The Effects of Exposure to Women Candidates on Political Attitudes

Download data and study materials from OSF

Principal investigators:

David Campbell

University of Notre Dame



Christina Wolbrecht

University of Notre Dame



Sample size: 1000

Field period: 10/03/2020-03/23/2021


Note: This experiment parallels our other TESS study "The Effects of Exposure to Women Candidates on Adolescents' Political Attitudes," which was among adolescent respondents instead of an adult sample. Accordingly, much of the content here is identical to the adolescent study.

From Hillary Clinton’s cracked ceiling to Elizabeth Warren’s pinky promises, women candidates often frame their candidacies as opportunities to advance the political empowerment and engagement of women. Advocates argue that more women political leaders signal a more open, fair, and representative political system and provide evidence of the capacity of women for political leadership to women and men alike. In the U.S., some observational studies have found that the presence of women candidates is associated with greater political engagement among women, although findings are mixed. However, previous research has paid less attention to the impact of women politicians on beliefs about democracy or about women as political leaders. In this paper, we employ a survey experiment to gauge the impact of exposure to women candidates on not only political engagement, but attitudes about the responsiveness of the political system and beliefs about the political capacities of women. In addition to questions about gender, we also address the role of partisanship in mediating these effects; in a time of severe partisanship, co-partisan women may have a bigger impact on adolescents than out-partisans. Like the previous observational studies, our experimental findings are mixed. The most consistent finding—consistent with other experimental research on adolescents—is that women candidates lead men to be more likely to view women as possessing leadership traits, and to think that there ought to be more women in elected office.

  1. 1. People who are exposed to women candidates who have run successful campaigns will:
    (a) express interest in being politically engaged and active
    (b) view the political system as responsive
    (c) hold egalitarian attitudes on gender roles, including that women are capable of political leadership.
  2. 2. These effects are greater for women; i.e., there is greater attitude change among women than men.
  3. 3. These effects are greater for co-partisans. That is, Democrats are more likely to respond when exposed to Democratic women candidates; and Republicans have a greater response to Republican women candidates.
Experimental Manipulations

The treatment consisted of a fictional news story highlighting the large number of women who ran in the 2020 elections across the country, and for many different offices.

Subjects were randomly assigned to one of three versions of the treatment or to a control group that received no story, in order to observe these attitudes in the absence of any information about the 2020 elections. In brief, the three versions of the story are:

  1. 1. Non-Partisan Story: The story did not specify whether the surge of women candidates were Republicans or Democrats. The party affiliation of the two candidates quoted in the story was not mentioned.
  2. 2. Democratic Story: Text was identical to (1), except that the story emphasized the number of Democratic women who ran in 2020. The two candidates quoted in the story were identified as Democrats.
  3. 3. Republican Story: The text was identical to (2), except that the emphasis was on Republican women, and the candidate quotes were attributed to Republicans.

The treatments all mention that these women candidates are bringing “new voices to politics,” with one candidate quoting as saying that she will bring a “fresh perspective to Washington.” Likewise, to ensure plausibility for both parties the story spoke generally of women running for offices at all levels of government.

As visual reinforcement the story included three photos of fictional women candidates with the caption: “Three candidates/Democratic candidates/Republican candidates running for seats in Congress, joining a national surge of women candidates across the country.” Also, both a headline and pull-put box underscored the theme of a “surge” in women candidates in 2020. The headline of the story was “A Big Year for Both Democratic and Republican Women/Democratic Women/Republican Women in the 2020 Elections.” The pull-out box contained a quote from a (fictional) woman candidate: “Women like me will get to work and make a difference for our communities and the rest of the nation,” with the candidate identified as “Charlene Washington, newly-elected Democratic/Republican member of Congress.” The text of the story included still more quotes, one from a winning woman candidate for the House of Representatives (“Amy Johnson,” from Michigan) and another from a scholar (“Maria Gonzalez”) commenting on the historic number of women running for office in recent years.

Political Engagement

Even if you’ve never thought about it, could you see yourself doing any of the following things in the future?

