Source: Dissertation Abstracts International, Volume: 75-09(E), Section: A.

Advisers: Alan Gerber; Gregory Huber.

Through the collection of novel data, new measurement strategies, survey experiments, and observational research, this dissertation considers the constraints that men and women face when considering a career in American politics, in the same light as those faced when choosing another career. Rather than explaining the gender gap in office-seeking as resulting from differences between men and women in personality or motivation, as does much of the current literature in political science, I consider how gender inequalities in the home contribute to the gender disparity in office-seeking. The first two essays focus on how work-life balance considerations affect the decision of men and women to run for public office. The third examines how voters update their beliefs about politicians when they receive new gender-coded information.

In Part 1, I survey a nationally representative sample of undergraduates, along with a sample of Yale undergraduates, and examine their desire to run for elected office, as an option among a variety of careers and life choices. I study students, because they still have a degree of choice about their future that people further into their careers lack. This allows me to study gender differences in career choices before they are constrained by later life experiences. I find that women students express less interest in inflexible careers, including those in politics, and that students who are more concerned with issues of family and workplace flexibility are less interested in running for political office. This suggests that gender disparities in government may be rooted in deep-seated differences in life plans and expectations between men and women. In addition, through qualitative research on the child-care center in Juneau, Alaska, which sets aside spaces for the children of state legislators, I find suggestive evidence that providing child care for state legislators would make it easier to recruit women to run for office.

In Part 2, I use two empirical strategies to measure the effect of travel distance on the decision to run for office. The expectation that women be primarily involved in childcare is deeply ingrained, and demonstrably affects whether women choose careers with shorter commute times. Political careers often require considerable travel, and therefore work travel preference may provide a way to measure how work-life balance preferences implicate the decision to run for office. First, I run a survey experiment in which subjects are asked to decide between a hypothetical career in the state legislature or in Congress. They are told that the national capital is five hours from home and that the state capital is either five hours or fifteen minutes from home. I find that women are twice as likely as men to forgo a hypothetical career in Congress if they can opt for a career in the state legislature that is located closer to home. Second, I use geographical and census data to determine whether, all else equal, districts further from state capitals are less likely to have women representatives. I calculate the distance from the mean center of population of every lower-house state-legislative district in the continental United States to its state capital. I find that, all else equal, districts further from state capitals are less likely to have women candidates for and women representatives in the state legislature, suggesting that the gender imbalance in representation may be deeply rooted in structural issues.

In Part 3, I examine whether people update their beliefs about men and women politicians according to Bayes' Rule and, if not, whether there are systematic differences in how voters update their beliefs about men and women politicians. Using a novel web application, I measure subjects' beliefs about the mean ideology of a hypothetical politician along a continuum, as well as subjects' 95% confidence intervals. I experimentally manipulate the party and gender of the politician, as well as the information that subjects receive about the politician. This allows me to test how a politician's gender and new information affect the location and strength of subjects' beliefs about a politician's ideology. Among other things, I find that Democratic women politicians, who make statements that violate party and gender stereotypes, drastically change voters' beliefs about them, and these beliefs change in a manner inconsistent with the normal-normal model of Bayes' Rule. This suggests that Democratic female politicians, in particular, may have to moderate their statements to avoid provoking drastic changes in voter opinion.