Date of Award

December 2020

Degree Type


Degree Name

Doctor of Philosophy (PhD)


Political Science


Shana K. Gadarian


candidate emergence, candidate recruitment, female representation, gendered civil society context, party strength, state legislatures

Subject Categories

Social and Behavioral Sciences


This dissertation examines the substantial and persistent variation in female representation observed across states and party caucuses in the US state legislatures. I develop and test the theory of gendered civil society context and party strength to explain why female candidates are more likely to run for office in some places rather than others. Gendered civil society context refers to the gender balance of civil society groups (e.g. unions, service organizations, advocacy groups, professional associations) in a party’s local strategic environment; party strength is defined as the capacity of political parties to influence candidate nomination.

Two main hypotheses guide the project. First, I hypothesize that when a party’s local gendered civil society context is comprised of female-dominated networks the emergence of female candidates is more likely compared to places where male-dominated networks wield more political power. Second, I expect that the relationship between gendered civil society context and female candidate emergence is moderated by party strength. I employ a multi-method approach to evaluate these hypotheses using regression analysis with nationwide data and case studies of Georgia and New York, based on 30 semi-structured interviews with Democratic and Republican county party leaders.

Taken together, the quantitative and qualitative results suggest that the theory of gendered civil society context does indeed provide a promising new explanation for candidate emergence. The quantitative results indicate that within both parties, more politically active female-dominated industry groups (e.g. teachers unions) are associated with more female candidates. Moreover, more politically active male-dominated industry groups (e.g. trade unions) are associated with fewer Democratic women running for office, but the same effect is not present on the Republican side.

Interviews generated rich data about how candidates come to run for office, and they indicate that gendered civil society contexts impact the likelihood of female candidate emergence in both parties. Party chairs in both Georgia and New York emphasized that prior community engagement is critical for successful candidacies. However, as expected, the types of engagement and organizations mentioned did vary across the states and parties in ways that point to gendered differences in candidate pipelines. Democratic chairs tend to mention professional and civic backgrounds of potential candidates that are more female-dominated (e.g. education, women’s rights) compared to those mentioned by Republican chairs (e.g. law enforcement, business owners). Moreover, the results suggest that candidate emergence within Republican networks is driven by personal connections within civil society groups in contrast to candidate emergence from the networks of advocacy and other stakeholder groups with strong institutional ties to Democratic party organizations. Finally, while potential candidates who come forward on their own (i.e. self-starters) are common in both parties, Republican chairs seem to rely more heavily on self-starters. This reliance on self-starters disadvantages women, who are less likely than men to run for office without being recruited.

In sum, the dissertation argues that the gendered civil society context represents an innovative conception of candidate pipelines, which acknowledges party-based and gender-based differences and accounts for a broad array of civil society actors that impact female candidate emergence.


Open Access