Gender in politics : an experimental study on gender and the evaluation of compe-tence and credibility

(1)

Gender in Politics:

An experimental study on gender and the evaluation of

compe-tence and credibility

University of Amsterdam: Graduate School of Communication

Master specialization: Political Communication

Name:

Lynnemore van Ommen

Student ID: 10009000

Supervisor: Dr. Marjolein Moorman

Date:

25 June 2015

(2)

ABSTRACT

The purpose of the current research was to explore whether gender of a politician is decisive in the evaluation of the competence and credibility of this politician. Specifically, whether respondents respond different to an angry female politician than to an angry male politician. Furthermore, it is examined whether the gender of the respondents is crucial in this evalua-tion.

A pre-test post-test 2 (gender: female vs male) x 2 (emotion: angry vs no emotion) x 2 (context: political vs not political) between-subjects factorial design was conducted among 160 respondents to test the hypothesis.

By means of the current experiment it could be argued that female politicians are not disadvantaged when it concerns the evaluation of their competence and credibility. Unex-pectedly male politicians appeared to be rather disadvantaged compared to female politicians. The female politician was evaluated as more competent and more trustworthy than the male politician. However, in a not political context man and woman have a greater possibility to be evaluated more positive on competence and credibility. Moreover, women rather than men are expected to show a gender based voting effect.

These findings have theoretical implications for the academic literature since it shows evidence for the fact that respondents react different to a political context than to a non-political context. There is an enduring discussion on whether there has to be a certain per-centage of women in the parliament, a “quota”. According to the results the female-quota is worthwhile considering.

(3)

INTRODUCTION

Today with the American elections and Hillary Clinton running for presidential candidate the focus on gender in politics is present more than ever. In 2008 Clinton weakened her historical role as the first female presidential candidate who seriously could win the election by means of highlighting the fact that ‘she did not nominated herself as a candidate because she is a women’ (Vrij Nederland, 15 April 2015). It turned out that her advisor had recommended her to weaken her womanhood. Unfortunately with no success. However, today Clinton is not only straightforward about her age, she is as well genuine about her womanhood. By means of #GrandmothersKnowBest she implicates that a political disadvantage – her age and gender – in fact appears to be an advantage. As for today, this strategy seems to be more successful since she is officially nominated as presidential candidate for the Democrats.

Not only in America women in the parliament seem a trending topic, in the Nether-lands too there is an enduring discussion on whether there has to be a certain percentage of women in the parliament, a “female-quota” (Financieel Dagblad, 12 September 2014). Ques-tions that are raised concern the fact whether the Netherlands has not have a female prime minister yet? Besides, why women are often second on the election-list, though are never chosen as chairman of a parliamentary party? However, the essence is: are female politicians taken even as serious as male politicians in that they are equally evaluated on their abilities?

It is often questioned whether voters rely on the gender of political candidates for the interpretation of the abilities of the candidate, the competence, credibility, and positions taken by the politicians concerning policies that are related to politics (Cook, 1998; Paolino, 1995; Zipp & Plutzer, 1985, as cited in Schneider, 2014). No research has focused on the influence of the gender of a politician on the evaluation of their abilities. In other words, a lack of re-search exists concerning whether the gender of a politician is criticized without being rele-vant to a specific context, thus only based on a pre-assumption. For this reason it would be

(4)

interesting to explore whether a gender effect will occur in politics.

When discussing critical topics or when emphasizing certain statements, in politics this is often accompanied by emotional expressions (i.e. anger, happiness, or sadness). It ap-pears that emotions like anger have the ability to actively change opinions (Lecheler, Schuck & de Vreese, 2013). Besides it is argued that expressing anger in a professional context can have a different effect for men than for women. When men express an angry emotion this can lead to a higher social status, while when women express an angry emotion this can have a reversed effect in that a lower social status will be ascribed ( Brescoll & Uhlmann, 2008). Since it is argued that once people identify the emotion of a person, they identify other char-acteristics of that person (Tiedens, 2001) it could be interesting to see whether emotions, es-pecially anger, are determining in the evaluation of politicians.

The aim of the current research is to explore whether the gender and the emotion ex-pressed by a politician is decisive in the evaluation of the competence and credibility of the politician:

RQ 1: To what extent do respondents evaluate a female politician expressing an angry

emo-tion as less competent and less credible than a male politician expressing an angry emoemo-tion?

Besides the lack of research on the gender of politicians, there is even less research on whether the gender of the voter plays a role in evaluating political candidates (Schneider, 2014). That is why it will additionally be explored whether the gender of the respondent (i.e. a potential voter) has an effect on the relation between the gender of the politician and the evaluation of competence and credibility. The gender of the respondents is in this case the moderating variable.

By exploring the main research question and the moderating effect, this study has po-tential to further understanding of gender differences in evaluating politicians and offers in-sight into why female politicians are possibly evaluated as less competent and credible than

(5)

male politicians. Furthermore the current research is renewing, as it explores not only in a political context whether gender of a person is decisive in the evaluation of competence and credibility, it also explores this possible gender effect in a not political context. As for today women in politics is a trending topic (e.g. the female-quota) expanding knowledge on the in-fluence of gender seems important. When it appears that women are taken less serious than men in politics, it is clear that the image of female politicians has to be changed otherwise it makes no sense to demand a female-quota.

THEORETICAL FRAMEWORK

Gender in politics

The Inter-Parliamentary Union (IPU, 1 April, 2015) recently published a list of women in the national parliament showing that in general the average number of women in the national par-liament is higher in multi-party systems (i.e. 37% in the Netherlands and 39% in Belgium) than in two-party systems (i.e. 19% in the US and 22% in the UK). The current study is con-ducted in the Netherlands. As the Netherlands appears to be a country where the percentage of women in the parliament is high, it is important to expand knowledge on the image of fe-male politicians.

Dolan (2010) and Sanbonmatsu (2002) have shown evidence for a “baseline gender preference”. In its purest form, baseline gender preference refers to the fact that voters in general have a preference for a male or female political candidate. Dolan (2010) and San-bonmatsu (2002) demonstrate the ways in which sex stereotypes are able or unable to influ-ence the willingness of the public to support women or men in electoral situations (Dolan, 2010). In their research the majority of American voters showed a clear baseline gender pref-erence for a male presidential candidate, while the minority argued that they would prefer a female presidential candidate (Dolan, 2010; Sanbonmatsu, 2002). The gender of the

(6)

candi-date appears to be decisive in the voters’ evaluation of the competence, ideological beliefs, and policy position of the candidate (Sanbonmatsu, 2002).

These findings are confirmed by the study of Erzeel and Caluwaerts (2015). Erzeel and Caluwaerts (2015) showed that female candidates are disadvantaged: the majority of the Belgium voters preferred a male candidate over a female candidate while the remaining mi-nority preferred a female candidate over a male candidate. In addition to the “baseline gender preference”, Holli and Wass (2010) found that age and party orientation mediate the gender-based voting effect. “Baseline gender preference” in Finland appeared to be more common among young women and older men. As for party orientation, left-wing party voters have a higher tendency to prefer a female candidate than right-wing party voters. Right-wing party voters have a higher tendency to vote for a male candidate (Erzeel & Caluwaerts, 2015).

