Amid increased attention toward legislator moderation and election reform, this paper examines the relationship between US primary election policies and electoral outcomes from 1976 to 2020. To accomplish this, I use a difference-in-differences approach to investigate whether adopting less restrictive primary systems impacts legislator extremism and voter turnout. I find that expanding ballot access causes legislator ideology to shift toward the median voter. This moderating effect is even more pronounced for newly elected representatives and is driven mainly by non-partisan primary systems. Over the same period, I estimate a decrease in general election participation following the adoption of "open-type" primary systems. I present evidence suggesting voter indifference as the likely cause of this negative effect, stemming from a smaller ideological gap between candidates. This paper offers a comprehensive view of primary election policies, underscoring the balance between enhancing representation and maintaining voter engagement.
Adult ADHD, Stimulant Medication, and Labor Market Outcomes
Treatment of Attention-deficit hyperactivity disorder (ADHD) --- a neurodevelopmental disorder associated with inattention and hyperactivity --- primarily consists of the prescription of stimulant medication. To study the relationship between stimulant medication and labor market outcomes in people with ADHD, I use linked employment and pharmaceutical data from the Medical Expenditure Panel Survey (MEPS) and leverage individual-level variation to estimate a two-way fixed effects regression. I find limited evidence to support a causal relationship between prescription behavior and employment, real wages, or weekly labor hours.
Does AI Facilitate Trust? An Experimental Study with ChatGPT (with Ethan Holdahl, Conner Weigand, and Jiabin Wu) (in progress)
In this study, we experimentally explore the impact of AI as a supportive tool for players in a two-player trust game. The game begins with the trustee sending a message to the trustor. In certain scenarios, the trustee is aided by the large language model (LLM) ChatGPT in composing this message. In other scenarios, the trustor uses GPT to interpret the message from the trustee, or both players may have access to GPT assistance. Our findings indicate that when the trustee utilizes GPT as a helper, it enhances cooperation with the trustor. