“I believe if I become Speaker of the house and in that highly visible role show the American people that women know how to use power, I think it helps all women in the political process or whatever field they're in.

-Nancy Pelosi (60 Minutes, Oct. 22, 2006).

Gender stereotypes affect perceptions of politicians and partially explain women’s underrepresentation in U.S. politics (Bauer, 2015; Huddy & Terkildsen, 1993; Kahn, 1994). Stereotypes are problematic for women who seek political office because voters may perceive them as incongruent with the leader role: that is, they might be seen as having communal traits typical of women and as not possessing the agentic traits desired in a leader (Eagly & Karau, 2002; Koenig et al., 2011). A study examining the content of stereotypes of politicians, now ten years old, supports the idea that the stereotype of women politiciansFootnote 1 is incongruent with leader stereotypes (Schneider & Bos, 2014).

During the past decade, however, there has been an infusion of women in prominent, political leadership roles in the U.S. Based on the increased descriptive representation of women in politics, the electorate is likely to be much more familiar with women politicians than before, as suggested by Nancy Pelosi’s words in the epigraph. And because stereotypes are social constructs that are dynamic in nature and evolve as social roles evolve, gendered political stereotypes may have shifted as a result (Diekman & Eagly, 2000; Eagly et al., 2020; Wilde & Diekman, 2005),Footnote 2 and in ways that further women’s opportunities in politics and beyond.

In this paper we theorize and test two specific ways the stereotype of women politicians may have changed. First, while Schneider and Bos’ study (2014) found no consensus on the traits characterizing women politicians, Americans nowadays may have a clearer picture of women politicians as a result of having more experience with them. Second, voters may now view women politicians more positively (Sweet-Cushman, 2022), and as more congruent with the traits desired in political leaders: leader competence, strong leadership, empathy and integrity (Kinder, 1986). This is because the stereotype of women now includes more agentic traits associated with competence (Eagly et al., 2020), while the leader stereotype now includes more communal traits (Koenig et al., 2011). In addition, women political leaders may be viewed as more competent, as are higher quality candidates than their men counterparts (e.g., Anzia & Berry, 2011; Bauer, 2020; Fulton, 2012). acknowledge.

In testing these ideas, we make a significant methodological contribution by using a list experiment, a new measurement approach in this context that better captures individual endorsement of stereotypes. Due to social desirability, the content of gendered politician stereotypes has typically been measured by asking respondents what “people in general think” of a particular social group (Schneider & Bos, 2014). We argue that this measures knowledge of cultural stereotypes, and not whether respondents actually endorse these stereotypes. Our innovative use of a list experiment to measure stereotype content offers researchers a new way to capture stereotype beliefs while avoiding social desirability bias.

We present the results of a large-N list experiment with U.S. participants as an improved measure of stereotype content, which we directly compare to the stereotype content study by Schneider and Bos (2014). This allows us to identify what stereotypes people have of politicians in general, women politicians, and men politicians, and how these stereotypes have changed over a decade. Our results indicate good news: the infusion of women into political leadership roles has given voters a clearer picture of women politicians, and that picture has also shifted to become more congruent with political leader roles. However, the findings also underscore the continuing masculinity of the political domain, by the strong overlap between the stereotype of men politicians and politicians in general. The stereotypes of these groups, unlike the stereotype of women politicians, evolved unfavorably, and people now are very distrusting towards them.

Gendered Politician Stereotypes

Despite recent increases in the number of women in political office, women remain numerically underrepresented worldwide (Inter-Parliamentary Union, 2022). Scholars have advanced numerous explanations for this, including gender differences in political opportunity structures (O’Brien, 2015), recruitment (Sweet-Cushman, 2020), and pipelines to office (Thomsen & King, 2020); the gender gap in political ambition (Crowder-Meyer, 2020; Schneider et al., 2016); gender differentiated media coverage (Van der Pas & Aaldering, 2020); how electoral contexts disadvantage women (Holman et al., 2016); and how candidate gender shapes voters’ information search (Ditonto et al., 2014) and their views of candidate qualifications (Bauer, 2020).