Competence & Credibility

Previous studies on the effects of gender in relation to the evaluation of the abilities or char-acter traits have presented evidence for the fact that gender is determining in the level of so-cial status a person gets allocated (Tiedens, 2001; Brescoll & Uhlmann, 2008). In the current study the effect of gender on the evaluation of competence and credibility will be tested. The reason for this is that competence and credibility are determinants of social status (Tiedens, 2001; Brescoll & Uhlmann, 2008). Competence is related to being strong, dominant, and smart and being less warm, friendly, and nice (Tiedens, 2001). The credibility of a source re-fers to positive characteristics of the communicator that can have an effect on the acceptance of the message by the receiver (Ohanian, 1990). Credibility is determined by means of three components: attractiveness, trustworthiness, and expertise. According to the “attractiveness model” by McGuire (1985) (as cited in Ohanian, 1990), the successfulness of a message is reliable upon the spectators’ judgment of the attractiveness of the source. Trustworthiness is

(7)

defined by the perceived confidence of the communicator in communication the statement he or she thinks is most valuable (Hovland, Janis, & Kelley, 1953, as cited in Ohanian, 1990). Expertise is defined by the perception of the communicator as being a source of valuable statements.

Hypothesis 1: In comparison to a non-political context, respondents will respond different to

a male politician than to a female politician. They will evaluate the male politician as more competent and more credible than the female politician.

Politicians expressing emotions

As the public forms an opinion by means of visual and verbal appearances of politicians, and as often in this visual and verbal appearances politicians express a particular emotion, it is interesting to see whether a difference exists in response to female politicians expressing a particular emotion and male politicians expressing the same emotion.

As argued before, not only it appears that anger is able to actively change opinions (Lecheler et al., 2013). It appears as well that, in a professional context, expressing anger has a positive effect on the evaluation of the social status of men while it has a negative effect on the evaluation of the social status of women ( Brescoll & Uhlmann, 2008; Tiedens, 2001; Broos, 2011). That is why anger is chosen as an emotion in the current research. It will be explored whether the public response differently to a female politician expressing anger than to a male politician expressing anger.

Brescoll and Uhlmann (2008) conducted three different studies, among American adults, on the relationship between anger, gender, and allocating status. Status, to be specific social status, is determined by positive and negative valued characteristics. A positive valued

(8)

characteristic is related to a higher status (Anderson, John, Keltner, & Kring, 2008). Express-ing anger indicates that a person is strong, competent and credible (Tiedens, 2001).

Tiedens (2001) found that, in a professional context during a job interview, men who expressed anger were more likely to be employed and to enjoy an increase in social status. However, in contrast, when women expressed anger during the interview, this could cause a decrease rather than an increase in their social status. A possible explanation for this differ-ence could be the fact that women are supposed to be more affectionate, and more modest than men. If women fail to correspond to this prescriptive stereotype, they can evoke negative responses among people (Heilman, 2001; Rudman, 1998). Brescoll and Uhlmann (2008) ar-gue that when a woman expresses anger the emotion will be ascribed to internal causes (e.g. caused by her personality: “she is an angry person”) rather than to external causes (e.g. caused by an external circumstance: “the situation was frustrating”). People will take the women less serious when she is angry because they assume that she is just an angry person and that she is not angry on behalf of the situation.

As stated above Tiedens (2001) conducted an experiment on anger and gender by means of job interviews. The present study looks at a different professional context: the ap-pearance of political candidates. The findings regarding the professional context, the job-interviews, could serve as a foundation for the current study.

The results of the three studies of Brescoll and Uhlmann (2008) present evidence for the fact that respondents, that were exposed to an angry man, viewed the man as more com-petent than the angry woman. Angry women were evaluated as more out of control than an-gry men. As a consequence the social status of anan-gry women appeared to be mediated by as-cribing anger to being out of control. Concluding, expressing anger can have a different effect for men than for women. Men expressing anger refers to a higher social status, while for

(9)

women expressing anger can have a reversed effect in that a lower social status will be as-cribed.

Proceeding on the findings of Brescoll and Uhlmann (2008), Broos (2011) conducted an experiment to test whether emotional appearances are likely to influence evaluations of female versus male politicians. In general, people expect men to be more likely to express anger than women (Grossman & Wood, 1993; Plant, Hyde, Keltner, & Devine, 2000). An explanation for the fact that it is expected that women are less likely to express anger is the fact that anger could be seen as an emotional version of toughness, and toughness is danger-ous for female candidates to show (Washington Post, 7 November 2007). In the article of the Washington Post it is argued that “a strong man is admired”, while a “strong woman is…it rhymes on rich”. In addition, Brescoll and Uhlmann (2008) present evidence for the fact that, concerning a professional context, female candidates are punished far more often for express-ing anger than male candidates. By means of his experiment, Broos (2011) presents evidence for the fact that the expectation, with respect to anger, among voters is different for female candidates than for male candidates. When a female candidate expresses anger, this will be evaluated as less appropriate and more emotional than a male candidate expressing anger.

To conclude, concerning a professional context it appears that women expressing an-ger enjoy a lower social status while men expressing anan-ger enjoy a higher social status. Anan-ger expressed by women is ascribed to personality while anger expressed by men is ascribed to an external cause.

Hypothesis 2: In comparison to a non-political context, the respondents will evaluate a male

politician expressing an angry emotion as more competent and more credible than a female politician expressing an angry emotion.

(10)

Moderating Effects

As argued before, a “baseline gender preference” exists when it comes to politics. This base-line gender preference not only refers to the fact that voters prefer a candidate of a particular gender, it also refers to the fact that voters may choose for a candidate of their own gender; “gender based voting” (Erzeel & Caluwaerts, 2015). Since the gender of the candidate plays a role in the evaluation of the angry emotion they express, it is interesting to explore whether the gender of the voter as well plays a role in these main effects, whether female voters show different tendencies than male voters towards political candidates and their emotional expres-sions.

Concerning the gender of voters in the evaluation of female or male politicians, vari-ous studies have presented evidence for a clear gender based voting effect in that voters pre-fer candidates of the same gender (Holli & Wass, 2010). To highlight, Sanbonmatsu (2002) presents evidence for the fact that most women express a preference for a female candidate, while most men prefer a male candidate. However, they present as well that “gender-based voting” is more evident for male voters than to female voters (Erzeel & Caluwaerts, 2015). This is as well confirmed by the results of Holli and Wass (2010). They highlight the fact that men have a higher tendency to choose a candidate of their own gender than women.

Concerning the evaluation of female or male candidates expressing an angry emotion, Hinsz and Tomhave (1991) argue that women respond different to positive or negative in-formation than men. In general women are more likely to react with a positive emotion to a picture of a woman or man than men. Men are more likely to react with a neutral emotion to a picture of a woman or man. In addition, Barret and Barrington (2005) showed evidence for the fact that women are less influenced by a negative picture of a political candidate, the chance that they would support the candidate when being exposed to a negative picture of the candidate was about the same as when not being exposed to a picture of the candidate. In

(11)

contrast, men were much less supportive to a candidate when being exposed to a negative picture of that candidate than when not being exposed to a picture. Since according to Barret and Barrington (2005) women react less negative to a negative photograph of a candidate than men it could be expected that women will respond less negative to a man or woman ex-pressing an angry emotion compared to men.

Hypothesis 3: A gender effect will occur in that compared to female respondents, male

re-spondents will evaluate the male politician more positive on competence and credibility than the female politician. This gender effect will be stronger for male respondents than for female respondents.

Hypothesis 4: Compared to male respondents, female respondents will evaluate both the

an-gry female and anan-gry male politician more positive on competence and credibility. This gen-der effect will be stronger for female respondents than for male respondents.

METHOD

Research method

A pre-test post-test 2 (gender: female vs male) x 2 (emotion: angry vs no emotion) x 2 (con-text: political vs not political) between-subjects factorial design was conducted. One experi-mental newspaper front page was designed for all 8 experiexperi-mental conditions, the front page differed in the picture that was displayed (Appendix A).

Participants

A total of 160 respondents were recruited for the experiment: 20 participants per condition (Table 1). Fifty percent of the participants were male and 50% were female. To generate a

(12)

representative sample of the Dutch society, the age of the participants ranged between 18 and 85 years old. Respondents that were aware of the experiment, that were familiar with the people that were displayed on the picture or that were younger than 18 years could not partic-ipate in the experiment.

Stimulus material

Since Todorov, Mandisodza, Goren, and Hall (2005) have revealed that when individuals are exposed to a photograph of a candidate for only a split second their evaluation of the ability of the candidate to perform can already be influenced, the stimulus material existed of a real looking newspaper front page. This front page, among others, displayed a male or female in a political context or in a non-political (scientific) context. The political context was accompa-nied by the phrase “Partij wil stem krijgen in kabinetsbeleid” (translation: “Party want to have a voice in the policy of the cabinet”) and the non-political context was accompanied by the phrase “De loopbaan. Hoe ver zijn we? [Wetenschap]” (translation: “The career. Where do we stand? [Science]”). Concerning the eight experimental conditions, participants in the first condition were exposed to a newspaper front page that showed a picture of a female poli-tician expressing a not angry emotion. Participants in the second condition were exposed to a newspaper front page that showed a picture of a male politician expressing a not angry emo-tion. Participants in the third condition were exposed to a newspaper front page that showed a picture of a female politician expressing an angry emotion. Participants in the fourth condi-tion were exposed to a newspaper front page that showed a picture of a male politician ex-pressing an angry emotion. Participants in the fifth condition were exposed to a newspaper front page that showed a picture of a woman expressing a not angry emotion. Participants in the sixth condition were exposed to a newspaper front page that showed a picture of a man expressing a not angry emotion. Participants in the seventh condition were exposed to a

(13)

newspaper front page that showed a picture of a woman expressing an angry emotion. Lastly, participants in the eight conditions were exposed to a newspaper front page that showed a picture of a man expressing an angry emotion (Table 1).

Table 1: Eight conditions

WOMAN MAN

Not Angry Angry Not Angry Angry Political Context 1 (N=20) 3 (N=20) 2 (N=20) 4 (N=20) Neutral

Context

5 (N=20) 7 (N=20) 6 (N=20) 8 (N=20)

Note: 2 (gender: female vs male) x 2 (emotion: angry vs not angry) x 2 (context: political vs neutral)

between-subjects factorial design.

A newspaper front page was chosen because still the majority of Dutch citizens re-ceive their news information through (online) newspapers (NOM, 2014). Furthermore, com-paring pictures shown in a newspaper with pictures shown on television, when reading a newspaper, readers are exposed to still pictures (Barret & Barrington, 2005). Readers could thus have more attention for the few unchanging pictures that are shown. In contrast, televi-sion watchers could have less attention for each single moving and changing picture that is displayed on television.

Pre-test

Participants were randomly assigned to one of the eight experimental conditions. To assure that participants were not biased regarding the politician covered on the picture, an ordinary couple, that is not involved or bound to politics, was used to act as real politicians. This way it could be excluded that participants would have predispositions regarding the politicians shown on the pictures. A picture was taken of the man and the woman expressing an angry emotion and a not angry emotion. The angry position was determined by two examples (Ap-pendix B).

(14)

A pre-test was conducted to measure whether the stimuli were interpreted the right way, whether the woman and man both when expressing an angry and not angry emotion were interpreted as angry and not angry. Ten respondents were exposed to the four pictures showing the man and woman in the political context and 10 respondents were exposed to the four pictures showing the man and woman in the non-political context (Appendix C, Ques-tionnaire). After being exposed to each of the four pictures respondents were asked to indi-cate on a 7-point Likert scale from (1) totally agree to (7) totally disagree whether they be-lieved the man or woman on the picture expressed an angry emotion.

The pre-test showed no significant results, except for the not angry emotions in the non-political context. The angry female politician (M=4,00, SD=1,56) was evaluated almost as angry as the angry male politician (M=4,90, SD=1,66). This was as well the case for the difference between the not angry female politician (M=1,60, SD=1,07) and the not angry male politician (M=1,80, SD=0,92). For both the angry (M=4,00, SD=1,56) and not angry (M=1,60, SD=1,07) female politician and the angry (M=4,90, SD=1,66) and the not angry (M=1,80, SD=0,92) male politician a difference was shown. When considering the evalua-tion of the four pictures in the non-political context, no difference was shown between the angry woman (M=5,10, SD=1,29) and the angry man (M=4,80, SD=1,6). Only the difference between the not angry woman (M=1,20, SD=0,63) and the not angry man (M=1,80,

SD=1,03) appeared to be significant χ2(3)=10,00, p<0,05. For the angry (M=5,10, SD=1,29)

and the not angry (M=1,20, SD=0,63) woman and the angry (M=4,80, SD=1,60) and the not angry man (M=1,80, SD=1,03) a difference was found (Appendix D, Table 2 & 3).

To conclude, besides the significant differences between the not angry woman and the not angry man in the non-political context, all other stimuli appeared not to be significant. Since the difference between the not angry woman and the not angry man appeared to be lim-ited, it was decided that the stimuli had a sufficient foundation to be used in the experiment.

(15)

Procedure

The data was collected online, through Facebook or e-mail, by means of snowball sampling. The questionnaire has been online for over six days, from the 23rd of May until the 28th of May 2015. In total 160 respondents, ranging from 18 to 85 years old (M=32,73, SD=16,52), are included in this research. When participating in the experiment, firstly the participants had to answer demographical questions concerning their gender, age, and education. Second-ly they were exposed to the stimulus material of one of the eight conditions. ThirdSecond-ly they had to answer questions concerning the stimulus material. As a manipulation check they had to answer a question concerning the content of the article. The manipulation check was followed by a question about the proficiency of the person shown on the picture. By means of the post-test, the dependent variables were measured. The dependent variables measured the compe-tence and credibility of the political candidate. Lastly, they had to answer a question about their political orientation and whether they knew the politician from the parliament (de Tweede Kamer). The political orientation was measured on a 7-point Likert scale from (1) left to (7) right oriented (Appendix C, Questionnaire).

Operationalization of concepts Dependent variables

To measure the effect of an angry politician and the effect of gender, two dependent variables were used: competence and credibility. Credibility is measured by means of three dependent variables: attractiveness, trustworthiness, and expertise.

Competence

Competence was measured by means of the following four items: competent-incompetent and

(16)

These four items were measured on a 7-point Likert scale, from for example extremely com-petent (7) to extremely incomcom-petent (1).

Credibility

There are three factors that are related to the perceived credibility of a source, those three fac-tors are: attractiveness, trustworthiness, and expertise. To measure credibility, the attractive-ness, trustworthiattractive-ness, and expertise of the candidate have been measured.

Attractiveness. According to Ohanian (1990), attractiveness can be measured by

means of the following five items: attractiveness-unattractiveness, classy-not classy,

beauti-ful-ugly, elegant-plain, sexy-not sexy. These items were measured on a 7-point Likert scale,

from extremely attractive (7) to extremely unattractive (1) (Schneider, 2014).

Trustworthiness. According to Ohanian (1990), trustworthiness can be measured by

means of the following five items: dependable-undependable, honest-dishonest,

reliable-unreliable, sincere-insincere, and trustworthy-untrustworthy. The items were measured on a

7-point Likert scale, from for example extremely dependable (7) to extremely undependable (1).

Expertise. According to Ohanian (1990), expertise can be measured by means of the

following five items: expert-not an expert, experienced-inexperienced,

professional-unprofessional, qualified-unqualified, and skilled-unskilled. The items were measured on a

7-point Likert scale, from for example extremely experienced (7) to extremely inexperienced (1).

(17)

To measure the validity of the manifest variables that measured the dependent variables, four factor analyses were conducted. The factor analyses showed that the different manifest varia-bles measured the latent variavaria-bles: competence, attractiveness, trustworthiness, and expertise.

The factor analysis showed that the four items that measured competence, the five items that measured attractiveness, the five items that measured trustworthiness, and the five items that measured expertise existed all of one component. The components were similar to previous research of Tiedens (2001) and Ohanian (1990). The construct-validity could thus be confirmed.

The reliability of the correlations between the different variables was measured by means of the cronbachs alpha (α). The reliability of the scales appeared to be highly reliable for all four components: competence (α=0,80), attractiveness (α=0,85), trustworthiness (α=0,94), and expertise (α=0,96).

Data Analysis

To test whether the descriptive variables: education, political orientation, and the recognition of the politician were related to the eight conditions, a Chi-squared test was conducted for all three variables.

To test for a possible priming effect regarding the allocation of the profession of the man and woman, an independent t-test was conducted. To measure whether there was a sig-nificant difference between the allocation of the profession in the political context and the non-political context, the variable that originally measured the profession of the man and woman by means of five different professions was divided into two groups: (1) politics and (2) other professions.

(18)

To test whether age and political orientation of the respondents correlated with the relation between the gender of the respondents and the evaluation of the man and woman, a bivariate analysis was conducted.

RESULTS

Descriptive Results

Fifty percent of the respondents that participated in the current experiment were women and 50% were men. Most of the respondents enjoyed a University level (WO) of education (54,4%), followed by a higher professional education (HBO) level (29,4%). There appeared to be no significant relation between education level and the different conditions

χ2(35)=22,61, p>0,05.

When comparing the respondents that were exposed to the political context with the respondents that were exposed to the non-political context, 85% (N=68) of the respondents in the political context indicated that the person on the picture was a politician against 7,5% (N=6) that indicated the person as being a professor. The remaining respondents (N=6) indi-cated that the person on the picture concerned a lawyer (N=3), doctor (N=2), or consultant (N=1). Comparing this to respondents in the non-political context, 43,8% (N=35) of the re-spondents in the non-political context indicated that the person on the picture was a politician against 41,3% (N=33) that indicated the person as being a professor. The remaining respond-ents (N=12) indicated that the person on the picture was a consultant (N=5), doctor (N=4), or lawyer (N=3). This remarkable finding might be a result of the introduction text. In the intro-duction it was stated that the questionnaire was conducted to expand knowledge in the field of political communication. Respondents could have been primed in thinking that in the non-political context the man or woman on the front page concerned a politician. However, the independent t-test showed a significant t (158)=-5,10, p<0,05 difference between the political

(19)

context and the non-political context in the expectations of the respondents regarding the pro-fession of the man and woman shown on the picture. Respondents in the political context (M=1,15, SD=0,36) were more likely to answer that the man or woman on the picture con-cerned a politician than respondents in the non-political context (M=1,56, SD=0,50). Priming is thus not definite.

On average the political orientation of the respondents was more right (M=4,40,

SD=1,45). There appeared to be no significant relation between political orientation and the

different conditions χ2(42)=45,85, p>0,05.

On average 15 of the 160 respondents answered that they recognized the politician from the parliament. In the non-political context (N=9) there were more respondents stating they recognized the politician than in the political context (N=6). As no significant result was shown χ2(1)=0,66, >0,05, there appeared to be no relation between the political or non-political context and recognition of the politician.

By means of bivariate analysis test it was measured whether age and political orienta-tion correlated with the relaorienta-tion between the gender of the respondents and the evaluaorienta-tion of the politicians. However, no significant results were found for age r=-0,03, p>0,05 and politi-cal orientation r=-0,07, p>0,05.

Manipulation check

Before testing the hypotheses, first for both the political and non-political context it was measured whether the stimuli were interpreted the right way, whether the man and woman both when expressing an angry and not angry emotion were interpreted as angry and not an-gry. Concerning the angry emotion, both for the political context F(1,38)=3,84, p>0,05 and the non-political context F(1,38)=0,84, p>0,05 the results were not significant. The angry female politician (M=4,55, SD=1,57) and the angry woman (M=4,65, SD=1,46) were

(20)

evalu-ated as angry as the angry male politician (M=5,25, SD=1,16) and the angry man (M=4,20,

SD=1,64). Likewise, concerning the not angry emotion both for the political context F(1,38)=2,56, p>0,05 and the non-political context F(1,38)=0,25, p>0,05 the results were not

significant. The not angry female politician (M=2,65, SD=1,31) and the not angry woman (M=2,35, SD=1,18) were evaluated similar as the not angry male politician (M=3,50,

SD=1,43) and the not angry man (M=2,55, SD=1,36) regarding the emotion they were

ex-pressing.

In addition it was tested whether the angry female differed in her expression from the not angry female and the angry man differed in his expression from the not angry man. Con-cerning the woman, for both the political context F(1,38)=17,26, p<0,05 and the non-political context F(1,38)=29,96, p<0,05 the difference appeared to be significant. The angry female politician (M=4,55, SD=1,57) and the angry woman (M=4,65, SD=1,46) differed from the not angry female politician (M=2,65, SD=1,31) and the not angry woman (M=2,35,

SD=1,18). Likewise, concerning the man in both the political context F(1,38)=17,97, p<0,05

and the non-political context F(1,38)=12,01, p<0,05 the results were significant. The angry male politician (M=5,25, SD=1,16) and the angry man (M=4,20, SD=1,64) differed from the not angry male politician (M=3,50, SD=1,43) and the not angry man (M=2,55, SD=1,36) (Appendix E, Table 4 & 5).

Examining Hypothesis 1

The first hypothesis predicted that, in comparison to a non-political context, respondents would respond different to a male than to a female politician. It was predicted that they would evaluate the male politician as more competent and more credible.

To examine whether the gender of the politician has an effect on the evaluation of competence and credibility, a univariate analysis of variance (ANOVA) was conducted. The

(21)

ANOVA test indicated that taken together for the political context as well as the non-political context the results of competence F(7,152)=7,75, p<0,05 and trustworthiness F(7,152)=5,92,

p<0,05 appeared to be significant. The not angry (M=4,68, SD=1,06) and angry (M=3,96, SD=0,81) female politician appeared to be more competent than the not angry (M=4,25, SD=0,86) and angry (M=3,74, SD=0,97) male politician. Likewise, the not angry (M=5,05, SD=0,88) and angry (M=4,49, SD=1,03) female politician appeared to be more trustworthy

than the not angry (M=4,39, SD=1,04) and angry (M=3,50, SD=1,30) male politician. The ANOVA indicated that no significant result existed for both attractiveness and expertise.

Remarkable is that it appeared that the not angry (M=5,24, SD=0,93) and angry (M=4,44, SD=0,65) woman and the not angry (M=5,01, SD=0,86) and angry (M=3,75,

SD=1,04) man in the non-political context were evaluated as more competent than the

wom-an wom-and the mwom-an in the political context. Furthermore, the not wom-angry (M=5,12, SD=0,97) wom-and angry (M=4,85, SD=0,83) woman and the not angry (M=5,08, SD=0,89) and angry (M=4,46, SD=0,95) man in the non-political context appeared to be more trustworthy than the man and woman in the political context (Table 6-9).

Hypothesis 1 could not be accepted. Only for competence and trustworthiness a sig-nificant result was found, the female politician was evaluated as more competent and more trustworthy than the male politician. When comparing these results to the findings of the non-political context, the man and woman in the non-non-political context were evaluated as more competent and trustworthy than in the political context.

Table 6: Means and standard deviations of competence

CONTEXT MAN WOMEN

Angry Not Angry Angry Not Angry

Political 3,74 (0,97) 4,25 (0,86) 3,96 (0,81) 4,68 (1,06) Not Political 3,75 (1,04) 5,01 (0,86) 4,44 (0,65) 5,24 (0,93)

(22)

Table 7: Means and standard deviations of attractiveness

CONTEXT MAN WOMEN

Political 3,34 (0,92) 3,39 (0,94) 3,55 (0,10) 4,00 (0.96) Not Political 3,63 (0,80) 3,94 (0,87) 3,63 (0,84) 4,03 (0,66)

Note. N=160 and F(7,152)=1,90, p=0,07

Table 8: Means and standard deviations of trustworthiness

CONTEXT MAN WOMEN

Angry Not Angry Angry Not Angry Political 3,50 (1,30) 4,39 (1,04) 4,49 (1,03) 5,05 (0,88) Not Political 4,46 (0,95) 5,08 (0,89) 4,85 (0,83) 5,12 (0,97)

Note. N=160 and F(7,152)=5,92, p=0,00

Table 9: Means and standard deviations of expertise

CONTEXT MAN WOMEN

Political 4,29 (1,47) 4,88 (1,25) 4,87 (1,24) 5,24 (0,86) Not Political 4,95 (1,07) 4,23 (0,86) 5,16 (0,65) 5,33 (1,07)

Note. N=160 and F(7,152)=1,89, p=0,08

The second hypothesis predicted that, in comparison to a non-political context, respondents would evaluate a male politician expressing an angry emotion as more competent and more credible than a female politician expressing an angry emotion.

To test whether expressing an angry emotion has an effect on the evaluation of com-petence and credibility of a politician, an ANOVA was conducted. The ANOVA test indicat-ed that in general the angry woman and angry man were significantly F(1,78)=7,26, p<0,05 evaluated as less trustworthy in the political context (M=4,00, SD=1,26) than in the non-political context (M=4,66, SD=0,90) (Table 10). When taking the gender in consideration, both in the political context as well as in the non-political context the angry female politician

(23)

petent F(3,76)=2,77, p<0,05 than the angry male politician (M=3,74, SD=0,97) and the an-gry man (M=3,75, SD=1,04). Likewise, the anan-gry (M=4,49, SD=1,03) female politician and the angry woman (M=4,85, SD=0,83) appeared to be more trustworthy F(3,76)=6,17, p<0,05 than the angry male politician (M=3,50, SD=1,30) and the angry man (M=4,46, SD=0,95). The ANOVA indicated that there was no significant result for both attractiveness and exper-tise.

Again, it is remarkable that the angry woman and the angry man in the non-political context were evaluated as more competent and more trustworthy than the angry woman and angry man in the political context (Table 11-14).

In general, when not correcting for the gender of the politicians, it appeared that re-spondents evaluated the man and woman in the political context as less trustworthy than in the non-political context. When controlling for the gender of the politicians, only for the evaluation of competence and trustworthiness a significant result was shown, in that the an-gry female politician was evaluated as more competent and more trustworthy than the anan-gry male politician. In addition, hypothesis 2 could not be accepted.

Table 10: Means and standard deviations of competence, attractiveness, trustworthiness, and expertise

CONTEXT Competence Attractiveness Trustworthiness Expertise Angry Emotion Angry Emotion Angry Emotion Angry Emotion

Political 3,85 (0,87)* 3,45 (0,95)** 4,00 (1,26)*** 4,58 (1,37)**** Not Political 4,09 (0,93)* 3,63 (0,81)** 4,66 (0,90)*** 5,06 (0,88)****

Note. N=80 and *F(1,78)=1,45, p=0,23, ** F(1,78)=0,87, p=0,35, *** F(1,78)=7,26, p=0,01, **** F(1,78)=3,39, p=0,07

CONTEXT MAN WOMEN

Angry Angry Political 3,74 (0,97) 3,96 (0,81) Not Political 3,75 (1,04) 4,44 (0,65)

(24)

CONTEXT MAN WOMEN

Angry Angry

Political 3,34 (0,92) 3,55 (0,10) Not Political 3,63 (0,84) 3,63 (0,80)

Note. N=80 and F(3,76)=0,47, p=0,70

CONTEXT MAN WOMEN

Angry Angry Political 3,50 (1,30) 4,49 (1,03) Not Political 4,46 (0,95) 4,85 (0,83)

Note. N=80 and F(3,76)=6,17, p=0,00

CONTEXT MAN WOMEN

Angry Angry

Political 4,29 (1,47) 4,87 (1,24) Not Political 4,95 (1,07) 5,16 (0,65)

Note. N=80 and F(3,76)=2,105 p=0,107

The third hypothesis predicted that, in comparison to a non-political context, male respond-ents would evaluate a male politician as more competent or credible than a female politician. For female politicians this gender-based voting effect would be less strong. In other words, the hypothesis predicts that a moderating effect of gender of the respondents will occur and will be stronger for male respondents than for female respondents.

To test for the moderating effect, an ANCOVA was conducted. By means of the AN-COVA it could be tested whether an interaction-effect would occur between the independent variable, the not angry condition, and the moderating variable, the gender of the respondents. An interaction effect will occur when the evaluation of competence and credibility of the not angry male or female politician is dependent upon the gender of the respondents.

(25)

The ACNOVA test for hypothesis 3 indicated that for both the political context as the non-political context there was no significant interaction effect found for competence attrac-tiveness, trustworthiness, and expertise (Table 15-18). Hypothesis 3 could thus not be accept-ed.

GENDER & EMOTION: AN-GRY MAN NOT ANGRY MAN ANGRY WOMAN NOT ANGRY WOMAN RESPOND-ENTS:

Men Women Men Women Men Women Men Women

Political 3,73 (0,95) 3,75 (1,03) 4,25 (0,54) 4,25 (1,13) 3,93 (0,10) 4,00 (0,76) 4,15 (0,85) 5,20 (1,03) Not Political 3,90 (0,96) 3,60 (1,15) 5,08 (0,99) 4,95 (0,76) 4,67 (0,76) 4,20 (0,44) 5,53 (0,80) 4,95 (1,01) Note. N=160 and F(15,144)=1,56, p=0,15

GENDER & EMOTION: ANGRY MAN NOT ANGRY MAN ANGRY WOMAN NOT ANGRY WOMAN RESPOND-ENTS:

(26)

Table 17: Means and standard deviations of trustworthiness GENDER & EMOTION: ANGRY MAN NOT ANGRY MAN ANGRY WOMAN NOT ANGRY WOMAN RESPOND-ENTS:

GENDER & EMOTION: ANGRY MAN NOT ANGRY MAN ANGRY WOMAN NOT ANGRY WOMAN RESPOND-ENTS:

Political 4,22 (1,52) 4,36 (1,50) 4,74 (0,91) 5,02 (1,57) 4,68 (1,46) 5,06 (1,01) 4,94 (0,91) 5,54 (0,72) Not Political 4,98 (1,31) 4,92 (0,84) 5,12 (0,81) 5,34 (0,94) 5,18 (0,54) 5,14 (0,78) 5,46 (0,80) 5,20 (1,33) Note. N=160 and F(15,144)=0,31, p=0,95 Examining Hypothesis 4

The last hypothesis predicted that, in comparison to a non-political context, female respond-ents would show a stronger, more positive response towards a male or female politician ex-pressing an angry emotion than male respondents. Again, it is predicted that a moderating effect of the gender of the respondents will occur and will be stronger for female respondents than for male respondents.

To test for the moderation affect, an ANCOVA was conducted. By means of the AN-COVA it could be tested whether an interaction-effect would occur between the independent

(27)

variable, the angry condition, and the moderating variable, the gender of the respondents. An interaction effect will occur when the evaluation of competence and credibility of the angry male or female politician is dependent upon the gender of the respondents.

The ANCOVA test for hypothesis 4 did not show any significant interaction effect be-tween gender of the respondents and competence, attractiveness, trustworthiness, and exper-tise of the woman and man (Table 19-22). In addition, hypothesis 4 could thus not be accept-ed.

GENDER & EMOTION: ANGRY MAN ANGRY WOMAN

RESPONDENTS: Men Women Men Women

Political 3,73 (0,95) 3,75 (1,03) 3,93 (0,10) 4,00 (0,76) Not Political 3,90 (0,96) 3,60 (1,15) 4,67 (0,76) 4,20 (0,44) Note. N=80 and F(7,72)=0,44, p=0,73

Political 3,44 (0,85) 3,24 (1,02) 3,04 (1,01) 4,06 (0,71) Not Political 3,52 (0,83) 3,74 (0.81) 3,58 (1,02) 4,68 (0,67) Note. N=80 and F(7,72)=1,78, p=0,16

(28)

Political 3,42 (1,17) 3,58 (1,48) 4,34 (1,16) 4,64 (0,93) Not Political 4,34 (1,21) 4,58 (0,621) 5,10 (0,76) 4,60 (0,85) Note. N=80 and F(7,72)=0,62, p=0,61

Political 4,22 (1,52) 4,36 (1,50) 4,68 (1,46) 5,06 (1,01) Not Political 4,98 (1,31) 4,92 (0,84) 5,18 (0,54) 5,14 (0,78) Note. N=80 and F(7,72)=0,15, p=0,93 DISCUSSION Conclusion

By means of four hypotheses it was predicted that in general respondents would evaluate the male politician as more competent and more credible than the female politician. Even when comparing a male politician expressing an angry emotion with a female politician expressing an angry emotion, it was predicted that the male politician would be evaluated as more com-petent and credible. Furthermore it was assumed that a moderating effect would exist, where-by the gender of the respondents would be crucial in the evaluation of the not angry or angry politicians.

According to the results all four hypotheses had to be rejected. Altogether, it could be argued that only for the evaluation of competence and trustworthiness the gender of a politi-cian is crucial. Against expectations, only a significant effect was found for the evaluation of

(29)

competence and trustworthiness regarding the female and male politician. No significant ef-fect was found for the evaluation of attractiveness or expertise. Moreover, no significant in-teraction effect was found between the gender of the respondents and the evaluation of com-petence, attractiveness, trustworthiness, and expertise of an angry or not angry female or male politician. However, based on the significant results it can be concluded that in general the woman is evaluated as more competent and trustworthy, regardless whether she expressed an angry or not angry emotion.

Although not all results are significant, the results hint to three general trends. First, it appeared that the woman in general is evaluated more positive on competence and credibility than the man. Second, in general the evaluation of competence and credibility is more posi-tive in the non-political context than in the political context. To illustrate, in the non-political context the (angry) woman and (angry) man were evaluated as more competent and more trustworthy than in the political context. A third trend that appeared concerns the gender based voting effect. The pattern shows that gender based voting would be more likely to oc-cur for female respondents. Female respondents evaluated the woman in both the political and the non-political context in most of the cases more positive than they evaluated the man. This positive relation between gender of the respondent and gender of the person shown on the picture was less evident for the male respondents.

According to the literature a “baseline gender preference” exists in that the gender of candidates is decisive in the voters’ evaluation of the competence, credibility, and policy po-sition of the candidate (Dolan, 2010; Sanbonmatsu, 2002). It was argued that female candi-dates would be disadvantaged in this baseline gender preference (Erzeel & Caluwaerts, 2015). In a professional context an angry man would be evaluated as more competent, and would thus enjoy an increase in social status, than an angry woman, who would enjoy a de-crease in social status (Tiends, 2001; Brescoll & Uhlmann, 2008). Furthermore, concerning

(30)

the gender of the voters, Holli and Wass (2010) argued that voters prefer candidates of the same gender (i.e. gender based voting). In addition, men should have a higher tendency to choose a candidate of their own gender than women. In contrast, women should react less negative when being exposed to a negative picture of a man or woman than men (Barret & Barrington, 2005).

By means of the current experiment it could be concluded that female politicians are not disadvantaged when it concerns the evaluation of their competence and credibility. Unex-pectedly male politicians appeared to be rather disadvantaged compared to female politicians. However in general the man and woman have a greater possibility to be evaluated more posi-tively in a not political context. Furthermore, women rather than men are expected to show a gender based voting effect.

Limitations and Future research

A few limitations that might explain the different results than expected could be related to this research. The finding that the angry female was evaluated as more competent than the angry male is contradicting to the findings of Brescoll and Uhlmann (2008). They present evidence for the fact that respondents, that were exposed to an angry man, viewed the man as more competent than the angry woman. An explanation could be that the female politician had possibly a stronger appearance, regarding to the reactions of the respondents here expres-sion was perceived as being more convincing and powerful. The ptest in the current re-search only tested for the angry emotion that was shown. The difference in strength of ap-pearance has not been pre-tested since that was not the goal of the research. Future research should control for the “strength of appearance” by conducting a pre-test.

Another limitation is the fact that, as argued before, respondents might have been primed by the introduction text of the questionnaire in which is stated that the goal of the

(31)

re-search is to expand the knowledge in the field of political communication. The fact that polit-ical communication was named could have directed the thoughts on the profession of the per-son on the picture towards politics because they were aware that the goal of the research had something to do with politics. Future research should prevent priming by not referring to politics before respondents are exposed to the stimuli.

The current study did not show a relation between gender, education level, political orientation, or age and the evaluation of the competence and credibility of the politician. For future research it might be interesting to see whether there are other underlying mechanisms like political knowledge or political interest (Erzeel & Caluwaerts, 2015; Holli & Wass, 2010) that correlate with the evaluation of female or male politicians.

