J., Iyengar, S., Lund, A. B., & Salovaara-Moring, I. (2009). Media System, Public Knowledge and Democracy A Comparative Study. European Journal of Communication, 24(1), 5-26.
Dalton, Russell J., I. McAllister & M.P. Wattenberg (2000) ‘The Consequences of Partisan Dealignment’, in J. Russell Dalton and M.P. Wattenberg (eds) Parties Without Partisans: Political Change in Advanced Industrial Democracies, pp. 37–63. Oxford: Oxford University Press.
De Vreese, C. H. (2007). Digital renaissance: young consumer and citizen? The Annals of the American Academy of Political and Social Science, 611(1), 207-216.
Druckman, J. N. (2003). The Power of Television Images: The First Kennedy‐Nixon Debate Revisited. Journal of Politics, 65(2), 559-571.
Eagly, A. H., & Johannesen‐Schmidt, M. C. (2001). The leadership styles of women and men. Journal of social issues, 57(4), 781-797.
Eagly, A. H., & Karau, S. J. (2002). Role congruity theory of prejudice toward female leaders. Psychological review, 109 (3), 573.
Entman, R. M. (1993). Framing: Toward clarification of a fractured paradigm. Journal of communication, 43(4), 51-58.
Field, A. (2013). Discovering statistics using IBM SPSS statistics. London: SAGE publications.
Gibson, R., Ward, S., & Lusoli, W. (2002). The internet and political campaigning: the new medium comes of age? Representation, 39 (3), 166-180.
Gilliam Jr, F. D., & Iyengar, S. (2000). Prime suspects: The influence of local television news on the viewing public. American Journal of Political Science, 560-573.
Graber, D. A. (1990). Seeing is remembering: How visuals contribute to learning from television news. Journal of communication, 40(3), 134-156.
Habermas, J. (2006). Political communication in media society: Does democracy still enjoy an epistemic dimension? the impact of normative theory on empirical research. Communication theory, 16(4), 411-426.
Harcup, T., & O'neill, D. (2001). What is news? Galtung and Ruge revisited. .Journalism studies, 2(2), 261-280.
Hart, W., Ottati, V. C., & Krumdick, N. D. (2011). Physical attractiveness and candidate evaluation: A model of correction. Political Psychology, 32(2), 181-203.
Henn, M., M. Weinstein, and D. Wring. 2002. A generation apart? Youth and political participation in Britain. British Journal of Politics and International Relation, 4 (2): 167-92.
Hess, U., Adams Jr, R. B., & Kleck, R. E. (2004). Facial appearance, gender, and emotion expression. Emotion, 4(4), 378.
Hox, J.J. (2002). Multilevel analysis: Techniques and applications. Mahwah, NJ/London: Lawrence Erlbaum Associates.
Kaase, M. (1994). Is there personalization in politics? Candidates and voting behavior in Germany. International Political Science Review, 15(3), 211-230.
King, A., & Leigh, A. (2009). Beautiful politicians. Kyklos, 62(4), 579–593.
Kriesi, H. (2012). Personalization of national election campaigns. Party Politics,18(6), 825-844.
Laustsen, L. (2013). Decomposing the Relationship Between Candidates’ Facial Appearance and Electoral Success. Political Behavior, 1-15.
Lawless, J. L. (2004). Women, war, and winning elections: Gender stereotyping in the post-September 11th era. Political Research Quarterly, 57(3), 479-490.
Lenz, G. S., & Lawson, C. (2011). Looking the part: Television leads less informed citizens to vote based on candidates’ appearance. American Journal of Political Science, 55(3), 574-589.
Little, A. C., Burriss, R. P., Jones, B. C., & Roberts, S. C. (2007). Facial appearance affects voting decisions. Evolution and Human Behavior, 28(1), 18-27.
McGregor, J. (2002). Restating news values: Contemporary criteria for selecting the news. In Refereed articles from the Proceedings of the ANZCA 2002 Conference, Coolangatta. Communication: Reconstructed for the 21st Century.
McQuail, D. (2010). McQuail's mass communication theory. London: Sage Publications.
Meindl, J. R. (1990). On leadership, an alternative to coventional wisdom. Research in organizational behavior, 12, 159-203.
Miller, J. M., & Krosnick, J. A. (1998). The impact of candidate name order on election outcomes. Public Opinion Quarterly, 291-330.
