The Effect of Gender Stereotyped framing and facial traits on support for female politicians

(1)

The Effect of Gender Stereotyped Framing and Facial Traits on Support for Female Politicians.

Judith Meijer | 10582053

Master Thesis Political Communication Communication Science

University of Amsterdam Supervisor: Dr. Bert Bakker

(2)

Abstract

Female politicians are vastly underrepresented in all forms of government, all over the world. This study tries to uncover whether and how this happens through media framing and facial traits. Specifically, I investigate to what extend and for whom gender stereotypical frames and gender stereotypical facial traits influence support for female politicians. Based on previous research, it was expected that a masculine frame and feminine face would yield the most positive attitudes and voting intentions in Dutch citizens. In addition, the moderating role of social dominance orientation, gender identity and sexism were examined to see whether these individual differences moderate this effect. To test the research question, an experiment was conducted amongst Dutch citizens. The results showed a positive effect of feminine facial traits on citizens’ attitude towards the politician. No effect was found for gender-stereotyped frames on support for female politicians. Moderating effects of social dominance orientation and sexism were found on citizens’ voting intention and attitude. No moderating effects were found for gender identity. This study helps to improve our broader understanding of gender stereotypes in political campaigns and could help battle the underrepresentation of female politicians in government.

Keywords: gender stereotypes, female politicians, social dominance

(3)

The Effect of Gender Stereotyped Framing and Facial Traits on Support for Female Politicians.

During the last ten years, more women have filled prominent political roles. Angela Merkel, Theresa May and Hillary Clinton are examples of this. But although the overall percentage of female politicians has increased, women are still vastly underrepresented in all forms of government, all over the world. As of March 2016, only 55 out of 150 members of Dutch Parliament are female (Tweede Kamer, 2016). The interaction between gender and political leadership thus deserves more attention. I theorize that the reason for this underrepresentation of female politicians is due to their facial traits and media framing.

To understand how framing can affect the underrepresentation of female politicians, I turned to gender stereotyping. The central idea of framing is “the selection, organization and emphasis of certain aspects of reality, to the exclusion of others” (de Vreese, Peter, & Semetko, 2001, p. 108). Gender stereotypes are defined as a “structured set of beliefs about personal attributes of women and men” (Ashmore and Del Boca 1979:222). Studies have shown that gender stereotypes are often used by citizens, in news coverage, as well as in election campaigns (Hayes & Lawless, 2015; Kahn, 1992). Among other things, feminine gender stereotyping is

characterised by compassion and empathy traits (Hayes, 2005). Likewise,

aggressiveness and competitiveness are examples of masculine gender stereotypes (Devere & Davies, 2006; Hayes, 2005). Most research on gender stereotypes focuses on identifying the different kinds of stereotypes people hold about men and women in politics (Bauer, 2015; Kahn, 1992; Little, Roberts, Jones, & Burriss, 2007). However, knowledge on the influence of stereotyped campaigns on political attitudes and voting intention is limited (Dolan, 2010). I expand the literature by examining whether it

(4)

would be beneficial for a female politician to incorporate gender stereotyped framing into her campaign.

How a politician presents herself or is represented by the media can be built quite easily. However, some characteristics of a person are more difficult to frame, such as the facial trait of a politician. Multiple studies show that feminine facial traits in male politicians positively influence voting intention (Daniel, DeBruine, Jones, & Perrett, 2013; Little et al., 2007; Spisak, Dekker, Krüger, & van Vugt, 2012). This indicates that facial traits play an important role in gaining voter support. However, not much is known about the effect of female politicians with manipulated masculine or feminine facial features. A study that did focus on female politicians found that the more feminine looking one was favoured more compared to the more masculine looking politician (Hehman, Carpinella, Johnson, Leitner, & Freeman, 2014). So although female politicians can act like men and bring masculine frames forward in their campaigns, if they are born with more or less feminine facial features, it would be very hard to change their face. This shows that female politicians ‘face’ a

predicament. It seems that in order to be most successful, they have to come across masculine in their framing style, whilst looking feminine in their facial traits.

To my best knowledge, there has been no previous research that combines gendered framing and gendered facial traits in order to see what yields the most support for female politicians. Accordingly, I will study the following research question:

RQ: To what extend and for whom do gender stereotypical frames and gender stereotypical facial traits influence support for female politicians?

Based on communication science research, the success of a campaign ad can be based on the voters’ attitude and voting intention towards a politician (Hitchon,

(5)

Chang, & Harris, 1997). Furthermore, the intention to vote seems to be missing from current research. Especially in a democratic society like the Netherlands, intention to vote is very important to politicians since they are elected based on the popular vote. Thus, making these variables scientifically relevant as well as societally.

Finally, this study is not only relevant to political communication research, but also the political communication profession. Spin-doctors, campaign teams and politicians themselves could use this research to find out whether masculine or feminine stereotyping would yield the most support for female politicians. In doing so, the current underrepresentation of women in politics could be battled. I will elaborate more on gender stereotypes in politics, gendered framing and facial traits in the theoretical framework. Lastly, the method and results of the experiment will be discussed and the thesis ends with the conclusion and discussion.

Theoretical framework Distinguishing gender from sex

It seems important to start with distinguishing sex from gender, as the majority of political communication research has failed to do so in the past and these terms are often interchanged in everyday life. In most social sciences, sex is defined as the biological differences between males and females (West & Zimmerman, 1987). Therefore, sex is difficult to change. Gender on the other hand describes the

characteristics that a society creates as masculine or feminine (Lippa, 2005). Because gender is constructed, this means that it can change over time and across different cultures, being more flexible than sex. For example, a female politician can look more masculine as opposed to feminine, making people associate her more with masculine stereotypes. Moreover, the media can easily frame gender, but sex is not something

(6)

that can be influenced by the media. Lastly, gender can develop independently of sex, which leads some women to identify more with the masculine gender whilst

biologically being a woman (McDermott & Hatemi, 2011).

It is important to differentiate gender from sex because even if a politician’s sex is female, this does not mean that she also has to look and act feminine.

Moreover, people can attribute different stereotypes to a woman if her gender is not in line with her sex. It seems interesting to see what happens to people’s evaluations when a female politician’s gender either compliments or counters her own sex. I will elaborate more on these gender stereotypes in the next section.

Gender stereotypes in politics

The central idea of framing is “the selection, organization and emphasis of certain aspects of reality, to the exclusion of others” (de Vreese, Peter, & Semetko, 2001, p. 108). Gender stereotypes could serve as a frame and are defined as a

“structured set of beliefs about personal attributes of women and men” (Ashmore and Del Boca 1979:222). Voters use these gender stereotypes as a form of shallow

decision heuristics to evaluate politicians. This is stated to have a direct impact on the evaluations people make about politicians (Dolan, 2014; Mcdermott, 1998). For example, studies have shown that citizens overall find female politicians to be more compassionate and warm than male politicians (Alexander & Andersen, 1993; Huddy & Terkildsen, 1993; Koch, 1999). In gender research, it seems that there are two ways of conceptualizing gender stereotypes. Voters tend to use either issue-ownership or personality traits to evaluate politicians based on their gender.