Two exceptional results are found that might be interesting to be further explored in future research. The first exceptional finding is that compared to the political context, in the non-political context the man and woman were evaluated more positive. This finding could signify that people are more distrustful when it comes to politics. This could be emphasized by the finding that when not considering the gender of the politicians, it appeared that re-spondents evaluated the man and woman in the political context as less trustworthy than in the non-political context. Another interesting finding refers to the gender based voting effect. The results hint to the fact that the gender based voting effect is potentially more existing in a political context than in a non-political context. This could indicate that concerning politics people feel more dedicated to their own gender. However, since this finding is not significant it should be further explored in future research.

The current study has offered insight into the differences in evaluation of competence and credibility concerning female or male politicians. It appears that the female politician is seen as more competent and trustworthy than the male politician. These findings are in con-trast to the findings of previous research and thus offer an interesting basis for a future study.

(32)

Furthermore, as the current study shows evidence for the fact that respondents react different to a political context than to a non-political context it could have theoretical implications for the academic literature. Concluding, according to the findings, a female-quota is worthwhile considering since women are taken as serious as man, or even more serious when it comes to competence and trustworthiness. Women do not have to feel insecure when it comes to lead-ing positions. However, more research is needed to confirm the findlead-ings of the current re-search.

(33)

References

Anderson, C., John, O.P., Keltner, D., & Kring, A.M. (2008). Who attains social status? Ef-fects of personality and physical attractiveness in social groups. Journal of Personality and Social Psychology, 81(1), 116-132. DOI: 10.1037//0022-3514.81.1.116.

Barret, A. W., & Barrington, L.W. (2005). Is a picture worth a thousand words? Newspaper photographs and voter evaluations of political candidates. Press/Politics, 10(4), 98-113. DOI: 10.1177/1081180X05281392.

Brescoll, V.L., & Uhlmann, E.L. (2008). Can an Angry Woman Get Ahead? Status Conferral, Gender, and Expression of Emotion in the Workplace. Psychological Science, 19(3), 268–75. DOI: 10.1111/j.1467-9280.2008.02079.x.

Broos, D.J. (2011). Testing the double standard for candidate emotionality: Voter reactions to the tears and anger of male and female politicians. The Journal of Politics, 73(2), 597-615. DOI:10.1017/S0022381611000053.

Bussemaker dreigt met wet vrouwenquota. (12 September 2014). Financieel Dagblad. Re trieved 20 June 2015, from.

Dolan, K. (2010). The impact of gender stereotyped evaluations on support for women candidates. Political Behavior, 32(1), 69-88. DOI: 10.1007/sl 1109-009-9090-.

Erzeel, S., & Caluwaerts, D. (2015). Is It Gender, Ideology or Resources? Individual-Level Determinants of Preferential Voting for Male or Female Candidates. Journal of

Elections, Public Opinion and Parties. DOI: 10.1080/17457289.2015.1008495.

Grossman, M., & Wood, W. (1993). Sex Differences in Intensity of Emotional Experience: A Social Role Interpretation. Journal of Personality and Social Psychology, 65, 1010– 22.

Harell, T.M.A., & Gosseling, T. (2014). Cuing the gap: gender and psychological orientations to politics. Working Paper.

(34)

Heilman, M.E. (2001). Description and prescription: How gender stereotypes prevent wom en’s ascent up the organizational ladder. Journal of Social Issues, 57, 657–674. DOI: 10.1111/0022-4537.00234.

Hillary Clintons sekse blijkt haar grootste troef. Als oma kan ze een presidentsverkiezing winnen. (15 april 2015). Vrij Nederland. Retrieved on 1 May 2015, from http://www.vn.nl/Archief/Buitenland/Artikel-Buitenland/Oma-in-het-Witte-Huis-.htm.

Hinsz, V.B., & Tomhave, J.A. (1991). Smile and (Half) the World Smiles with You, Frown and You Frown Alone. Personality and Social Psychology Bulletin, 17, 586–92. DOI: 10.1177/0146167291175014.

Holli, A.M., & Wass, H. (2010). Gender-based voting in the parliamentary elections of 2007 in Finland. European Journal of Political Research, 49, 598-630. DOI: 10.1111/j.1475-6765.2009.01910.x

Lecheler, S., Schuck, A., & de Vreese, C.H. (2013). Dealing with feelings: Positive and negative discrete emotions as mediators of news framing effects. Communications,

38(2), 189-209. DOI: 10.1515/commun-2013-0011.

NOM Print Monitor. (2014). Retrieved 2 June 2015, from

www.nommedia.nl/printmonitor.html.

Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers’ perceived expertise, trustworthiness, and attitude. Journal of Advertising, 19(3), 39-52.

Plant, E.A., Hyde, J.S., Keltner, D., & Devine, P.G. (2000). ‘‘The Gender Stereotyping of Emotions.’’ Psychology of Women Quarterly, 24, 81–92. DOI: 10.1111/j.1471-6402.2000.tb01024.x.

(35)

Rudman, L.A. (1998). Self-promotion as a risk factor for women: The costs and benefits of counterstereotypical impression management. Journal of Personality and Social Psy

chology, 74, 629–645.

Sanbonmatsu, K. (2002). Gender Stereotypes and Vote Choice. American Journal of Political

Science 46(1), 20–34. DOI: 10.2307/3088412.

Schneider, M.C. (2014). The effects of gender-bending on candidate evaluations. Journal of

Women, Politics & Policym, 35, 55-77. DOI: 10.1080/1554477X.2014.863697.

The Rules for Female Candidates. (7 November 2007). Washington Post. Retrieved 25 April

2015, from

http://www.washingtonpost.com/wp-yn/content/article/2007/11/06/AR2007110602489.html.

Tiedens, L.Z. (2001). Anger and advancement versus sadness and subjugation: The effect of negative emotion expressions on social status conferral. Journal of Personality and

Social Psychology, 80, 86–94.

Todorov, A., Mandisodza, A.N., Goren, A., & Hall, C.C. (2005). Inferences of Competence from Faces Predict Election Outcomes. Science, 308(16), 23-26. DOI:10.1126/science.1110589.

Women in National Parliament. (1 April 2015). Inter-Parliamentary Union (IPU). Retrieved 12 May 2015, from http://www.ipu.org/wmn-e/classif.htm.

(36)

Appendixes

Appendix A

(37)

(38)

(39)

(40)

Appendix B

Examples of an angry expression

Appendix C

Questionnaires

1.Pre-test

Beste deelnemer,

Bedankt voor uw deelname aan dit onderzoek. U wordt allereerst gevraagd een drietal demografische vragen te beantwoorden. Vervolgens wordt u blootgesteld aan een voorpagina van het NRC Handels-blad. Ik wil u vragen de voorpagina zorgvuldig te bekijken alvorens u de vraag met betrekking tot de voorpagina beantwoordt. Het is belangrijk om te weten dat er geen goede en foute antwoorden be-staan. Het doel van dit onderzoek is om meer inzicht te krijgen op het gebied van politieke communi-catie.

Alvorens u de vragenlijst invult, dient u onderstaande tekst aandachtig door te lezen.

 Ik verklaar hierbij op voor mij duidelijke wijze te zijn ingelicht over de aard en methode van dit onderzoek.

 Ik stem geheel vrijwillig in met deelname aan dit onderzoek en ik behoud daarbij het recht deze instemming weer in te trekken zonder dat ik daarvoor een reden hoef op te geven.