Please rate how likely you are to do any of the following in the future using a scale of 0 to 10 where 0 is not likely at all and 10 is very likely.
(order randomized)

  1. A. Volunteer in my community
  2. B. Vote in a public election
  3. C. Work in a political campaign
  4. D. Participate in a lawful demonstration
  5. E. Write to public officials
  6. F. Run for political office
In the analysis, items C-F are combined into an additive index, as exploratory factor analysis indicates that they form a common factor. Because they do not load well with the other items, "vote in a public election" and "volunteer in my community" are modeled separately.

Political Responsiveness

Using the slider, please rate how much you agree or disagree with the following statement using a scale of 0 to 10 where 0 is disagree and 10 is agree.
(order randomized)

  1. A. The political system helps people with their genuine needs
  2. B. Politicians do not listen to people like me
  3. C. I can make a difference solving problems in my community
  4. D. Democracy is the best way to run this country

Gender Roles

Please use the slider to finish the statement below using a 0-10 scale where 0 is worse, 5 is the same, and 10 is better.

If we had more women in elected office, America would be…
Worse . . .. The same . . . Better

Leadership Traits

Many people think men and women political leaders have different strengths. Others think that there are no gender differences among our elected officials. What do you think?

In general, do you think men or women in high political office are better at each of the following, or is there no difference? Please use the slider below, using a 0-10 scale where 0 is men are generally better, 5 is no difference, and 10 is women are generally better.
(order randomized)

  1. A. Working out compromises
  2. B. Being a strong leader
  3. C. Being honest and ethical
  4. D. Working well under pressure
Items A and C are combined into an index of stereotypically feminine leadership traits, while B and D are similarly combined into a masculine traits index.

Sexist Attitudes

Using the slider, please rate how much you agree or disagree with the following statement using a scale of 0 to 10 where 0 is disagree and 10 is agree.
(order randomized)

  1. A. It is usually better for everyone involved if the man is the achiever outside the home and the woman takes care of the home and family
  2. B. Discrimination against women is no longer a problem in the United States
  3. C. Women should be cherished and protected by men
  4. D. Most women interpret innocent remarks or acts as sexist
Items A,B, and D are combined into a Non-Sexism Index, coded so that higher values correspond to less-sexist attitudes.. Item C (benevolent sexism) does not load with the other items and is thus modeled separately.

Summary of Results

Here we summarize the findings, but for the complete results (including the magnitude of effects and standard errors), please consult the table.

In the following description the non-partisan treatment refers to the news story that describes a surge of women candidates in both the Republican and Democratic parties, while the co-partisan treatment refers to how subjects respond when exposed to a story about women candidates from their party (e.g. Democratic subjects, Democratic candidates). Note that the partisan groups (Republicans, Democrats) include partisan leaners, or those subjects who initially indicate that they are independent but, upon follow-up, acknowledge leaning toward one party.

Political Engagement

We do not observe any effects for the non-partisan treatment on either women or men.

For the co-partisan treatments, results are mixed. Democratic women are more likely to vote when exposed to Democratic women candidates (p < .01), while both Republican men (p <.01) and Democratic men (p <.05) are less likely to vote when exposed to co-partisans running for office. Democratic women are less likely to volunteer when Democratic women run, while co-partisan candidates lead Republican men to be more likely to volunteer.

Political Responsiveness

The non-partisan treatment causes women to be more likely to say that politicians listen to people like them (p <.05), but the non-partisan treatment has no other effects on either men or women.

Co-partisan candidates lead Democratic women (p <.05) and Democratic men (p <.10) to be more likely to say that politicians listen to people like them. It also leads Democratic women to think that democracy is the best way to govern the country. However, co-partisan women cause Democratic women to be less likely to report that they can make a difference in their community, as well as a decrease in the belief that democracy is the best way to govern the country among both Republican and Democratic men.

Gender Roles

The non-partisan treatment—women candidates of both parties—leads men in general to say that America would be better off with more women, as well as score higher on indices of masculine and feminine leadership traits, respectively (both p <.10).

Republican men are more likely to say that women have masculine leadership traits when exposed to a co-partisan treatment (p <.05). A story about co-partisan candidates also leads Democratic women to score higher on the non-sexism index (p <.05).


Results Summary Table