"What's Her Job?" Agentic Women, Sexism, and the Consequences for Political Candidate Emergence

By Admin on August 1, 2024

ABSTRACT

Agentic women are frequently stereotyped negatively in politics. While public opinion suggests higher levels of egalitarian gender attitudes toward women in politics, these analyses miss important variations in explicit sexism. Using a survey experiment in the Canadian multi-party context, we show that for those reporting high levels of explicit sexism, agentic women are, as potential candidates, disadvantaged compared to similarly agentic men. Given the uneven distribution of sexist attitudes among partisans, these results suggest that gender representation differences between parties can be traced both to party recruitment strategies and the political behavior of party members.

1. We refer to "sexist attitudes" and "sexism" as general terms for our core concept of explicit sexism.

2. This research was funded by the Social Sciences and Humanities Research Council of Canada, grant #435–2015–1155.

3. These findings mimic work on racial resentment and support for Barack Obama (Tesler and Sears Citation2010), whereby underlying racial animus decreased support for Obama, while those who endorsed racial egalitarianism showed increased support for Obama.

4. While objectification and self-objectification are not a direct measure of sexism, they mimic some of the underlying essentialization of various forms of sexism, especially benevolent sexism.

5. While our sample includes both Canadian citizens and permanent residents, the vast majority (96.2%) of the sample consisted of Canadian citizens. The study was approved by the University of Calgary Research Ethics Board (Study #REB17–0861). Respondents consented to participate in the study online prior to beginning the survey.

6. For example, the occupations of winning candidates in 2019—that is, candidates who secured a nomination and then were elected to the House of Commons – include an air traffic controller, customer service (retail) workers, a lunch supervisor, and several mid- to lower-level managers in small businesses (see Johnson et al. Citation2021).

7. The other main avenue to party membership is a leadership selection contest. While party members are able to discuss policy, the overall policy direction of the party is the discretion of the party leader, central party, and party in public office (Cross, Pruysers, and Currie-Wood Citation2022).

8. This generalization is tempered in more recent elections, with left-leaning parties explicitly committing to gender parity in candidate nominations and/or cabinet appointments in ways more conservative parties have not. Specifically, the federal New Democratic Party of Canada, or NDP, has nominated an equal number of women and men since the 2015 federal election. Similarly, while the party has not achieved gender parity in candidate nominations (Ouellet, Shiab, and Gilchrist Citation2021), the federal Liberal Party of Canada has appointed a gender parity cabinet since they formed government in 2015. In contrast, the federal Conservative Party of Canada has consistently viewed gender inequalities in political engagement and descriptive representation as an individual-level rather than systemic phenomenon. Indeed, while the Liberals and NDP ensure groups are represented in central party structures, the Conservatives do not (Cross, Pruysers, and Currie-Wood Citation2022, 164). The Liberals and NDP also vary here, as only the NDP guarantees internal party representation for LBGTQ2+ members, members with disabilities, and labor groups (Cross, Pruysers, and Currie-Wood Citation2022).

9. The full text of all vignettes is available in the appendix.

10. The short synopsis for each candidate gave their name, education level, occupation, and two accomplishments. For example, one might read "Rebecca Adams Education: B.A. (Economics) MBA Occupation: CEO of Advanced Strategies Accomplishments: Author of In the Market; 'Top 100 Canadian Business Owner.'"

11. We rely on a politically-oriented measure of explicit sexism as a continuation of Mo's (Citation2015) work on explicit and implicit sexism in politics. While we recognize the value of other scales such as the Ambivalent Sexism Inventory (Glick and Fiske Citation1996), we decided to use the shortened Mo scale because our concern is less with the specifics of hostile and benevolent sexism, but rather with how perceptions of women in power affect the candidate emergence process. Mo derived these measures from existing public opinion surveys as well as work from scholars who focused on gender roles and gender stereotyping. These measures ask respondents to specifically consider their beliefs about the "proper" role of women in society, with specific references to politics, homemaking, and financial support. We recognize the concern that some of these measures (in particular, the Equal Candidates question) may tap a similar concept to our dependent variable. To address this concern, we rely on the eight-item additive index, which covers a broad range of explicit attitudes about women's role in politics, society, and leadership positions.

12. In the past, scholars have differentiated between hostile and benevolent sexism (Glick and Fiske Citation1996, Citation2001) and, while we find this differentiation useful, here we rely instead on a scale designed to measure preferences for men over women in leadership positions.

13. All questions used in the explicit sexism scale can be found in the appendix.

14. Since respondents evaluated all four prospective candidates, we refrain from using language that implies a true experimental manipulation. However, since prospective candidate gender was randomized within the high-agency and low-agency vignette pairs, we can experimentally test whether women, as prospective candidates, were evaluated more positively (or negatively, dependent on sexism) than men as prospective candidates in the same vignette. When we perform this analysis, the results are consistent with the results presented here. We present these results in the appendix (Appendix Table A1) for interested readers.

15. Since we cannot experimentally manipulate a respondent's level of explicit sexism, we recognize that additional co-variates may be driving these effects. We re-ran the models with a host of political and demographic controls and included these in Appendix Table 2. The effects we discuss below are unchanged by the addition of controls.

16. While significant results emerge for explicit measures of sexism, we also measured implicit sexism using an IAT, which largely produced null findings. This is perhaps unsurprising, given that the relationship between implicit associations and conscious attitudes and behavior tends to be relatively weak (Greenwald, McGhee, and Schwartz Citation1998, Cameron et al. Citation2012). Nonetheless, we present the IAT stimuli (occupations and names), which were shown in French or English depending on the respondent's preference, in Appendix Table 3. Given the lack of clear expectations, we focus on the explicit sexism measure in our discussion of the results, though we replicate our overall analyses with the implicit measure included as well in Appendix Table 4. As noted, the implicit measures offered little predictive value in our models.

17. See footnote 7 above.

18. Range and mean data appear in Appendix Table 7.

19. In addition to their presence in federal politics, all three of these parties have a corresponding provincial party in most provinces, if not in name, then in party family. The exception is the province of Quebec, where party formation also follows cleavages defined by nationalism and sovereignty.