Herrnson et al. (2003) developed the concept of “gender ownership” of issues, meaning that voters believe that male and female politicians are better at handling different issues. Due to the female gender stereotype of being caretakers, citizens

(7)

believe female politicians are better at handling “feminine issues” like health care, poverty and education. In contrast, due to the male stereotype of being a protector and breadwinner, male politicians are evaluated as better equipped to handle “masculine issues” such as national security, foreign policy and the economy (Herrnson, Lay, & Stokes, 2003; Lawless, 2004).

Gender stereotypes based on personality reflect beliefs about the traits men and women are expected to represent (Kaid, 2004). Therefore, gender stereotyping based on personality traits is more in line with the definition where citizens base their beliefs on personal attributes or traits of men and women (Ashmore & Del Bocca, 1979). In addition, looking at gender stereotypes from a personality trait standpoint has a significant impact on elections since voters’ assessments of personality traits are known to affect their evaluations of politicians (Kahn, 1993). I will thus use the second concept of gender stereotypes for this study, personality traits. I will elaborate more on the different feminine and masculine traits in the following paragraph.

According to social role theory, gender stereotypes relate to either communal or agency attributes (Eagly, Wood, & Diekman, 2000). Different studies portray a picture of the masculine frame consisting the following traits: strength, decisiveness, individualism, aggressiveness, toughness, assertiveness, controversy and conflict, also known as agency traits (Alexander & Andersen, 1993; Eagly & Karau, 2002; Hayes, 2011; Johnston Wadsworth et al., 1987; Kittilson & Fridkin, 2008; Lippa, 2005). Frames that focus on feminine stereotypes emphasize beliefs about the personality traits that women are expected to have. These feminine traits portray an image of compassion, honesty, warmth, sympathy, gentleness, affectionateness, helpfulness and sensitiveness, also known as communal traits (Alexander & Andersen, 1993; Eagly & Karau, 2002; Johnston Wadsworth et al., 1987; Kittilson & Fridkin, 2008).

(8)

When it comes to gendered framing, there seems to be some conflict in the literature. Most studies describe masculine and feminine frames in comparable ways, but there is not a universal way of looking at the construct. Some scholars see

gendered framing as two dimensions, where masculine and feminine frames are seen as separate concepts (Alexander & Andersen, 1993). However, more recent studies describe them as one dimension (Bauer, 2015; Hayes, 2011; Rudman, Greenwald, & McGhee, 2001). For example, Rudman et al., (2001) suggest that the concepts are bipolar, ranging from strong (masculine) to weak (feminine). In addition, Hayes (2011) noted that the masculine frame portrays an image of toughness whilst the feminine frame describes a lack of toughness. After examining the different studies, it becomes clear that aggressiveness and competitiveness are the most commonly used and most impactful frames to portray masculinity (Eagly & Karau, 2002; Gordon, Shafie, & Crigler, 2003). Gentleness and helpfulness are most used to portray

femininity (Bauer, 2015; Deaux & Lewis, 1984; Hayes, 2011). I believe that these can be seen as one dimension, since gentleness is an antonym for aggressiveness and helpfulness is polar-opposite of competitiveness (Thesaurus, 2016). Moreover, lay people see masculinity and femininity as opposites (Deaux, 1987). I therefore think that using one dimension to describe gendered framing is required for this research. This leaves us with two bipolar concepts for gendered framing: aggressive to gentle and competitive to helpful.

Effects of gender stereotypical framing in ads

A lot of research on gender stereotypes has focused on identifying the different kinds of stereotypes people hold about women and men in politics, and the visibility of these stereotypes in campaigns. However, knowledge about the effect of stereotyped campaigns on attitudes and voting intention is limited. As

(9)

aforementioned, research has shown that voters tend to stereotype based on their gender. Sadly, this can pose obstacles for female politicians. Specifically, the feminine frame aligns with the role of being caretakers instead of leaders whilst masculine traits correspond with leadership traits (Eagly & Mladinic, 1994). Since feminine stereotypes are not congruent with the male dominated political traits, I expect that the feminine frame will disadvantage female politicians. I build upon Hayes (2011) who suggests that if women were to portray themselves in a counter stereotypical way, they could receive more support. Margaret Thatcher’s advisors must have thought the same, as they gave her speech lessons to come across more aggressive and less gentle in the media (Loyd, 2011).

In addition, studies show that voters prefer masculine to feminine traits in their political leaders, specifically toughness and ambitiousness (Hayes, 2011; Huddy & Terkildsen, 1993; Johnston Wadsworth et al., 1987; Lawless, 2004). However, many of these studies use survey data, whilst my research will be conducted by an

experiment. Since my research design has not been used before, an experiment will uncover new information. There was however one study that did use experimental research. It showed that female politicians who portray a strong masculine message (by being portrayed as aggressive and career oriented) in political video ads, are evaluated more positive than female politicians who portray a feminine frame (Johnston Wadsworth et al., 1987). These previous findings lead me to expect that female politicians will be evaluated best when they portray a strong masculine frame. Thus, the following hypothesis can be formulated:

H1: Masculine framed female politicians receive more support from voters than female politicians with a feminine frame.

(10)

How a politician presents herself or is represented by the media can be built quite easily. However, some characteristics of a person are not so easily framed, such as the facial traits of a politician. For millions of years, faces have been an important source of information for humans when evaluating others. Although female

politicians can influence the way they portray themselves, or the way the media portray them, changing one’s facial structure is difficult to adjust.

Facial sexual dimorphism (facial traits) is the difference in people’s faces that varies only because they were born male or female. Figure 1 illustrates this. First, the eyes differ, whereby a feminine face has notably larger eyes and eyebrows are raised. Second, the facial structure differs. A feminine face has a small chin and rounded features. A masculine face on the other hand has thicker upper brows, a larger chin, thinner lips, more squared forehead and a more prominent nose (Cunningham, 1986; Perrett et al., 1998). See figure 1 for an example of the same female with adjusted masculinity and femininity.

Figure 1. Differences between a feminine face (left) and a masculine face (right).

Some scholars demonstrated that people preferred to vote for male politicians with feminine facial features over masculine features in peace-time, but not during war (Daniel et al., 2013; Little et al., 2007; Spisak et al., 2012). These scholars only

(11)

focused on males, so how does this play out for females? Not much is known about the effect of female politicians as opposed to male politicians with manipulated masculine or feminine facial features. One recent study did look at female politicians and voting intention. The results show that female politicians with more feminine features tend to win elections, while females with more masculine features tend to lose (Hehman et al., 2014). Yet, this study did not focus on sexual facial dimorphism. They showed two completely different female politicians. For instance, one had short blonde hair and was seen as more masculine, whilst the other had long dark hair and was seen as more feminine. This makes the results difficult to interpret as multiple factors could account for the effect. Their take on femininity is therefore different than mine, as I only alter facial features within the same female politician.

Since it is difficult to alter facial features, some female politicians try to enhance their femininity in other ways, as we saw with Angela Merkel in 2005. When she was elected Prime Minister, she got a makeover. Her hair and make-up were altered, trying to make her look more feminine. But did this help her in becoming more popular? It seems interesting to examine whether female politicians who look more feminine are rated better than those who look more masculine. I expect that some women have an advantage that they are born with. Therefore, the following hypothesis can be formulated:

H2: Female politicians with feminine facial traits receive more support than female politicians with masculine facial traits.

