Skip to main content

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

A Correction to this article was published on 20 January 2022

This article has been updated


In this paper, we examine whether women candidates are more likely to spur turnout in election years when gender-related issues are central to the national debate. We argue that having women on the ballot in a gendered electoral environment mobilizes specific groups of voters. Utilizing voter files in Pennsylvania and Washington for 2014 and the more gender focused 2018 election, we evaluate this potential mobilizing effect in both primary and general midterm elections. Our results show that both female and male voters were more likely to turn out in the 2018 midterm elections when a woman was on the ballot for the U.S. House of Representatives. In Pennsylvania, which tracks registrants’ party affiliation, Democrats, members of third parties, and independents were particularly impacted by the presence of a female candidate. Moreover, in both states, a woman on the ballot was especially important for young people, a group that is traditionally less engaged. Utilizing a difference-in-difference approach, we confirm these results are not due to the endogenous selection of where women choose to run. These findings demonstrate that the mobilizing effect of women candidates is dependent on political context.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6

Change history


  1. For example, see FiveThirtyEight special on “When Women Run” (January 2020).

  2. Additional details are available in Supplementary Appendix E.

  3. Difference-in-difference was not performed for Pennsylvania because court ordered redistricting changed the district lines between 2014 and 2018.

  4. A total of 176 women were candidates for Congress in 2014, compared to 257 women in 2018 (CAWP, 2018c). The 2018 election resembles the 1992 election, dubbed the “Year of the Woman” for the surge in women candidates, particularly Democratic women, and focus on gender-related issues.

  5. Millennials are those born between 1980 and 1996. Generation Z encompasses those born between 1997 and 2012. Only a small portion of Generation Z (those born 1997–2000) were eligible to vote in 2018.

  6. A perspective also seen in the smaller group of adults that are part of Generation Z (Parker et al., 2019).

  7. For an assessment of state polling accuracy see Guskin and Santamariña (2020). For additional details on differences between survey and voter file data as well as an analysis showing how survey data yields inaccurate estimates in this context, see online Appendix E.

  8. The voter files contain key attributes of individual registrants including vote history, gender, age, and (for Pennsylvania) party registration.

  9. Due to the maintenance of VFs as snapshots in time, we use separate VFs for the 2018 and 2014 models. 2018 models use a January 31, 2019 version of the Washington VF and a February 18, 2019 Pennsylvania VF. 2014 models use a December 2014 Washington VF and a February 6, 2017 export of the Pennsylvania VF. Ideally, we would have preferred a 2014 version of the Pennsylvania VF, unfortunately the state does not maintain old versions and the 2017 version was the oldest the authors had in their possession. Luckily, the 2017 version had voter history for the past 40 elections and district designations prior to the 2018 court ordered redistricting that altered district lines.

  10. Data and replication codes for all analyses are available at the Political Behavior Dataverse page:

  11. This is done by computing the predicted probability of voting when female candidate = 1 minus the probability of voting when female candidate = 0 with all constituent interaction terms set accordingly. See Brambor et al. (2006).

  12. While we include every House primary, regardless of whether there is a female on the ballot, in the case of the 2014 Democratic primary, we exclude districts 15 and 18 because no candidates ran for their party’s nomination in those districts. In these excluded districts, no House candidate, male or female, would have impacted turnout.


Download references


We thank Michael Hanmer, Janelle Wong, Patrick Wohlfarth, Will Bishop, Tiago Ventura, Sebastian Vallejo Vera, and the participants at the University of Maryland American Politics Workshop for helpful feedback on earlier versions of this paper. We also thank Jennifer Wolak, Christina Wolbrecht, Melissa Deckman, and Daniel Smith for helpful conversations about this project. Thoughtful comments from Political Behavior editor Geoffrey Layman and the anonymous reviewers challenged us to improve the manuscript. Any errors or omissions remain our own.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Stella M. Rouse.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (DOCX 743 kb)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Safarpour, A.C., Gaynor, S.W., Rouse, S.M. et al. When Women Run, Voters Will Follow (Sometimes): Examining the Mobilizing Effect of Female Candidates in the 2014 and 2018 Midterm Elections. Polit Behav 44, 365–388 (2022).

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:


  • Female candidates
  • Gendered electoral context
  • Midterm elections
  • Mobilization
  • Voter turnout
  • Young voters
  • Voter files