A Test of Three Explanations for Gender-Office Congruency in US Elections

• In Event: The Gendered State and Women's Equality

Abstract

Women continue to be underrepresented in political office in the U.S. and other countries. There are multiple reasons for the underrepresentation. On the “supply side,” explanations include differences between potential male and female candidates in ambition for public office and access to resources to support their campaigns. We focus on the “demand side” to understand if voters intentionally or inadvertently discriminate against female candidates at the ballot box. Building on prior studies, we seek to understand why the nature of voter bias appears to vary in both magnitude and direction depending on the office being sought.

We consider this question within the theory of gender-office congruency. The theory suggests that voters associate the demands of specific political offices with gender stereotypes. Male candidates are thought to be advantaged for executive offices that emphasize unilateral leadership while female candidates are be more favored for legislative offices that emphasize cooperation with other politicians. Although the logic is intuitive, the theory has mostly been applied to candidates running for president or Congress, two federal offices that both have responsibility for foreign affairs, a stereotypically male area of competence. In addition, this focus conflates congruency with experience by contrasting an office that has never been held by a woman with an office that many women have held.

To unpack the theory more fully, we consider several hypotheses. First, voters may discriminate more against female candidates when they are seeking executive office rather than legislative office. Second, voters may discriminate more against female candidates when they are seeking an office with national security and foreign policy responsibilities than when seeking an office with a purely domestic focus. Third, voters may discriminate more against female candidates when they are seeking offices in which few if any women have served than when seeking offices that have been regularly held by women.

We resolve these three explanations for gender-office congruency by conducting a conjoint experiment in a national survey whose sample (N = 3,000) is drawn from YouGov’s panel. The survey asks each respondent to evaluate multiple pairs of hypothetical candidates for office. The conjoint experiment is appealing because it allows us, via randomization, to control a dozen other factors such as candidates’ policy positions, levels of experience in office, and character traits that voters might believe are otherwise correlated with candidate gender. The experiment reveals the degree of baseline gender bias after controlling away stereotypes about candidates that might lead to inadvertent favoritism toward male or female candidates.

To test the first hypothesis, the survey asks respondents to choose among pairs of candidates for four different offices: president, member of Congress, governor, and state legislator. If voters penalize female candidates when they run for executive office, then the analysis will show larger biases when women run for president or governor. To test the second hypothesis, we compare the estimated gender bias when a candidate run for president (the office most responsible for foreign affairs), Congress (an office with some responsibility for foreign affairs), and both governor and state legislator (neither of which have responsibility for foreign affairs). If voters penalize women seeking offices with a foreign affairs focus, then the bias should be largest in presidential elections and smallest in gubernatorial and state legislative elections. Finally, to test the third hypothesis, after the experiment we ask respondents for their perceptions of the prevalence of women in elective office. Specifically, we ask whether they recall being represented at any point by a female member of Congress or state legislature. We also ask them to estimate what percentages seats in the House of Representatives and their state legislature are held by women. If the exposure hypothesis is correct, then the degree of gender bias ought to be largest for respondents who do not remember being represented by women and/or believe that fewer offices are held by women.

The results of the study are valuable because they help to resolve ambiguity in a prominent explanation for the underrepresentation of women in public life. In addition, they point to when factors are most likely to result in the increasing prominent of women in elective office, especially those with executive responsibilities.

Authors

• Barry C. Burden, University of Wisconsin

• Yoshikuni Ono, Tohoku University