Not only Angela Merkel, but also Margaret Thatcher is a great example of a female politician who had to strategize her way into popularity. Before she was elected Prime Minister, she got vocal lessons to sound more masculine in her

(12)

and acting more masculine is presumably the way to the top. When being framed as masculine, voters will ascribe the masculine stereotypes of strong leadership to her (Hayes, 2011). And when she looks feminine, voter support will be greater (Hehman et al., 2014). Therefore, I expect that when a female politician comes across as masculine in her framing style but looks feminine, she combines the best of both worlds. Thus, the following hypothesis is formulated:

H3: Female politicians framed in a masculine way with feminine facial traits receive more support than female politicians framed in a feminine way whilst having

masculine facial traits.

Moderators

While many scientists have studied the effects of frames on attitude and behavioural intention, the effects are usually small (Potter & Riddle, 2007). One explanation is because framing effects depend on individual differences (Aarøe, 2011; Choma, Hodson, & Costello, 2012; David & Olatunji, 2011; Valkenburg & Peter, 2013). These individual differences are known to influence citizens’ voting intentions and attitude towards politicians (Chong & Druckman, 2007). Since female politicians are not treated the same as their male counterparts (Bauer, 2015), I expect individual differences to account for the effects of the frames. Thus, I examine three individual difference variables that can moderate the effect of the campaign ads.

First, social dominance orientation (SDO). SDO is a persons’ preference for inequality amongst social groups (Pratto, Sidanius, Stallworth, & Malle, 1994). SDO is found to be a powerful predictor of generalized prejudice against groups such as gays, Jews, and women (Sidanius & Pratto, 1999). Social dominance theory (SDT) is the theory behind SDO that helps us understand how the process of group-based hierarchy works (Pratto, Sidanius, & Levin, 2006). SDT states that “legitimising

(13)

myths” forms citizens’ hierarchy based behaviour. Hierarchy-enhancing legitimising myths are generally held beliefs, which include stereotypes. Citizens with SDO use these myths to justify that inequality is legitimate. And since women are stereotyped as gentle and nurturing and men are stereotyped as better leaders, this seems to account for the prejudice that citizens with high SDO hold against women (Lee, 2014). Based on SDT, I theorize that those with high SDO will rate the more feminine looking politicians as less competent because they are prejudiced towards women. They will also have less support for the masculine frame since the politician is not acting in line with her sex. So the more feminine a woman looks and the more masculine she acts, the more prejudice a person with high SDO becomes. This accounts for them rating her as less competent for the job. Thus, the following hypothesis is formulated.

H4: The effect of H3 will be less strong for citizens with high levels of SDO than for those with low levels of SDO.

Secondly, a person’s gender identity can also moderate the effect of the ad. The notion that men and women predictably differ in political attitudes and behaviour is well supported. However, this study advocates the belief that someone can

biologically be female but have a stronger masculine gender identity than someone who is actually born male, and vice versa. And since this research illustrates the importance of differentiating sex from gender, gender identity is used as a moderator and not sex. Gender identity represents a persons’ disposition or attitude about themselves (Abdelal et al. 2006). According to McDermott & Hatemi (2011), gender identity has a bigger impact on voter preference than sex and it is debated whether a persons’ sex is the reason for his or hers political behaviour (Hatemi, McDermott, Bailey, & Martin, 2012). In addition, focusing only on the sex of voters can be risky,

(14)

as sex does not always correspond with one’s gender identity (Dolan, 2008). Burn (1996) suggested that gender identity could fuel gender equality efforts within women whilst triggering criticisms from men against gender equality. I therefore predict that men with a strong gender identity will be less supportive of the female politicians than men with a weak gender identity. I also expect that women with a strong gender identity will evaluate the female politicians better than those with a weak gender identity. Thus, the following hypotheses are formed:

H5a: The effect of H3 will be less strong for male voters who have a strong gender identity, compared to males with a weak gender identity.

H5b: The effect of H3 will be stronger for female voters who have a strong gender identity, compared to females with a weak gender identity.

Sexism

Sexism is the prejudice people hold about someone else’s sex. And although sexism can happen to both men and women, it is more common to negatively affect females. I therefore expect sexist people to be less in favour of the female politician compared to those who are less sexist. They will rate the more feminine looking politician as less competent because they have prejudice against women. If she looks more feminine it is thus expected this will play out negative amongst voters with high sexism.

H6: The effect of H3 will be less strong for citizens with high sexism than for those with low levels of sexism.

Method Sample

(15)

amongst Dutch citizens. Participants were recruited through a convenience sample, making sure the sample was as diverse as possible in terms of gender and political ideology. Recruiting participants far outside of my own circle accounted for this. The data collection ranged from November18th till December 5. In that period, 210 respondents ranging from 18 to 81 years old filled out the survey and 176 saw one of the conditions. This led to a completion rate of 83.8% with an average of 44

respondents per condition. The average age was 39 years old (SD = 17.16) and 52.8% was male. The data in this study is analysed with IBM SPSS.24.

Design and procedure

The experiment followed a 2 (masculine frame/feminine frame) x 2

(feminine/masculine facial traits) between subjects factorial design. The participants started filling out questions about age, social dominance orientation, gender, gender identity and sexism. Afterwards, participants were randomly assigned to one of four conditions. In the conditions, participants were exposed to either a masculine or feminine framed campaign message in combination with either a masculine or feminine looking female politician. Then respondents were asked about their attitude and voting intention towards this politician. The survey ended with questions about level of education, political interest and political ideology. See appendix A for the item wording.

Stimulus material

Participants received a political campaign poster with the face of a politician, accompanied by a short text. The frames contained gender-stereotyped statements. The masculine frame consisted of the female politician stating that she will fight whoever gets in her way (aggressive and competitive) and the feminine frame consists of her saying that she wants to help other people who are in need (gentle and helpful).

(16)

Pictures of a female politician accompanied the frames. The facial expression was neutral in each of the variations of the face. I chose an image from Google Images which best represented the descriptions of masculine and feminine facial features. The feminine face had a small chin and rounded features whilst the masculine face has thick upper brows, a large chin, thin lips, more squared forehead and a more prominent nose. Turn to Appendix B for the stimulus material.

Pre-test

The stimulus material was pretested to a sample of 14 respondents, 7 males and 7 females. Participants read the masculine and feminine frames and were asked whether they found the frames to come across masculine feminine. This was

measured on two 5-point bi-polar scales ranging from aggressive (1) to gentle (5) and from competitive (1) to helpful (5). The pre-test showed that the stimulus material was successful for both aggressiveness/gentleness t(11) = -6.19, p < .001, 95% CI [-2.25; -1.04] as well as for competitiveness/helpfulness t(11) = -7.65, p < .001, 95% CI [-3.53; -1.95]. On average, respondents in the masculine condition found the frame to be more aggressive (Mdifference = 1.64, SDdifference= .27) and more competitive (Mdifference

= 2.73, SDdifference= .36) compared to the participants in the feminine condition. To

check whether participants also saw a difference in the masculine and feminine faces, they were shown images of two fictional politicians, one male and one female with altered masculine and feminine facial traits. Participants were asked to what extent they saw the politicians as being masculine or feminine. This was measured on a 5-point scale ranging from masculine (1) to feminine (5). An independent sample t-test was performed to see if the images worked as they were intended t(11) = .65, p = .95, 95% CI [-.83; .88]. Participants in the masculine condition did not find the face of the woman to be more masculine compared to the respondents in the feminine condition

(17)

(Mdifference = .02, SDdifference= .39). Appendix C shows the pictures used in the first

pre-test.