Nagel, F., Maurer, M., & Reinemann, C. (2012). Is there a visual dominance in political communication? How verbal, visual, and vocal communication shape viewers' impressions of political candidates. Journal of Communication, 62(5), 833-850.
Olivola, C. Y., & Todorov, A. (2010). Elected in 100 milliseconds: Appearance-based trait inferences and voting. Journal of Nonverbal Behavior, 34(2), 83-110.
Olivola, C. Y., Sussman, A. B., Tsetsos, K., Kang, O. E., & Todorov, A. (2012). Republicans Prefer Republican-Looking Leaders Political Facial Stereotypes Predict Candidate Electoral Success Among Right-Leaning Voters. Social Psychological and Personality Science, 3(5), 605-613.
Ottati, V. C., & Deiger, M. (2002). Visual cues and the candidate evaluation process. In The social psychology of politics (pp. 75-87). Springer US.
O’Toole, T., M. Lister, D. Marsh, J. Jones, and A. McDonagh. 2003. Tuning out or left out? Participation and nonparticipation among young people. Contemporary Politics 9 (1): 45-61
Reinemann, C., and J. Wilke (2007) ‘It’s the Debates, Stupid! How the Introduction of Televised Debates Changed the Portrayal of Chancellor Candidates in the German Press, 1949–2005’, The Harvard International Journal of Press/Politics, 12(4): 92–111.
Rosar, U., Klein, M. & Beckers, T. (2008) The frog pond beauty contest: Physical attractiveness and electoral success of the constituency candidates at the North Rhine-Westphalia state election of 2005. European Journal of Political Research, 47, 64-79.
Rosar, U., Klein, M., & Beckers, T. (2012). Magic Mayors: Predicting Electoral Success from Candidates' Physical Attractiveness under the Conditions of a Presidential Electoral System. German Politics, 21(4), 372-391.
Scheufele, D. A., and M. C. Nisbet. 2002. Being a citizen online: New opportunities and dead ends. Harvard International Journal of Press/Politics 7, (3), 55-75.
Shavitt, S., Swan, S., Lowrey, T. M., & Wänke, M. (1994). The interaction of endorser attractiveness and involvement in persuasion depends on the goal that guides message processing. Journal of Consumer Psychology, 3(2), 137-162.
Surawski, M. K., & Ossoff, E. P. (2006). The effects of physical and vocal attractiveness on impression formation of politicians. Current Psychology, 25(1), 15-27.
Swanson, D. L., & Mancini, P. (Eds.). (1996). Politics, media, and modern democracy: An international study of innovations in electoral campaigning and their consequences. Greenwood Publishing Group.
Till, B. D., & Busler, M. (2000). The match-up hypothesis: physical attractiveness, expertise, and the role of fit on brand attitude, purchase intent and brand beliefs. Journal of
Advertising, 29(3), 1-13.
Todorov, A., Mandisodza, A. N., Goren, A., & Hall, C. C. (2005). Inferences of competence from faces predict election outcomes. Science, 308(5728), 1623-1626.
Weber, M. (1968). On charisma and institution building. In: Awamleh, R., & Gardner, W. L. (1999). Perceptions of leader charisma and effectiveness: The effects of vision content, delivery, and organizational performance. The Leadership Quarterly, 10, (3), 345-373.
Appendix 1 – Tables. Table 1 Model 1 N=205 R2= 0.302 Model 2 N=205 R2= 0.308 Model 3 N=205 R2= 0.334 Candidates’ features Intercept 4.158 (0.349)*** 4.159*** 2.788*** Candidates gender -0.086 (0.109) -0.791 (0.313)** -0.804 (0.160)** Candidates attractiveness 0.137 (0.114) -0.483 (0.323) -0.554 (0.324)* Candidates competence 1.646 (0.119)*** 0.993 (0.322)** 0.935 (0.321)**
Candidates gender x attractiveness 0.501 (0.253)** 0.526 (0.254)**
Candidates gender x competence 0.542 (0.262)** 0.541 (0.259)**
Attractiveness x competence 0.382 (0.280) 0.456 (0.280)
Control variables
Participants age 0.024 (0.008)**
Participants gender 0.361 (0.160) **
Participants political trust 0.121 (0.076)
Table 1. Multi-level regression model of gender, attractiveness, competence on perceived competence.