All of these explanations tie to or originate from gender stereotypes or beliefs that people hold about how men and women are or should be (Prentice & Carranza, 2002; Schneider et al., 2022). These stereotypes can affect whether candidates consider running for office, whether they are positioned to do so, and how they are viewed as candidates by the media and voters. Decades of research shows that men are generally assumed to have agentic qualities like assertive, ambitious, dominant, confident and competitive, while women are assumed to possess communal traits like warm, compassionate, sensitive, emotional and honest (e.g., Eagly & Karau, 2002). Social role theory argues that gender stereotypes derive from the sex-based division of labor in society, leading men and women to develop the traits they needed to succeed in their sex-segregated roles (Eagly & Wood, 2012; Eagly et al., 2000). Men and women continue to exhibit agentic and communal traits, respectively, due to social pressures to conform to gender roles (Eagly et al., 2000).

Gender stereotypes are assumed to disadvantage women politicians and contribute to their political underrepresentation. Extending social role theory, role congruity theory (RCT) posits that women leaders experience prejudice because the communal traits ascribed to women are not congruent with the more agentic expectations of leaders (Eagly & Karau, 2002). Men seeking leadership roles, on the other hand, benefit from agentic stereotypes since those traits align with those expected of leadership roles (Koenig et al., 2011). Applying RCT to politics, women political leaders are disadvantaged because their prescribed communality is incongruent with the agentic expectations shown to define political leader roles (Bos et al., 2017; Schneider et al., 2016, 2022). Meanwhile, when women leaders exhibit agentic traits, they can face backlash for violating the proscriptive expectations of them as women (Jamieson, 1995). Thus, women politicians experience a double bind: they either violate the expectations based on political leadership or the ones based on femininity.

To understand whether and how gender stereotypes affect women’s representation in politics, voters’ stereotypes of politicians must first be established. More precisely, we need to know whether gender stereotypes are applied to men and women politicians. Schneider and Bos’ (2014) influential work on this topic offers and in-depth examination of the content of the traits people associate with men, women, politicians, men politicians and women politicians respectively. One main finding was that there is little agreement among respondents on which traits women politicians possess, indicating that the women politician stereotype is “nebulous and lacks clarity” (Schneider & Bos, 2014, p. 261).

In contrast, voters have a clear stereotype of men politicians that overlaps largely with agentic stereotypes of men and of politicians in general. The traits most often attributed to men politicians overlap with agentic traits and highlight their capability as leaders: well-educated, charismatic, well-spoken, leader, ambitious, and driven (Schneider & Bos, 2014). The authors further examine how much respondents characterize each politician group as related to the four well-established leadership dimensions that voters want politicians to possess: leader competence, strong leadership, integrity and empathy (Kinder, 1986). The data revealed that both men politicians and politicians in general were strongly associated with the two stereotypically “masculine” trait dimensions of political leadership: strong leadership and leader competence (Schneider & Bos, 2014). And while women politicians scored considerably lower on strong leadership and leader competence, they also did not outscore men politicians and politicians in general on empathy and integrity (which relate more to feminine stereotypical qualities). The stereotype of women politicians, then, did not reflect any of the four leadership dimensions that voters value in political leaders.

Stereotype Change and Politician Stereotypes

While Schneider and Bos’ (2014)Footnote 3 study lays a foundation for gendered politician stereotype content, there are reasons to suspect that politician stereotypes, and the stereotype of women politicians in particular, have changed. As social role theory posits that gender stereotypes originate from the social roles men and women hold, these social constructs dynamically change as roles shift (Koenig & Eagly, 2014). That is, when the social roles that the stereotypes are based on change, the related stereotypes update accordingly (Diekman & Eagly, 2000; Eagly et al., 2020; Wilde & Diekman, 2005).

The stereotype of women politicians may have updated in response to notable boosts in women serving in prominent political leader roles. Globally, the percentage of women in parliament increased 5 percentage points between 2011 and 2021 (Inter-Parliamentary Union, 2011, 2022), while the number of women cabinet ministers increased 6 percentage points between 2010 and 2021 (Women in Politics:, 2010, 2010; Women in Politics: 2021, 2021). In the US, high profile women political leaders such as Nancy Pelosi and Hillary Clinton received considerable attention from both the media and the electorate. Additionally, the number of women in the US Congress increased appreciably: between the 111th Congress (2009–2011) to the 117th Congress (2021–2022), the overall number of women in Congress increased from 90 to 145 (an increase of 61%) (History of Women in the U.S. Congress, 2022).

