The Effects of Descriptive Representation in American Legislatures: Evidence from Committee Hearings and Legislative Effectiveness

One of the most significant shifts in Congress over the past 30 years is the growing diversity of its members. In 1980, fewer than 20 nonwhite legislators served in the House and Senate; today, that number exceeds 130. Women’s representation has followed a similar trajectory, rising from just over 20 in the 1980s to more than 150 today. LGBTQ+ lawmakers, who until 2011 never exceeded three seats per term, now hold 13 seats. And while working-class candidates are rarely elected to Congress, they do occupy nearly 6% of state legislative seats.

Importantly, these descriptive representatives are not just rank-and-file members—they occupy the most consequential leadership positions in Congress. Nancy Pelosi (D-CA) became the first woman Speaker of the House in 2007 and served four nonconsecutive terms, making her one of the longest-serving Speakers in history. Under her leadership, Congress passed landmark legislation, including the Affordable Care Act, Dodd-Frank Wall Street Reform, the American Rescue Plan, and the Bipartisan Infrastructure Law. Women and nonwhite legislators have also taken on key committee leadership roles. In the 117th Congress, they chaired 12 of the 20 standing House committees, including two of the most powerful economic committees— Appropriations, led by Rosa DeLauro (D-CT), and Financial Services, led by Maxine Waters (D-CA). These shifts signify more than just increasing diversity in Congress; they show that descriptive representatives now play a central role in shaping policy and wielding institutional power.

How, then, have descriptive representatives—and their positions in leadership—shaped policymaking in American legislatures? In this dissertation, I focus on the effects of nonwhite and working- class representation in the U.S. Congress and state legislatures. In doing so, I extend the descriptive representation literature with three key contributions: two theoretical and one methodological. First, I extend intergroup contact theory from social and political psychology to legislatures (Allport, Clark and Pettigrew 1954; Pettigrew 1998; Lollis 2024a). Given that the institutional design of legislative committees is well-suited to facilitate contact effects, I argue that sustained cross-group interaction between white and nonwhite legislators influences how white lawmakers discuss race in hearings. Specifically, white legislators in racially diverse committees not only mention race more often but also incorporate more evidence into their race-based discussions.

Second, I expand Anzia and Berry’s (2011) theoretical framework on electoral bias and legislative effectiveness to an understudied minority group: working-class state legislators. Because fundraising and party recruitment biases discourage working-class candidates from running, those who do win tend to be exceptionally qualified and politically skilled. As a result, despite structural disadvantages rooted in class bias, working-class state legislators are just as effective as their white-collar peers at advancing and passing legislation.

Alongside these theoretical contributions, I address a methodological limitation in the study of descriptive representation. Most research in this area relies on sponsorship, cosponsorship, and roll-call vote data. Committee hearings, in contrast, are speech-based and substantively rich. However, they have largely been inaccessible to scholars due to the sheer volume of hearings and the challenges associated with operationalizing measures from text. To fill this gap, I analyze 1.4 million committee hearing statements spanning 25 years and develop a systematic method—a bigram dictionary classification approach—to identify references to race. Unlike prior studies that focused on select committees within a single term (Gamble 2007, 2011), this method allows for the creation of systematic, highly accurate measures of identity references in legislative speech. While this dissertation applies it to race, the methodological approach can also be extended to study discussions of gender, sexual orientation, and class.