In the second pre-test, a different image was chosen with a more obvious distinction of masculinity and femininity. This time I only used a female face and no male face. Based on previous facial trait research, the faces were outlined with black so only the actual face was visible and chest or clothes would not interfere (Spisak et al., 2012). In the second pre-test, 20 participants, 10 males and 10 females, completed the online survey and were asked the same questions as in the first pre-test. The new pictures proved to be successful t(21) = 4.15, p < .001, 95% CI [.64; 1.92]. On

average, participants in the masculine condition evaluated the face of the woman to be more masculine and more feminine in the feminine condition (Mdifference = 1.28,

SDdifference= .31).

Operationalization

The dependent variables of this study, attitude and voting intention towards the female politician were both measured using one item. Attitude was measured by competence, on a 7-point Likert scale ranging from 1 (not at all competent) to 7 (very

competent).

Voting intention was measured with one question, asking participants to indicate how likely it is they would vote for the person in the ad. The question was measured on a 7-point scale ranging from 1 (very unlikely) to 7 (very likely). In addition, participants were asked about demographics such as gender, age, education, political ideology and political interest.

The moderating variables in this study; SDO, gender identity and sexism, were measured by multiple items. All scales were translated to Dutch by three Master students and the most common translations were used in order to secure internal

(18)

validity. The short SDO scale was used to measure SDO (Ho et al., 2015). The scale consists of two components; trait dominance and anti-egalitarianism. Respondents were shown eight items such as “Some groups of people are simply not the equals of others”. Items were measured on a 7-point scale ranging from 1 (strongly disagree) to 7 (strongly agree). A factor analysis of all eight items used to measure SDO showed two components with an eigenvalue higher than 1. The two components explain 58,8% of the total variance in the eight items. A reliability analysis showed that all eight items combined had a Cronbach’s alpha of .75 and that removing any items would not significantly improve the alpha. Thus, a reliable scale was constructed for SDO using all eight items (M = 20.61, SD = 7.83), with 1 indicating a very low social dominance orientation and 7 indicating a very high social dominance orientation. Table 1D shows the factor loadings for each of the items.

Gender identity was measured by 7 items that were different for males and females (Hatemi et al., 2012). An example of a gender identity question for females was “I often think I would rather be a man” whilst for males the question would be “I often think I would rather be a woman”. Items were originally measured on a 2-point scale, ranging from 1 (yes) to 2 (no). But I measured them on a 7-point scale ranging from 1 (very much disagree) to 7 (very much agree), as I believe gender identity is not as black and white as simple yes or no answer. The items were first recoded to make sure a high score on the GI scale would mean a high gender identity. Separate principal components factor analyses were performed for each sex. GIF for females and GIM for males.

A factor analysis was used to test whether the seven items GIF items formed a reliable scale. Three components had an eigenvalue higher than one, which explained 66,4% of the total variance in the seven items. However, as shown in table 2D, the

(19)

seven items proved to form a reasonably reliable scale to measure GIF (α = .69) and removing any items from the scale would not improve Cronbach’s alpha. Thus, a scale was constructed for GIF using all seven items (M = 38.12, SD = 6.07), with 1 indicating a very low female gender identity and 7 indicating a very high female gender identity.

Another confirmatory factor analysis was performed for the seven items used to measure gender identity for males (GIM). This showed two components with an eigenvalue higher than 1. These components explain 55,7% of the total variance in the seven items. A reliability analysis showed that all seven combined had a Cronbach’s alpha of .70 and removing any items from the scale would not improve the alpha. Thus, a reasonably reliable scale was constructed only for GIM using all seven items (M = 37.90, SD = 6.54), with a score of 1 indicating very low gender identity and a score of 7 indicating high gender identity. Table 3D shows the factor loadings for GIM.

Finally, respondents were asked about sexism. The modern sexism (MS) scale was used for this as it “measures covert or subtle forms of sexism (sexism that is either hidden and clan- destine or unnoticed because it is built into cultural and societal norms)” (Swim & Cohen, 1997, p. 103). Participants were asked to fill out 8 questions such as “Women often miss out on good jobs due to sexual discrimination.” The 8 items were measured on a 7-point scale ranging from 1 (very much disagree) to 7 (very much agree). Again, the items were recoded to ensure a high score on the sexism scale would mean that a person is sexist. A factor analysis of the eight items used to measure sexism showed two components with an eigenvalue higher than 1. These explain 62,1% of the total variance in the eight items. Table 4D shows the factor loadings for each of the items. A reliability analysis showed that all eight items

(20)

combined had a Cronbach’s alpha of .85 and that removing any items would not improve the alpha. Thus, a reliable scale was constructed for sexism using all eight items (M = 27.36, SD = 8.80), with a score of 1 indicating low sexism and a score of 7 indicating high sexism.

Results Randomisation check

In order to test whether the variables were evenly distributed over the

conditions, randomization checks were performed for age, gender, level of education, political interest and political ideology. Using one-way ANOVA’s, no difference was found across the four conditions for age (F(3, 174) = 0.31, p = .821), political

ideology (F(3, 171) = .88, p = .451) and political interest (F(3, 173) = 1.24, p = .296). Pearson’s Chi-squared found no difference across the four conditions for gender (χ²(3) = 2.28, p = .516) and level of education (χ²(3) = 2.11, p = .550). This means that the randomisation was successful. It can thus be concluded that age, gender, level of education, political ideology and political interest and are equally distributed across the four conditions.

Manipulation checks

Even though the pre-test showed that the stimulus material was effective, manipulation checks were carried out to make sure this was also the case for the actual survey. The manipulations of the frame and face were again found to be effective in portraying masculinity and femininity. Manipulation check questions were measured ranging from 1= very masculine/aggressive/competitive to 5= very feminine/gentle/helpful. A one-way ANOVA was used to perform the manipulation check on the effect of the frame on evaluations of the frame. There was a statistically

(21)

significant difference between the masculine frame and feminine frame, F(1, 175) = 39.73, p < .001. Participants in the feminine condition found the text to be

significantly more gentle and helpful (M = 4.18, M = 4.00) and participants in the masculine frame condition evaluated the text as more aggressive and competitive (M

= 2.35, M = 2.44). As for the manipulations of the face of the female politician, a

t-test showed that the manipulations were successful in portraying either a more masculine face, or a more feminine looking face. There was a statistically significant difference between respondents’ perception of the feminine face and masculine face,

t(165.97) = -5.97, p = < .001, 95% CI [-1.07; -.54]. Respondents in the feminine

condition evaluated the face of the politician as more feminine (M = 4.25, SD = .73) whilst those in the masculine condition evaluated the face as more masculine (M = 3.44, SD = 1.05).