As the number of elected women politicians increases, and as the mass media provide coverage of them, voters gain familiarity with women politicians as a social group. This familiarity informs stereotypes, which are based on directly observing members of a social group or through seeing the group’s representation in the media (Koenig & Eagly, 2014). Thus, compared to the nebulous stereotype of women politicians in 2011in 2021, citizens should exhibit a stronger consensus on the stereotypic traits of women politicians (Hypothesis 1).

As voters gained familiarity with women politicians, they may have also refined their stereotypes about them. We consider how voters may have updated their stereotype of women politicians related to agentic-communal traits as well as the common leadership trait dimensions voters use to evaluate candidates for office: competence, leadership, integrity and empathy (Kinder, 1986). This enables us to answer the question: Are women politicians less incongruent with political leaders now than a decade ago?

The 2011 data showed that women politicians did not align with the communal, agentic and competence traits generally used to describe men and women respectively, nor with the four leadership dimensions that voters value in political leaders (Schneider & Bos, 2014). Below, we lay out three reasons why we expect this has changed.

First, the stereotype of women in general has changed over time as women’s roles in society shifted with their increased workplace participation (Women in the Labor Force: A Databook, 2022) and educational attainment (Okahana & Zhou, 2018). Due to their changing social roles, the stereotype of women has become more agentic over time (Diekman & Eagly, 2000; Wilde & Diekman, 2005), both in terms of general agentic traits (e.g., dominance, aggressive and competitive) as well as the more specific competence-related agentic traits that constitute a subcategory of agency.Footnote 4 Women, for instance, are currently being perceived as having more capabilities, as similarly competent to men (Eagly et al., 2020), and as more powerful than before (Diekman et al., 2004). At the same time, women’s relative advantage over men on communal traits increased (Eagly et al., 2020). Thus, women have gained in stereotypically masculine strengths, including competence, while also retaining stereotypically feminine communal traits. In contrast, the stereotype of men has not changed over time: men’s stereotype did not come to include more communal traits (Diekman & Eagly, 2000)and their perceived levels of power remain stable (Diekman et al., 2004). Overall, the general stereotype of women was updated to them being seen as more agentic generally and more competent specifically, while the stereotype of men remains unchanged.

Second, the stereotype of leaders has also evolved over time. As more and more women held positions of power, not only in politics, but also in business, the ideas people have about what constitutes a “good” leader became less tied to masculinity (e.g., Beaman et al., 2009; Eagly & Sczesny, 2009). Koenig et al.’s (2011) meta-analysis demonstrates that the leader stereotype has become increasingly androgynous over time. They state that “leadership now, more than in the past, appears to incorporate more feminine relational qualities, such as sensitivity, warmth, and understanding, thus adding them to the masculine dominance and strength qualities traditionally associated with leadership” (Koenig et al., 2011, p. 634). Thus, the stereotype of leaders has updated to incorporate more communal leadership qualities alongside agentic traits.

Third, both aforementioned trends might be exacerbated in the political context where women politicians are shown to excel compared to their men colleagues. Women face more obstacles on their way to the top, and the women who ‘make it’ are more qualified than the men in these positions: Women candidates encounter a gendered qualifications gap where they are held to higher standards (Bauer, 2020) and must be more qualified than men to win elections (Fulton, 2012). Once elected, women legislators outperform men legislators by passing more legislation and bringing home more federal dollars to their districts (Anzia & Berry, 2011; Lazarus & Steigerwalt, 2018). Observing highly qualified and productive women political leaders may shape voters’ stereotype of women politicians such that they are viewed as having higher competence, and as possessing masculine leadership traits (leader competence and strong leadership). And having knowledge of more accomplished women seeking political office and in political leader roles, voters may now integrate feminine leadership traits (empathy and integrity) in their stereotype of politicians.