Table 2 Model 1 N=205 R2= 0.105 Model 2 N=205 R2= 0.115 Model 3 N=205 R2= 0.154 Candidates’ features Intercept 3.370 (0.160)*** 3.305 (0.354)*** 2.510 (0.541)*** Candidates gender -0.001 (0.121) -0.142 (0.338)** -0.192 (0.334) Candidates attractiveness 0.240 (0.122)** -0.441 (0.340) -0.351 (0.335) Candidates competence 0.848 (0.128)*** 0.966 (0.334)** 0.876 (0.335)**
Candidates gender x attractiveness 0.073 (0.286) 0.123 (0.279)
Candidates gender x competence 0.231 (0.269)** 0.241 (0.267)
Attractiveness x competence -0.436 (0.288) -0.364 (0.292)
Control variables
Participants age -0.009 (0.010)
Participants gender 0.282 (0.159) *
Participants political trust 0.182 (0.072)**
Table 2. Multi-level model regression model of gender, attractiveness and competence on perceived leadership.
Note. Coefficients are unstandardized, standard errors are robust. *p<0.10 ** p<0.05, **** p<0.01.
Table 3 Model 1 N=205 R2 = 0.139 Model 2 N=205 R2 = 0.142 Model 3 N=205 R2 = 0.213 Candidates’ features Intercept 2.875*** 3.184*** 2.241*** Candidates gender -0.149** -0.450** -0.478** Candidates attractiveness 0.492*** 0.194 0.121 Candidates competence 0.323*** 0.079 0.019
Candidates gender x attractiveness 0.282 0.317**
Candidates gender x competence 0.161 0.163
Attractiveness x competence 0.162 0.221
Control variables
Participants age 0.003
Participants gender 0.286**
Participants political trust 0.149**
Table 3. Multi-level regression model of gender, attractiveness and competence on attitude towards candidate.
Table 4.
Manipulation Perceived competence Perceived leadership Attitude towards candidate Female, attractive, competent 5.00 (1.199) 4.31(1.143) 3.61 (0.649) Female, attractive, incompetent 3.72 (1.293) 3.74 (1.317) 3.34 (0.769) Female, unattractive, competent 5.16 (1.247) 4.258 (1.290) 3.30 (0.762) Male, attractive, competent 5.27 (0.970) 4.41 (1.045) 3.59 (0.652) Male, attractive, incompetent 3.41 (1.428) 3.68 (1.330) 3.20 (0.812) Male, unattractive, competent \ 4.96 (1.095) 4.37 (1.316) 3.02 (0.674) Male, unattractive, incompetent 3.35 (1.306) 3.16 (1.095) 2.73 (0,686)
Table 4 Model 1 N=168 R2= 0.156 Model 2 N=168 R2= 0.222 Model 3 N=168 R2= 0.334 Candidates’ features Intercept 0.585 (0.051)*** 0.527 (0.063)*** 0.664 (0.185)*** Candidates is a male 0.069 (0.029)** 0.627 (0.783) 0.544 (0.689) Candidates attractiveness 0. 016 (0.114) 0.034 (0.046) 0.019 (0.041)
Competence: Holds an academic title 0.024 (0.032) 0.030 (0.033) 0.048 (0.028)*
Competence: (Has been) incumbent 0.126 (0.028)*** 0.121 (0.028)*** 0.095 (0.024)***
Interactions Male * attractiveness 0.203 (0.197) 0.158 (0.173) Male * incumbency 0.234 (0.172)** 0.195 (0.151) Male * degree 0.006 (0.020) 0.098 (0.003) Incumbency * attractiveness 0.004 (0.008) 0.049 (0.007) Degree * attractiveness 0.020 (0.012) 0.135 (0.037)
Control background variables
Candidate has immigration background -0.031 (0.026)
Candidate came out as homosexual 0.017 (0.050)
Candidates number of opponents 0.461 (0.308)
Candidates place on list -0.550 (0.150)***
Candidates political party:
Partij van de Arbeid 0.041 (0.117)
Volkspartij Vrijheid en Democratie 0.058 (0.115)
Democraten 66 0.160 (0.105)
GroenLinks 0.009 (0.129)
ChistenUnie 0.067 (0.118)
Christen Democratisch Appèl 0.002 (0.104)
Table 4 . Regression for attractiveness and background features on electoral outcome.
Manipulation condition 1
Manipulation condition 2
Manipulation condition 3
Manipulation condition 4
Manipulation condition 5
Manipulation condition 6*
Female, unattractive, not competent
Manipulation condition 7
Manipulation condition 8