Effects of gender-stereotyped frames and faces

Two-factor ANOVA’s were performed for voting intention and attitude to measure the first three hypotheses. I expect the masculine frame to have a more positive effect voting intention and attitude towards the politician than the feminine frame (H1). As table 5 shows, the first two-factor ANOVA was performed with frame (masculine/feminine) as the independent variable and voting intention (1 = low, 7 = high) as the dependent variable. The test revealed no significant difference between participants’ voting intention towards the politician in the masculine frame condition compared to those in the feminine condition, F(1, 161) = .87, p = .352. The second ANOVA with face as independent variable and attitude (1 = negative, 7 = positive) as dependent variable can be found in table 6. Results showed no significant difference between participants’ attitude towards the politician in the masculine condition compared to the feminine condition. The results suggest that a masculine frame does

(22)

not have a more positive effect on Dutch citizens’ attitude towards a female politician than a feminine, F(1, 173) = .06, p = .796. Contrary to my expectation, the results indicate that a masculine framed campaign does not affect Dutch citizens’ voting intention and attitude towards a female politician more positive than the feminine framed campaign. Therefore, H1 is rejected.

For H2 it was expected that the campaign ad of a female politician with feminine facial traits generates a higher voting intention (H2a) and more positive attitude (H2b) amongst Dutch citizens than does a female politician with masculine facial features. The first univariate ANOVA in table 5 revealed a non-significant effect of independent variable facial traits (masculine/feminine) on dependent variable voting intention (1 = low, 7 = high), F(1, 161) = .42, p = .516. Consequently, a female politician with feminine facial traits does not have a more positive effect on Dutch citizens’ voting intention than a female politician with masculine facial traits. Therefore, the results do not support hypothesis H2a. However, the univariate ANOVA performed for H2b did show a small significant effect of independent variable facial traits on dependent variable attitude towards the politician (1 =

negative, 7 = positive), F(1, 173) = 629, p = .013, η2 = 0.01. The results can be found in table 6. Respondents in the feminine condition had a more positive attitude towards the politician (M = 4.45, SD = 1.17) than participants in the masculine condition (M = 3.95, SD = 1.46). In other words, a female politician with feminine facial traits does have a more positive effect on Dutch citizens’ attitude than a female politician with masculine facial traits. The results thus support hypothesis 2b.

Next, I expect the campaign ad of a female politician framed in a masculine way whilst having feminine facial traits to generate a higher voting intention (1 = low, 7 = high) and a more positive attitude (1 = negative, 7 = positive) than a female

(23)

politician framed in a feminine way whilst having masculine facial features (H3). The two-factor ANOVA in table 5 showed no significant interaction effect of independent variables masculine frame and feminine face on the dependent variable voting

intention, F(1, 161) = 1.34, p = .249. The respondents who saw the ad with the masculine framed message and feminine face did not have a higher voting intention towards the politician than those who were in the other conditions. As table 7 shows, the means barely differed across the conditions. The second two-factor ANOVA for dependent variable attitude also showed a marginally significant effect, F(1, 173) = 3.23, p = .074. Participants who saw the masculine frame in combination with the feminine face had a more positive attitude towards the politician (M = 4.28, SD = .21) than those who saw the feminine frame and masculine face (M = 3.73, SD = .20). As table 7 shows, the campaign ad of a female politician framed in a masculine way whilst having feminine facial traits (condition 2) does generate a slightly more positive attitude than a female politician framed in a feminine way whilst having masculine facial features (condition 4). However, this difference is not significant. Therefore, the results do not support hypothesis 3.

Moderators

The moderators SDO, GIF and GIM were initially measured on 7-point scales but had to be recoded in to 1 (low) and 2 (high), in order to make sure enough

participants were in each of the groups. Sexism was also measured on a 7-point scale, but was recoded into 1 (low), 2 (medium) and 3 (high). As for SDO, I expect the campaign ads of a female politician with a masculine framed and a feminine face to generate a higher voting intention and more positive attitude than vice versa.

However, this effect is expected to be less strong for citizens with high levels of SDO than for those with low levels of SDO. A two-factor ANOVA was conducted to

(24)

examine the effect of independent variables campaign ads and SDO on dependent variable voting intention (1 = low, 2 = high). There was a small statistically significant interaction between the effects of the ads and SDO on voting intention,

F(3, 157) = 2.92, p = .036, η2 = 0.01. Post hoc comparisons with Bonferoni correction indicated that in the masculine frame/feminine face condition, participants with high SDO had a lower voting intention (M = 3.38, SD = .55) than those with low SDO (M = 4.76, SD = .48). The next two-factor ANOVA was conducted to examine the effect of independent variables campaign ads and SDO on dependent variable attitude (1 = negative, 7 = positive). Contrary to the expectation, the results were not significant,

F(3, 169) = .85, p = .467. In sum, SDO moderates the effect of the ad for Dutch

citizens’ voting intention but not for attitude on a female politician. Therefore, the results support H4 but only for voting intention. See table 8D and 9D for a clear overview of the results of H4.

In addition, I expect the campaign ad with the masculine frame and feminine face to generate a higher voting intention (1 = low, 7 = high) and more positive attitude (1 = negative, 7 = positive) than vice versa. But this effect is expected to be less strong for male citizens who have a strong gender identity compared to those with a weak gender identity (1 = weak, 2 = strong). A two-factor ANOVA showed no significant interaction effect for voting intention (F(3, 79) = 1.21, p = .311) and attitude (F(3, 85) = .63, p = .597). See table 10D for the results. In sum, gender identity does not moderate the effect of the ad for Dutch males’ voting intention and attitude about a female politician. Therefore, the results do not support the hypothesis and H5a is rejected.

It was expected that the campaign ad would have a more positive effect on attitude and voting intention amongst female respondent who have a strong gender

(25)

identity (GI) compared to those with a weak gender identity (H5b). The two-factor ANOVA in table 12D shows a marginally significant interaction effect for voting intention, F(3, 69) = 2.66, p = .056. Post hoc comparisons with Bonferoni correction indicated that in the masculine frame/feminine face condition, females with high GI had an overall higher voting intention (M = 5.86, SD = .87) than those with low GI (M = 4.36, SD = .69). As table 13D shows, no significant effect was found for and

attitude, F(3, 175) = .41, p = .745. In sum, gender identity does not moderate the effect of the ad for Dutch females’ voting intention and attitude towards a female politician. However, it must be noted that for voting intention, the results were marginally significant.

Lastly, the interaction effect for H6 was measured with a two-factor ANOVA. I expect the campaign ad with the feminine face to generate a higher voting intention (1 = low, 7 = high) and more positive attitude (1 = negative, 7 = positive) than vice versa. But this effect is expected to be less strong for citizens who score high on the sexism scale compared to those with low levels of sexism (1 = low, 2 = medium, 3 = high). The ANOVA with independent variables face and sexism and dependent variable voting intention showed a small significant interaction effect, F(2, 159) = 4.03, p = .020, η2 = 0.01. Participants high in sexism in the feminine face condition had a lower voting intention (M = 3.56, SD = .44) than those with low levels of sexism (M = 5.65, SD = .41). The two-factor ANOVA with independent variables face and sexism and dependent variable attitude was not significant, F(2, 171) = 1.71,

p = .182. See table 14D and 15D for the results of H6. In sum, sexism does moderate

the effect of a female politicians’ face on citizens’ voting intention, but not for

attitude towards a female politician. Therefore, the results only support the part of the hypothesis about voting intention, but not for attitude.