To summarize, the stereotype of women now includes more agentic traits, the stereotype of men did not change, and the stereotype of leaders now includes more communal traits. Based on these observations, we develop a set of expectations about how gendered politician’s stereotypes may have changed over the past decade. First, when comparing 2011 and 2021, we predict no change in the stereotype of men politicians and expect to find that they still are strongly associated with agentic traits and competence, and weakly with communal traits (Hypothesis 2a). Moreover, we expect men politicians still to be strongly linked to masculine leadership traits (leadership competence and strong leadership) and weakly linked to feminine leadership traits (integrity and empathy) (Hypothesis 2b).

As a result of the infusion of more qualified women politicians, we predict voters to have updated their stereotype of politicians in general to include more communal traits. More precisely, when comparing 2011 and 2021, we expect that politicians in general still are strongly associated with agentic traits (general agentic traits and agentic-competence), but are now more strongly associated with communal traits as well (Hypothesis 3a). Moreover, we expect politicians in general still are strongly linked to masculine leadership traits, but now are also more strongly connected to feminine leadership traits (Hypothesis 3b).

Lastly, we expect that the stereotype of women politicians is now more congruent with the political leader role. Thus, when comparing 2011 and 2021, we expect women politicians to now be more strongly associated with agentic traits (general agentic traits and agentic-competence) and with communal traits (Hypothesis 4a). Additionally, we expect women politicians now to be more strongly linked to masculine leadership traits and to feminine leadership traits (Hypothesis 4b).

Methods

A New Way to Measure Stereotype Content

Our ability to test these ideas about stereotype content hinges on the quality of our measures of stereotypes. Capturing stereotype content is not straightforward, in large part because respondents may feel pressured by social desirability to underreport stereotypical beliefs. This is especially the case when respondents hesitate to associate a social group with negative characteristics (e.g., Davison and Birch 2004, Weber, Lavine, Huddy and Federico 2014, see Sigall and Page 1971). To circumvent such bias, stereotype content has often been measured by asking respondents to indicate whether people in general would consider traits or adjectives to describe a social group (Devine and Baker 1991, Garcia-Marques, Santos and Mackie 2006, Schneider and Bos 2011; 201). In asking people to give other people’s opinion, rather than their own, they should not feel pressured to report a socially acceptable answer.

There are reasons to suspect that these measures do not provide an accurate reflection of stereotype content. People often display “pluralistic ignorance,” or inaccurate perceptions of other people’s opinions (Allport, 1924), and conform their behavior to the incorrectly perceived norm, which then reaffirms that norm in the eyes of others (Katz et al., 1931; Prentice & Miller, 1993). Particularly when it comes to group attitudes, people tend to overestimate other people’s prejudice (Bateson, 2020; Lucas & Ossoff, 2021; O’Gorman, 1975), via a so-called “conservative lag,” where their own prejudice has diminished, but they have yet to notice that others have done the same (Shamir & Shamir, 1997, p. 230). Thus, measures that ask respondents for other people’s opinions likely present a picture of stereotype content that is inaccurate and outdated.

List experiments provide an alternative to measuring stereotype content that avoids social desirability bias while also sidestepping the conservative lag of pluralistic ignorance. Relying on aggregated data instead of individual responses to measure sensitive items (Coutts & Jann, 2011; Glynn, 2013; Holbrook & Krosnick, 2010),list experiments randomly assign respondents to the control group or the experimental group and indicate from a list how many items apply to them. The control group receives a list of non-sensitive items and the experimental group chooses from a list with the same non-sensitive items plus the sensitive item that is the focus of the study. In comparing the number of items that respondents choose between the two groups, one can observe the extent to which respondents chose the sensitive item. For example, Streb et al. (2008) use this method to show that roughly 26% of the U.S. public is “angry or upset” about the prospect of a female president, while only around 10% admits so in a direct question (see also Burden et al., 2017).Footnote 5 We developed a novel list experiment in which we ask respondents how many traits from a list they think describe a particular social group (see Appendix A). By asking them to indicate the number of traits that describe a social group, rather than which traits, we allow respondents to mask which specific traits they link to the social group.

