The Gendered Politics of Congressional Elections

Are female candidates less likely than male candidates to attract votes or win elections? We conduct a large-n longitudinal analysis employing survey and observational data from every two-party congressional race over a 12-year period (2006–2018) and connect individual-level theory and evidence with aggregate-level results. We demonstrate that candidate gender significantly influences congressional vote-choice and election outcomes. Holding other variables constant, we show that male Republican and male independent voters are significantly less likely to vote for female Democratic candidates, but do not assess a similar penalty on female Republican candidates. Perceived ideological distance does not explain the lack of support for female Democrats—however, variation in candidate quality does: Female Democratic candidates can attract the support of male Republican and male independent voters when they have a qualifications advantage, but are penalized when they are merely “as qualified.” At the aggregate-level, female Democratic candidates with a qualifications advantage are as likely as males to win elections; but are significantly less likely than males to win when qualifications are held constant. The proportion of male Republicans and male independents in a district determines the extent of the penalty, with women’s electoral prospects declining as this proportion increases. Women can win, but they need to be highly qualified and strategic about the races in which they emerge. These findings contribute to our understanding of the micro- and macro-level factors that shape women’s electoral fortunes; and advance the goal of representational equality by helping candidates and campaigns concentrate their efforts on the most winnable voters and districts.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Subscribe and save

Springer+ Basic €32.70 /Month

Buy Now

Price includes VAT (France)

Instant access to the full article PDF.

Rent this article via DeepDyve

Similar content being viewed by others

When Women Run, Voters Will Follow (Sometimes): Examining the Mobilizing Effect of Female Candidates in the 2014 and 2018 Midterm Elections

Article 07 January 2022

Ideology and Gender in U.S. House Elections

Article 28 September 2018

Voting for women in recent Canadian elections

Article 05 February 2024

Notes

Previous research suggests that women voters can also be sexist (Cassese and Barnes 2018). For instance, white women may adopt sexist beliefs in order to gain acceptance, privilege, and protection from men. Our argument is not that women voters will not be sexist at all, but rather, that they are less likely than men to be sexist due to personal experiences with discrimination and self-interest in gender egalitarian policies.

To be clear, our argument about women’s quality advantage stems not from any kind of “natural superiority” of women candidates. Rather, our theory is rooted in differential competitive pressures faced by male and female candidates, and gendered self-selection.

Ansolabehere (2010a, b, 2012); Ansolabehere and Schaffner (2013, 2017); Shaffner and Ansolabehere (2015); Shaffner et al. (2019).

It is worth noting that this period covers times of good and bad fortune for both parties. Democrats dominated the elections of 2006, 2008, 2012 and 2018; but Republicans fared better than Democrats in 2010, 2014 and 2016. Only for four years (2008–2010 and 2016–2018) was there unified party control of government. Across the period under study, neither party had a distinct advantage.

Uncontested races where the incumbent ran unopposed, and races where two co-partisans ran against each other (e.g. California’s “top-two” primary system occasionally produces general elections with either two Democrats or two Republicans running) are excluded from the analysis.

Online Appendix A provides the summary statistics for all of the variables included in our individual-level models.

We omit respondents who voted for third-party candidates, since we are interested in the two-party vote.

Choosing to code the dependent variable in the direction of Democrats is arbitrary as the two-party vote share adds to 100. Changing the dependent variable to the Republican’s vote-share simply flips the sign of the coefficients in the model, but does not alter the coefficient’s size, significance or any of the other model statistics except for the intercept.

Male (0) or Female (1).

For ease of interpretation in our interactive models, we collapse the seven-point partisanship variable into three categories: − 1 = Strong/Weak/Lean Democrat, 0 = Pure Independent, 1 = Strong/Weak/Lean Republican. Our results remain unchanged if we include partisan leaners as independents.

White (1) or non-white (0). Married (1) or not married (0). Employed full-time (1) or not (0). College-educated (1) or not (0). Extremely Liberal (1) to Extremely Conservative (7).

To see whether the gap in relative spending or relative experience contributes more to female Democrats’ disadvantage, we estimate two separate models: one with only relative spending, and one with only relative experience. The female Democratic candidate coefficient emerges as negative and statistically significant in both the models: ß = − 0.13 (p < 0.001) in the model with only relative spending and ß = − 0.05 (p < 0.10) in the model with only relative experience. Although the two variables are significantly correlated (r = 0.74, p < 0.001), they each contribute unique variance when they are included together as shown in Models 2–4 in Table 1.

It is notable that the number of cases drops significantly (from 195 K respondents to a little over 120 K) when the perceived ideological distance measure is included. This is because a large portion of respondents could not place the Democratic candidate on an ideological scale. To evaluate whether these omitted cases were introducing bias, we ran Model 3 without the perceived distance variable. Our results are unchanged when we omit perceived distance: MR + MI voters are significantly less supportive of female than male Democratic candidates. Online Appendix B reports this analysis in greater detail.

An alternative explanation for our findings is that female Democrats are penalized for being more racially diverse than female Republicans. To fully answer whether this question, we would need to know the racial identity of every candidate, which we do not have. As a preliminary test of the racism hypothesis, we evaluate whether white voters are more punitive towards female Democrats than the full sample of voters, but we find that this is not the case. See Online Appendix C for the full analysis.

For example, scholars of micro- and macro-level partisanship debate about the nature and stability of partisan attachments: Green et al. (2002); Erikson et al. (2002)

Uncontested races where the incumbent ran unopposed and races where two co-partisans ran against each other are excluded from the analysis.

Online Appendix D provides the summary statistics for all of the variables included in our models.

References

Author information

Authors and Affiliations

  1. Department of Political Science, Texas A&M University, 2936 Research Parkway, College Station, TX, 77845, USA Sarah A. Fulton & Kostanca Dhima
  1. Sarah A. Fulton