(26)

Discussion

The aim of this study was to find out whether, and for whom, gender

stereotypical frames and gender stereotypical facial traits influence voting intention and attitude of citizens towards female politicians. This study suggests that only a feminine face positively affects citizen’s attitude towards a female politician. Gender stereotypical frames however, do not influence voting intention or attitude. Next, gender stereotypical facial traits positively influence attitudes towards female politicians, but do not influence the voting intention. A more feminine face yields a more positive attitude than a masculine looking face. Secondly, this study intended to research which individual differences could moderate the effect of the campaign ads. Social dominance orientation moderated the effect of the ad on citizens’ attitude towards the politician. Those with high social dominance orientation had a less positive attitude towards the feminine looking politicians. The other variables did not moderate the effect of the ad or had a marginally significant effect.

The positive effect that a more feminine looking face has on citizens’ attitudes is in accordance with previous research (Hehman et al., 2014). In doing so, this study confirms previous findings and contributes to the limited amount of research on facial traits of female politicians. So although studies show that voters prefer masculine to feminine traits in their political leaders, this is not the case for facial traits (Hayes, 2011; Huddy & Terkildsen, 1993; Johnston Wadsworth et al., 1987; Lawless, 2004). This means that when research is conducted on politicians, it should be done

separately for male and female politicians.

An explanation for the non-significance of the gender-stereotyped frames could be that the frame text was very short. Namely, the text that was used was only

(27)

one sentence long. Even though the manipulation check was successful in portraying masculinity and femininity, a longer text might have had more effect. According to social role theory, communal or agency attributes are assigned to men and women. But it could be the case that these attributes were not assigned to the female

politicians in the stimulus material used in this study. Perhaps the Dutch people, who are typically known for being open-minded, are not holding on to stereotypes as much as Americans. Since this is the first study with a Dutch sample and most previous research on gender stereotypes is U.S. based, this could be an explanation for the findings.

While this research was performed with great care, there are some limitations. First, the sample was somewhat small for the experimental design. Because of this, the moderating variables social dominance orientation, male gender identity and female gender identity could not be treated as continuous variables or be recoded into a low/medium/high variable. To ensure that the groups would not be too small, these moderating variables were turned into dummy variables (low/high). A big

disadvantage of turning these continuous variables into two category variables is as follows. Some people in the low category could actually be closer to people in the high category. If there would have been three categories these people could have been in the same group, which is neater than how it was done now. However, I chose this method to ensure large enough groups.

Second, when looking at the results that were significant, none of them were significant for voting intention as well as for attitude. It was expected that when someone’s voting intention was positive, his or hers attitude would match the voting intention. This did not turn out to be the case. A cause of this could be the way in which attitudes and voting intention were measured. Both were measured with only

(28)

one question instead of multiple questions, as previous research tends to do (BRON). In addition, self reported attitude and voting intention can be tricky since people tend to answer in socially desirable manner, especially when they personally know the researcher, which was the case here.

Despite the limitations, this study has uncovered some important elements. Citizens prefer a more feminine looking face in female politicians, which is unlike previous studies who have studied male politicians (Laustsen & Petersen, 2015). This means that female politicians are treated differently from their male counterparts and that perhaps they should also be treated differently in research. The

underrepresentation of female politicians could be the source of the problem. For centuries, men have been the ones in powerful position. Perhaps voters are not used to seeing women in high leadership roles, and therefore stereotype them straightaway. This study could serve as a starting point for other researchers and maybe this will become the era of much needed research and focus on female politicians.

To conclude, this study showed that female politicians are positively affected by having a stereotypical feminine look. Policy makers can implement the findings into their work and pollsters could possibly predict popularity of female party leaders. And although I do not advocate judging a book by it’s cover, female politicians can use this information to increase voter support, by enhancing their feminine side. This research could thus help battle the underrepresentation of female politicians in governments.

(29)

References

Alexander, D., & Andersen, K. (1993). Gender as factor in the attribution of leadership traits. Political Research Quarterly 46: 527-545.

Ashmore, R., & Del Boca. F. (1979). Sex stereotypes and implicit personality theory: Towards a cognitive-social psychological conceptualization. Sex Roles, 5, 219–48.

Bystrom, D. & Kaid, L.L. (2002). Are women candidates transforming campaign communication? A comparison of advertising videostyles in the 1990s. In C.R. Rosenthal (ed.) Women Transforming Congress. Norman, OK: University of Oklahoma Press, 489–504.

Devere, H., & Davies, S. G. (2006). The Don and Helen New Zealand Election 2005: A media a- gender?. Pacific journalism review, 12(1).

Aarøe, L. (2011). Investigating frame strength: The case of episodic and thematic frames. Political Communication, 28(2), 207–226.

http://doi.org/10.1080/10584609.2011.568041

Alexander, D., & Andersen, K. (1993). Gender as a factor in the attribution of leadership traits. Political Research Quarterly, 46(3), 527–545.

http://doi.org/10.1177/106591299304600305

Bauer, N. M. (2015). Emotional, sensitive, and unfit for Office? Gender stereotype activation and support female candidates. Political Psychology, 36(6), 691–708. http://doi.org/10.1111/pops.12186

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).

(30)

http://doi.org/10.1371/journal.pone.0003666

Choma, B. L., Hodson, G., & Costello, K. (2012). Intergroup disgust sensitivity as a predictor of islamophobia: The modulating effect of fear. Journal of

Experimental Social Psychology, 48(2), 499–506.

http://doi.org/10.1016/j.jesp.2011.10.014

Chong, D., & Druckman, J. N. (2007). Framing public opinion in competitive democracies. American Political Science Review, 101(4), 637–655. http://doi.org/10.1017/S0003055407070554

Daniel, E., DeBruine, L. M., Jones, B. C., & Perrett, D. I. (2013). Facial cues to perceived height influence leadership choices in simulated war and peace contexts. Evolutionary Psychology, 11(1), 89–103.

http://doi.org/10.1177/147470491301100109

David, B., & Olatunji, B. O. (2011). The effect of disgust conditioning and disgust sensitivity on appraisals of moral transgressions. Personality and Individual

Differences, 50(7), 1142–1146. http://doi.org/10.1016/j.paid.2011.02.004

Deaux, K., & Lewis, L. L. (1984). Structure of gender stereotypes: Interrelationships among components and gender label. Journal of Personality and Social

Psychology, 46(No. 5), 991–1004. http://doi.org/10.1037/0022-3514.46.5.991

Devere, H., & Davies, S. G. (2006). The Don and Helen New Zealand election 2005: A media a-gender? Pacific Journalism Review, 12(1), 65–86.

Dolan, K. (2008). Is there are ‘gender affinity effect’ in American politics? Political

Research Quarterly, 61(1), 79–89.