A recent critique of traditional list experiments is the problem of “satisficing,” where respondents choose a number of items that they deem reasonable, without actually reading the items carefully and picking the accurate number (Kuhn & Vivyan, 2018). This can lead to a systematic bias between the experimental and control group, as those who satisfice on average choose a higher number in the experimental condition than in the control condition, simply because the list is longer (Kuhn & Vivyan, 2018). To avoid this bias, we present every respondent with an equally long list of five randomly drawn trait from a master list of traits, similar to a conjoint experiment (see e.g., Hainmueller et al., 2014). By comparing, in the aggregate, how popular each trait is for a social group, we can determine the percentage of respondents linking a trait to a specific social group.

Data and Operationalization

Data were collected from 1200 respondents in December 2020 and January 2021 via the online survey agency Lucid. While Lucid respondents do not constitute a probability sample of the U.S. population, studies show that their demographic and experimental effect size estimates are similar to those of probability samples (Coppock & McClellan, 2019). In addition, we used quotas to ensure our sample is representative of U.S. voters with regard to sex, age and party identification.Footnote 6 Every respondent answered 12 list questions; each was randomly drawn to be about one of these five social groups: politicians, men politicians, women politicians, men and women.Footnote 7 For each question, a different, randomly drawn list of five traits was shown, and respondents were asked to indicate the number of traits on the list they thought were characteristic of the particular social group. These traits were drawn from a master list of 99 traits based on Schneider and Bos (2014). From the answers to the list question, in combination with which traits were shown, we calculated the percentage of respondents that select a trait as fitting with a particular social group.Footnote 8

We analyze our data in two ways, per trait and using scales. First, we compare the trait percentages in our list experiment with Schneider and Bos’ (2014) data from 2011, with the goal of understanding the change in gendered political stereotypes. However, there are differences in the methods: in the list experiment respondents are asked to evaluate 5 traits at a time, while in the original survey respondents judged the complete list of 99 traits all at once.Footnote 9 It is to be expected that respondent would pick a higher relative share of items out of a list of five than out of a list of about a hundred. Consistent with this, participants selected an average 31% of traits when faced with 99 traits, whereas traits were selected at an average of 54% in our list experiment, as displayed in Table 1. Therefore, comparing the percentages directly between these datasets would artificially inflate support for Hypothesis 1. In order to make the two datasets comparable and to ensure that the change we observe is not driven by the difference in methods, we calculate the standardized score, in which 0 represents a trait that is picked by an average share of people that year over all target groups, and 1 is one standard deviation above that average.

Table 1 Descriptive statistics for trait percentages and standardized trait score
Full size table

Second, we construct three scales based on the psychology literature and four based on Kinder’s classic work (1986). For each scale, we select relevant traits and take the average standardized percentage of respondents selecting those traits for a target social group. The three psychology-based scales are for communion, agency and competence. Communion reflects people-oriented traits, such as caring, compassionate, and warm, while agency taps into self-mastery and goal-orientation, with traits like active, ambitious and assertive (e.g., Eagly et al., 2020). Competence, with traits like intelligent, organized and smart, is sometimes considered part of agency, but follows a different trend in recent years (Eagly et al., 2020, p. 302), which is why we examine it separately. The four scales based on the Kinder (1986) trait dimensions reflecting desirable leadership are strong leadership, leadership competence, integrity and empathy.Footnote 10 These dimensions are particularly suited for our purpose because they cover two more agentic dimensions desired in leadership (strong leadership and leadership competence) and, importantly, also two more communal dimensions (empathy and integrity). As such, these dimensions should provide a comprehensive picture of changes in politician stereotypes.

To ensure that any apparent shifts over time are not driven by the difference in method between 2011 and 2021, we included a replication of the original survey question in the recent data collection, asking for the perceived opinion of people in general. Appendix E juxtaposes the scale scores of the 2011 data, the 2021 replication, and the 2021 list experiment. This shows that the 2021 replication follows a similar but slightly attenuated pattern of the 2021 list experiment, confirming that we can interpret the differences between the years as shifts over time, rather than measurement artifacts.