 Ik besef dat ik op elk moment mag stoppen met het experiment.

 Als mijn onderzoeksresultaten gebruikt worden in wetenschappelijke publicaties, of op een andere manier openbaar worden gemaakt, dan zal dit volledig geanonimiseerd gebeuren.

 Mijn persoonsgegevens worden niet door derden ingezien zonder mijn uitdrukkelijke toe-stemming.

(41)

(lynnemore.vanommen@student.uva.nl).

 Voor eventuele klachten over dit onderzoek kan ik contact opnemen met het lid van de Com-missie Ethiek van de afdeling Communicatiewetenschap, per adres: ASCoR secretariaat, Commissie Ethiek, Universiteit van Amsterdam, Postbus 15793, 1001 NG Amsterdam; 020‐ 525 3680; ascor‐ secr‐ fmg@uva.nl.

1. Ik stem hierbij in aan dit onderzoek deel te nemen en op de hoogte te zijn van de aard van het onderzoek.

o Ja o Nee 2. Wat is uw geslacht? o man o vrouw 3. Wat is uw leeftijd? o ____

4. Wat is uw hoogst genoten opleiding?

o Mavo/ vmbo o Havo o Vwo o Mbo o Hbo o Wo

5. De persoon op de foto uit een boze emotie.

Geheel mee eens 0 0 0 0 0 0 0 Geheel mee oneens

6. U bent aan het einde van de enquête gekomen. Hartelijk dan voor uw medewerking!

(42)

Klik op ‘volgende’ om uw antwoorden op te slaan en te verzenden.

2.Experiment

Beste deelnemer,

Bedankt voor uw deelname aan dit onderzoek. U wordt allereerst gevraagd een drietal demografische vragen te beantwoorden. Vervolgens wordt u blootgesteld aan een voorpagina van het NRC Handels-blad. Ik wil u vragen de voorpagina zorgvuldig te bekijken alvorens u de vragen met betrekking tot de voorpagina beantwoordt. Het is belangrijk om te weten dat er geen goede en foute antwoorden be-staan. Het doel van dit onderzoek is om meer inzicht te krijgen op het gebied van politieke communi-catie.

Alvorens u de vragenlijst invult, dient u onderstaande tekst aandachtig door te lezen.

 Ik verklaar hierbij op voor mij duidelijke wijze te zijn ingelicht over de aard en methode van dit onderzoek.

 Ik stem geheel vrijwillig in met deelname aan dit onderzoek en ik behoud daarbij het recht deze instemming weer in te trekken zonder dat ik daarvoor een reden hoef op te geven.

 Ik besef dat ik op elk moment mag stoppen met het experiment.

 Als mijn onderzoeksresultaten gebruikt worden in wetenschappelijke publicaties, of op een andere manier openbaar worden gemaakt, dan zal dit volledig geanonimiseerd gebeuren.

 Mijn persoonsgegevens worden niet door derden ingezien zonder mijn uitdrukkelijke toe-stemming.

 Mijn vragen zijn naar tevredenheid beantwoord.

 Voor meer informatie, nu of in de toekomst, kan ik mij wenden tot Lynnemore van Ommen (lynnemore.vanommen@student.uva.nl).

 Voor eventuele klachten over dit onderzoek kan ik contact opnemen met het lid van de Com-missie Ethiek van de afdeling Communicatiewetenschap, per adres: ASCoR secretariaat, Commissie Ethiek, Universiteit van Amsterdam, Postbus 15793, 1001 NG Amsterdam; 020‐ 525 3680; ascor‐ secr‐ fmg@uva.nl.

1. Ik stem hierbij in aan dit onderzoek deel te nemen en op de hoogte te zijn van de aard van het onderzoek.

o Ja o Nee 2. Wat is uw geslacht? o man o vrouw 3. Wat is uw leeftijd? o ____

(43)

o Mavo/ vmbo o Havo o Vwo o Mbo o Hbo o Wo 5. Stimuli Voorbeeld:

6. Wat was er afgebeeld op de foto?

o Een Man o Een Vrouw o Er was geen foto

6. Wat was volgens u het beroep van de persoon op de foto?

o Hoogleraar o Advocaat o Arts o Politicus o Consultant

7. De persoon op de foto uit een boze emotie.

Geheel mee eens 0 0 0 0 0 0 0 Geheel mee oneens 8. De persoon op de foto is…..

Competent 0 0 0 0 0 0 0 Incompetent

Intelligent 0 0 0 0 0 0 0 Onintelligent

(44)

Sympathiek 0 0 0 0 0 0 0 Onsympathiek 9. De persoon op de foto is…..

Aantrekkelijk 0 0 0 0 0 0 0 Onaantrekkelijk

Stijlvol 0 0 0 0 0 0 0 Niet stijlvol

Mooi 0 0 0 0 0 0 0 Lelijk

Elegant 0 0 0 0 0 0 0 Niet elegant

Sexy 0 0 0 0 0 0 0 Niet sexy 10. De persoon op de foto is…..

Betrouwbaar 0 0 0 0 0 0 0 Onbetrouwbaar

Eerlijk 0 0 0 0 0 0 0 Niet eerlijk

Zeker/Stabiel 0 0 0 0 0 0 0 Onzeker/Onstabiel

Oprecht 0 0 0 0 0 0 0 Niet oprecht

Geloofwaardig 0 0 0 0 0 0 0 Ongeloofwaardig 11. De persoon op de foto is…..

Expert 0 0 0 0 0 0 0 Geen expert

Ervaren 0 0 0 0 0 0 0 Onervaren

Deskundig 0 0 0 0 0 0 0 Niet deskundig

Gekwalificeerd 0 0 0 0 0 0 0 Niet gekwalificeerd

Vaardig 0 0 0 0 0 0 0 Niet vaardig

12. Hoe zou u uw politieke oriëntatie omschrijven?

Links 0 0 0 0 0 0 0 Rechts

13. Herkende u de politicus op de foto uit de Tweede Kamer?

o Ja o Nee

14. U bent aan het einde van de enquête gekomen. Hartelijk dan voor uw medewerking!

(45)

Appendix D

1.Pre-test

Table 2: Political context

TYPE MEAN SD

Angry Female 4,00 1,56

Angry Male 4,90 1,66

Note. N=10 and χ2(15)=12,00, p=0,68

TYPE MEAN SD

Not Angry Female 1,60 1,07 Not Angry Male 1,80 0,92

Note. N=10 and χ2(6)=12,39, p=0,06

TYPE MEAN SD

Angry Female 4,00 1,56 Not Angry Female 1,60 1,07

Note. N=10 and χ2(15)=20,00, p=0,17

TYPE MEAN SD

Not Angry Male 1,80 0,92

Note. N=10 and χ2(6)=6,71, p=0,35 Table 3: Non-political context

TYPE MEAN SD

Angry Female 5,10 1,29

Note. N=10 and χ2(15)=15,28, p=0,43

TYPE MEAN SD

Not Angry Female 1,20 0,63 Not Angry Male 1,80 1,03

Note. N=10 and χ2(3)=10,00, p=0,02

TYPE MEAN SD

Angry Female 5,10 1,29 Not Angry Female 1,20 0,63

Note. N=10 and χ2(3)=1,67, p=0,64

TYPE MEAN SD

Not Angry Male 1,80 1,03