Dolan, K. (2010). The impact of gender stereotyped evaluations on support for women candidates. Political Behavior, 32(1), 69–88.

(31)

Dolan, K. (2014). Gender stereotypes, candidate evaluations, and voting for women candidates: What really matters? Political Research Quarterly, 67(1), 96–107. http://doi.org/10.1177/1065912913487949

Eagly, A. H., & Karau, S. J. (2002). Role congruity theory of prejudice toward female leaders. Psychological Review, 109(3), 573–598. http://doi.org/10.1037/0033-295X.109.3.573

Eagly, A. H., & Mladinic, A. (1994). Are people prejudiced against women? Some answers from reserch on attitudes, gender stereotypes, and judgement of competence. European Review of Social Psychology, 5(January), 1–34. http://doi.org/10.1080/14792779543000002

Eagly, A. H., Wood, W., & Diekman, A. B. (2000). Social role theory of sex

differences and similarities: A current appraisal. The developmental social psychology of gender. Retrieved from

https://books.google.nl/books?hl=nl&lr=lang_nl&id=03A4363JQd4C&oi=fnd&

pg=PA7&dq=online+marketing&ots=ODnpzynscc&sig=-ImcjM4fZTYvTV7egVnPpdXYXB4

Gordon, A., Shafie, D. M., & Crigler, A. N. (2003). Is Negative Advertising Effective for Female Candidates?: An Experiment in Voters’ Uses of Gender Stereotypes.

The International Journal of Press/Politics, 8(3), 35–53.

http://doi.org/10.1177/1081180x03008003003

Hatemi, P. K., McDermott, R., Bailey, J. M., & Martin, N. G. (2012). The different effects of gender and sex on vote choice. Political Research Quarterly, 65(1), 76–92. http://doi.org/10.1177/1065912910391475

Hayes, D. (2005). Candidate qualities theory through a partisan of trait ownership.

(32)

http://doi.org/10.2307/3647705

Hayes, D. (2011). When gender and party collide: Stereotyping in candidate trait attribution. Politics & Gender, 7(2), 133–165.

http://doi.org/10.1017/S1743923X11000055

Hayes, D., & Lawless, J. L. (2015). A non-gendered lens? Media, voters, and female candidates in contemporary congressional elections. Perspectives on Politics,

13(1), 95–118. http://doi.org/10.1017/S1537592714003156

Hehman, E., Carpinella, C. M., Johnson, K. L., Leitner, J. B., & Freeman, J. B. (2014). Early processing of gendered facial cues predicts the electoral success of female politicians. Social Psychological and Personality Science, 5(7), 815–824. http://doi.org/10.1177/1948550614534701

Herrnson, P. S., Lay, C. J., & Stokes, A. K. (2003). Women running ‘as women’: Candidate gender, campaign issues and voter-targeting strategies. The Journal of

Politics, 65(1), 244–255. http://doi.org/10.1196/annals.1340.018

Hitchon, J. C., Chang, C., & Harris, R. (1997). Should women emote? Perceptual bias and opinion change in response to political ads for candidates of different

genders. Political Communication, 14(September), 49–69. http://doi.org/10.1080/105846097199533

Ho, A. K., Sidanius, J., Kteily, N., Sheehy-Skeffington, J., Pratto, F., Henkel, K. E., … Stewart, A. L. (2015). The nature of social dominance orientation: Theorizing and measuring preferences for intergroup inequality using the new SDO 7 scale.

Journal of Personality and Social Psychology, 109(6), 1003–1028.

http://doi.org/10.1037/pspi0000033

Huddy, L., & Terkildsen, N. (1993). Gender stereotypes and the perception of male and female candidates. American Journal of Political Science, 37(1), 119–147.

(33)

Johnston Wadsworth, A., Patterson, P., Lee Kaid, L., Cullers, G., Malcomb, D., & Lamirand, L. (1987). ‘Masculine’ vs ‘feminine’ strategies in political ads: Implications for female candidates. Journal of Applied Communication

Research, 15(2), 77–94.

Kahn, K. F. (1992). Does being male help? An investigation of the effects of candidate gender and campaign coverage on evaluations of U.S. Senate

candidates. Journal of Politics, 54(2), 497–517. http://doi.org/10.2307/2132036 Kahn, K. F. (1993). Gender differences in campaign messages: The political

advertisements of men and women candidates for U.S. Senate. Political

Research Quarterly, 46(3), 481–502.

http://doi.org/10.1177/106591299304600303

Kaid, L. L. (2004). Handbook of political communication research. Mahwah, NJ: Lawrence Erlbaum Associates, Inc. http://doi.org/10.4135/9781412953993.n498 Kittilson, M. C., & Fridkin, K. (2008). Gender, candidate portrayals and election

campaigns: A comparative perspective. Politics and Gender, 4(3), 371–392. http://doi.org/10.1017/S1743923X08000330

Koch, J. W. (1999). Candidate gender and assesments of Senate candidates. Social

Science Quarterly, 80(1), 84–96.

Laustsen, L., & Petersen, M. B. (2015). Winning faces Vary by ideology: How Nonverbal Source Cues Influence Election and Communication Success in Politics. Political Communication, (September), 1–24.

http://doi.org/10.1080/10584609.2015.1050565

Lawless, J. L. (2004). Women, war, and winning elections: Gender stereotyping in the post-September 11th era. Political Research Quarterly, 57(3), 479–490.

(34)

Lee, Y.-K. (2014). Gender stereotypes as a double-edged sword in political advertising. International Journal of Advertising, 33(2), 203–234. http://doi.org/10.2501/IJA-33-2-203-234

Lippa, R. (2005). Gender, nature and nurture (Second edi). Mahway, New Jersey: Lawrence Erlbaum Associates.

Little, A., Roberts, C., Jones, B., & Burriss, R. (2007). The faces of leaders: Sexual dimorphism, perceived traits, and voting in context. Evolution and Human

Behavior, 28(1), 18–27.

Mcdermott, M. L. (1998). Race and gender cues in low-information elections.

Political Research Quarterly, 51(4), 895–918.

http://doi.org/10.1177/106591299805100403

McDermott, R., & Hatemi, P. K. (2011). Distinguishing sex and gender. PS: Political

Science & Politics, 44(1), 89–92. http://doi.org/10.1017/S1049096510001939

Pratto, F., Sidanius, J., & Levin, S. (2006). Social dominance theory and the dynamics of intergroup relations: Taking stock and looking forward. European Review of

Social Psychology, 17(1), 271–320. http://doi.org/10.1080/10463280601055772

Pratto, F., Sidanius, J., Stallworth, L. M., & Malle, B. F. (1994). Social dominance orientation: A personality variable predicting social and political attitudes.

Journal of Personality and Social Psychology, 67(4), 741–763.

http://doi.org/10.1037/0022-3514.67.4.741

Rudman, L. a., Greenwald, a. G., & McGhee, D. E. (2001). Implicit self-concept and evaluative implicit gender stereotypes: Self and ingroup share desirable traits.