Results

What are the current gendered stereotypes of politicians? We start with hypothesis 1, which predicts more consensus on the traits that characterize women politicians. Table 2 compares the 2011 and the 2021 data, listing the most and least picked traits for women politicians, along with the percentages of respondents who chose each trait.Footnote 11 To compare the data over time, the standardized trait percentages are also listed. Relative to 2011, there are more traits in 2021 that a large number of respondents agree characterize women politicians. In the recent data, ten traits have a standardized score of one or higher, meaning that the percentage of respondents choosing the trait for women politicians is at least one standard deviation above the average percentage of people attributing a trait to a social group. By contrast, in 2011, this held for only five traits. Thus, there is support for Hypothesis 1, as there is more consensus on which traits are typical of women politicians now than a decade ago.

Table 2 Top twenty and bottom five traits picked for women politicians
Full size table

But how much consensus is enough to consider a stereotype clear? To put this in perspective, Table 3 lists the number of traits on which voters reach consensus (sd > 1) that they apply to women politicians, men politicians, and politicians generally in both datasets. It also lists the number of traits respondents agree are not typical of a social group (sd < − 1). On both counts, agreement about typical women politicians has grown: the number of traits associated with women politicians has increased from five to ten, while the number of traits clearly not associated with women politicians increased from 7 to 13. Table 3, however, also shows that there is more consensus on what men politicians and politicians in general are like, compared to women politicians, in both 2011 and 2021. Men politicians are attributed 26 and 20 traits in the two years respectively, and politicians 23 and 14 traits. Likewise, people agree on the traits these groups lack, with 18 to 28 traits falling into the category with a standardized score of less than -1. In conclusion, while the stereotype of women politicians has become clearer over time, it is not yet as crystallized as those of men politicians and politicians generally.

Table 3 Characteristic and uncharacteristic traits compared over time
Full size table

Besides showing more clarity on the stereotype of women politicians, Tables 2 and 3 also paint the picture people have of the groups of politicians. Notably, women politicians in 2021 are described by positive qualities like intelligent, analytical, moral, ambitious, rational and ethical, while the trait least associated with them is weak. In contrast, men politicians and politicians are linked to a remarkable number of negative traits, like self-interested, manipulative, selfish, and power-hungry, while being disassociated from gentle, sympathetic, and really cares about people like me. We come back to these cynical perceptions in the conclusion.

Next, we test the expectations regarding the agency, competence, communality and leadership abilities of the social groups. Our second hypothesis is that the stereotype of men politicians has not changed over the past ten years in their perceived agency, competence, and communality (H2a), nor in their perceived leadership qualities (H2b). Figure 1 addresses H2a, with the red squares conveying the standardized agency, competence and communality scores for men politicians. As the top-left panel shows, men politicians were considered high on agency in 2011, but this has dropped considerably by 2021 (difference test between years: t = 3.67; p < 0.000). In the recent data, men politicians are considered about as agentic as women politicians. Likewise, men politicians perceived competence dropped considerably, a shift that is also statistically significant (t = 2.60, p = 0.010). The loss in agency and competence is not compensated with more perceived communality, as men politicians are nowadays considered about as (un)communal as they were ten years ago (t = 0.74; p = 0.460).

Fig. 1
figure 1

Agency, competence and communality per social group in 2011 and 2021, with 95%CI

Full size image

Figure 2 displays the standardized scores on the four Kinder (1986) dimensions of leadership. In 2011, men politicians scored high on the two arguably masculine dimensions, leadership competence and strong leadership. By 2021, these qualities are no longer considered typical of men politicians, with leadership competence dropping below zero (difference test between years: t = 6.46; p < 0.000) and leadership to just above zero (t = 5.15; p =  < 0.000). Empathy and integrity were considered even less characteristic of men politicians in 2021 than ten years earlier, though the decline is not statistically significant (empathy: t = 1.09; p = 0.274; integrity: t = 1.38; p = 0.168). As such, the second hypothesis is not supported: men politicians are now viewed substantially lower in competence and agency, and on the masculine leadership dimensions competence and strong leadership.

Fig. 2
figure 2

Leadership scale scores per social group (Kinder, 1986), with 95% CI

Full size image

Our third hypothesis concerns politicians in general. The expectation is that politicians in general are attributed more feminine traits (communality, empathy and integrity). Figure 1 shows in the bottom panel that the general politician stereotype has not increased in communality (t = 0.98; p = 0.327). Contrary to expectation, however, the stereotype did lose substantially in agency (t = 4.62; p < 0.000). As such, politicians track closely with men politicians, as both have lost in agency and remained markedly low on communal traits. Hypothesis 3a is thus rejected.

We see the same pattern in Fig. 2, in testing Hypothesis 3b about leadership dimensions. In the 2011 data, the stereotype of politicians was high in the ‘masculine’ leadership traits of competence and strong leadership, but this is no longer the case (leader competence: t = 3.73; p < 0.000; strong leadership: t = 5.66; p < 0.000). Meanwhile, empathy and integrity have decreased even further, although just not significantly so (empathy: t = 1.45; p = 0.147; integrity: t = 1.62; p = 0.105). In sum, we reject Hypothesis 3a and 3b: the stereotype of politicians in general has not evolved to incorporate more traditionally feminine qualities like communality, empathy and integrity.

Finally, our fourth hypothesis addresses women politicians. Here, we expect that the current stereotype of women politicians includes more agency, competence and communion (H4a), as well as more of all four leadership dimensions (H4b). Looking again at Fig. 1, we see on the bottom panel that women politicians have gone up in communality (t = − 2.22; p = 0.026), in line with the expectation. The level of general agency attributed to women politicians has remained about the same (t = − 0.59; p = 0.554), but as expected women politicians are nowadays considered higher in competence than before (t = − 2.29; p = 0.022).

The leadership dimensions in Fig. 2 paint a similar picture. The stereotype of women politicians has gone up in the Kinder dimensions of leadership competence and strong leadership, although not statistically significantly (leader competence: t = − 1.55; p = 0.121; strong leadership: t = − 052; p = 0.600). Integrity is now more linked to the women politician stereotype (t = − 3.14, p = 0.002), while empathy does not achieve statistical significance (t = − 1.58; p = 0.114). Hypothesis 4 is therefore partially substantiated: women politicians have come to be seen as more communal (part of H4a) and higher in integrity (part of H4b).

Causal Mechanism

We argued that increased exposure to women politicians over the past decade has driven stereotype change. We split the results out by respondent party identification to tentatively shed light on this mechanism. As the increase in women politicians mainly took place within the Democratic Party, Democrats likely have been more attentive to the increase in women politicians. Figure 3 shows that on most dimensions, voters of the different parties share similar stereotypes, but with two notable exceptions. Democrats attribute significantly more leader competence and more empathy to women politicians than do Republicans or independents (full results in Appendix F). Thus, Democrats indeed show somewhat more positive stereotypes about the leadership capabilities of women politicians. However, while these additional analyses are consistent with the proposed mechanism, there are other possible reasons why democrats have updated their views more, and therefore they are only suggestive at best.

Fig. 3
figure 3

Results split out by respondent party identification and respondent gender

Full size image

Conclusion

This paper set out to examine how gendered stereotypes of politicians have evolved over the last decade. As part of this, we developed a novel application of a list experiment to observe stereotype content, and specifically stereotype endorsement, while avoiding social desirability and satisficing biases (Kuhn & Vivyan, 2018). As this innovative measure can be applied to stereotype endorsement for any social group, and because of the particular strength of list experiments to address social desirability, we imagine especially promising applications to stereotypes based on a range of marginalized identities such as race and ethnicity (Carew, 2016, Kelly et al. 2018) as well as in the study of intersectional stereotypes (Cargile, 2016).

Our findings show important changes in gendered political stereotypes over the last ten years. Notably, the evolution of the stereotype of women politicians is positive. While ten years ago people could not agree on the traits that describe women politicians (Schneider & Bos, 2014), now there is a clear stereotype of women politicians, who are generally thought to be intelligent, analytical, moral, ambitious and rational. Moreover, women politicians score higher on communality, competence, and integrity than a decade ago. Overall, the traits that people associate with women politicians are mostly positive and desirable for politicians. By contrast, the evolution of the stereotypes of men politicians and politicians in general is not favorable. While both groups ten years ago were associated with traits as being confident, well-educated, charismatic, well-spoken, ambitious, and driven (Schneider & Bos, 2014), now they are linked to traits as power-hungry, selfish, and manipulative. Additionally, while they previously scored high on agentic traits and masculine leadership traits, they have lost considerably on the trait dimensions of agency, strong leadership and leader competence. The latter development was not compensated with an increase on communality or feminine leadership traits.

The favorable development of the women politicians’ stereotype goes hand in hand with the evolution of the women stereotype in general, which now includes more agentic traits than before while retaining an advantage on communal traits compared to men (Diekman & Eagly, 2000; Diekman et al., 2004; Eagly et al., 2020). This also aligns with other research showing a slight preference for women over men candidates in vote choice (Bridgewater & Nagel, 2020; Schwarz & Coppock, 2022). These developments show that women might no longer be seen as incongruent with leadership roles, which could help considerably with increasing their representation in politics.

At the same time, politics is still seen as a masculine domain (Koenig et al., 2011; Schneider & Bos, 2019), and our findings show a continuing strong overlap in how people think about men politicians and about politicians. This implies that when people are asked about politicians in general, they think of men politicians, which corroborates the think-leader-think-male phenomenon we know from management studies (e.g., Schein, 1975; Schein et al., 1996), and how children mostly draw men when asked to draw a political leader (Bos et al., 2022). The masculinity of the political domain dissuades women to run for political office (e.g., Schneider et al., 2016), and parties to appoint women in leadership positions, thereby sustaining the underrepresentation of women in politics.

More pronounced than the positive trend for women politicians is how our results show increasing negativity toward men politicians and politicians generally. The fact that people strongly associate traits like power-hungry, manipulative, selfish and sleazy with them, while descriptions like moral and honest are hardly associated with them is very worrisome. Two factors might help explain this surprising bankruptcy of the men politician stereotype.

First, the timing of the 2021 data collection might influence our findings. Data collection took place in late December 2020 and early January 2021, in the midst of the second wave of the COVID-19 pandemic. The U.S. was experiencing a surge in positive COVID tests, hospitals were full and COVID deaths were on the rise (Centers for Disease Control & Prevention, 2022). The handling of the pandemic by the U.S. government was heavily criticized, and there was widespread frustration with how politicians dealt with the crisis (Altiparmakis et al., 2021; Fetzer et al., 2020). This exceptional political environment might have exacerbated the association of negative traits with (men) politicians, while activating positive stereotypes of women as honest and competent, as agents of change (Piazza & Diaz, 2020; Piazza & Schneider, 2021). Further underlining this was a prevalent narrative of women heads of state such as Merkel in Germany and Ardern in New Zealand responding exceptionally well to the crisis (Coscieme et al., 2020; Dada et al., 2021; Sergent & Stajkovic, 2020).

Second, the negative associations with (men) politicians fit into broader trends in American politics of increasing political cynicism and distrust (Citrin & Stoker, 2018) and affective polarization (Mason, 2016). Part of the electorate adheres to a populist ideology (Mudde, 2007; Norris & Inglehart, 2019), believing that the corrupt political elite are the enemy of the people. Another part of the electorate is strongly disappointed in the authoritative turn American democracy made during the Trump presidency (and before) (Abramowitz et al., 2018; Levitsky & Ziblatt, 2018). These divergent views on political reality increase the distrust towards the other side of the political spectrum and lead to a growing polarization in affect toward out-partisans (Finkel et al., 2020; Iyengar et al., 2019; Pierson & Schickler, 2020). And contributing to cynical views of politicians, at the time of our data collection Pres. Trump’s was undertaking divisive, high-profile efforts to question, deny, and overturn the results of the 2020 presidential election. In this political climate, in which political elites are disliked and distrusted, the stereotype of (men) politicians seems to have become tainted. Interestingly, this negative sentiment about politicians did not spill-over to women politicians, highlighting the fact that women politicians are still seen as a category apart from politicians.