Personality and Social Psychology Bulletin, 27(9), 1164–1178.

http://doi.org/10.1177/0146167201279009

(35)

peacekeepers: Testing a biosocial implicit leadership hypothesis of intergroup relations using masculine and feminine faces. PLoS ONE, 7(1), 1–8.

http://doi.org/10.1371/journal.pone.0030399

Swim, J. K., & Cohen, L. L. (1997). Overt, covert, and subtle sexism. Psychology of

Women Quarterly, 21(1), 103–118.

http://doi.org/10.1111/j.1471-6402.1997.tb00103.x

Tweede Kamer (2016) Vrouwen in de Tweede Kamer. Retrieved online from: https://www.parlement.com/id/vh8lnhrre0zv/vrouwen_in_de_tweede_kamer Valkenburg, P. M., & Peter, J. (2013). The differential susceptibility to media effects

model. Journal of Communication, 63(2), 221–243. http://doi.org/10.1111/jcom.12024

West, C., & Zimmerman, D. H. (1987). Doing gender. Gender and Society (Vol. 1). Retrieved from http://www.jstor.org/stable/189945

(36)

Tables

Table 5

Main effect and interaction effect of frame and face on voting intention

Source Sum of Squares df MS F p

Frameconditie 4.508 1 4.508 .870 .352 Faceconditie 2.191 1 2.191 .423 .516 Frameconditie * Faceconditie 6.924 1 6.924 1.337 .249 Error 833.860 161 5.179 Total 4440.000 165 Note. N = 166.

(37)

Table 6

Main effect and interaction effect of frame and face on attitude

Source Sum of Squares df MS F p

Frameconditie .118 1 .118 .067 .796 Faceconditie 11.103 1 11.103 6.292 .013 Frameconditie * Faceconditie 5.716 1 5.716 3.239 .074 Error 305.281 173 1.765 Total 3416.000 177 Note. N = 178.

(38)

Table 7

Means and standard deviations for voting intention and attitude across conditions

Voting intention Attitude

Condition N M SD N M SD

1. masculine frame x masculine face 46 4.80 2.29 50 4.14 .19 2. masculine frame x feminine face 37 4.16 2.25 39 4.28 .21 3. feminine frame x feminine face 42 4.90 2.20 44 4.59 .20 4. feminine frame x masculine face 40 4.72 2.36 44 3.73 .20

(39)

Appendix A Online survey experiment

Voordat u eventueel aan de enquête kunt beginnen moet u omwille van mogelijke ethische bezwaren eerst de volgende vraag beantwoorden:

Wat is uw leeftijd in jaren? (bijv: 34)

U krijgt nu een aantal stellingen te zien. Geef aan in hoeverre u het hiermee eens of oneens bent.

SDO_1 Een ideale samenleving vereist dat sommige groepen aan de top staan en andere op de bodem.

 Volledig mee oneens (1)  Mee oneens (2)

 Een beetje mee oneens (3)  Niet oneens of eens (4)  Een beetje mee eens (5)  Mee eens (6)

 Volledig mee eens (7)

SDO_2 Sommige groepen zijn gewoon ondergeschikt ten opzichte van andere groepen.

Q15 Geen enkele groep zou de samenleving moeten domineren.  Volledig mee oneens (1)

 Mee oneens (2)

(40)

SDO_4 Groepen aan de onderkant van de samenleving zijn net zo verdienstelijk als groepen aan de top.

SDO_5 Gelijkheid van groepen zou NIET ons primaire doel moeten zijn.  Volledig mee oneens (1)

 Mee oneens (2)

SDO_6 Het is onrechtvaardig om te proberen alle groepen gelijk te laten zijn.  Volledig mee oneens (1)

 Mee oneens (2)

SDO_7 We zouden alles op alles moeten zetten om de condities tussen de verschillende groepen gelijk te maken.

SDO_8 We moeten er voor zorgen dat alle groepen een gelijke kans hebben op succes.

(41)

Q22 Wat is uw geslacht?  Man (1)

 Vrouw (2)

Q63 U krijgt wederom een aantal stellingen te zien. Geef aan in hoeverre u het hiermee eens of oneens bent.

GIF_1 Ik vraag mij vaak af hoe het zou zijn, om een man te zijn.  Volledig mee oneens (1)

 Mee oneens (2)

 Een beetje mee oneens (3)  Niet oneens of een (4)  Een beetje mee eens (5)  Mee eens (6)

GIF_2 Ik heb het gevoel dat ik in veel opzichten meer overeenkomsten heb met mannen dan met vrouwen.

GIF_3 Mensen vinden dat ik mij vrouwelijker zou moeten gedragen dan ik doe.  Volledig mee oneens (1)

 Mee oneens (2)

GIF_4 Ik heb het gevoel dat een deel van mij vrouwelijk is en een deel mannelijk.  Volledig mee oneens (1)

 Mee oneens (2)

(42)

GIF_5 Ik denk vaak dat ik liever een man was geweest.  Volledig mee oneens (1)

 Mee oneens (2)

GIF_6 Mensen in winkels en restaurants hebben mij wel eens voor een man aangezien.

GIF_7 Ik ben er trots op dat ik vrouwelijk ben.  Volledig mee oneens (1)

 Mee oneens (2)

GIM_1 Ik vraag mij vaak af hoe het zou zijn, om een vrouw te zijn.  Volledig mee oneens (1)

 Mee oneens (2)

GIM_2 Ik heb het gevoel dat ik in veel opzichten meer overeenkomsten heb met vrouwen dan met mannen.

(43)

GIM_3 Over het algemeen begrijp ik vrouwen beter dan mannen.  Volledig mee oneens (1)

 Mee oneens (2)

GIM_4 Ik heb het gevoel dat een deel van mij mannelijk is en een deel vrouwelijk.  Volledig mee oneens (1)

 Mee oneens (2)

GIM_5 Het lijkt mij leuk om als vrouw verkleed te gaan naar een verkleedfeestje.  Volledig mee oneens (1)

 Mee oneens (2)

GIM_6 Ik denk vaak dat ik liever een vrouw zou zijn.  Volledig mee oneens (1)

 Mee oneens (2)

GIM_7 Ik voel mij niet erg mannelijk.  Volledig mee oneens (1)

 Mee oneens (2)

U krijgt nu voor de laatste keer een aantal stellingen te zien. Geef aan in hoeverre u het hiermee eens of oneens bent.

(44)

Sex_1 Vrouwen lopen geregeld goede banen mis vanwege seksuele discriminatie.  Volledig mee oneens (1)

 Mee oneens (2)

 Een beetje mee oneens (3)  Niet eens of oneens (4)  Een beetje mee eens (5)  Mee eens (6)

Sex_2 Op televisie worden vrouwen zelden op een seksistische manier behandeld.  Volledig mee oneens (1)

 Mee oneens (2)

Sex_3 De samenleving heeft een punt bereikt waarop vrouwen en mannen gelijke kansen krijgen.

Sex_4 De woede onder vrouwenrechtenorganisaties in Nederland is begrijpelijk.  Volledig mee oneens (1)

 Mee oneens (2)

Sex_5 In de afgelopen jaren hebben de overheid en de media meer zorgen getoond over de behandeling van vrouwen, terwijl dit verschilt van de daadwerkelijke ervaringen van vrouwen.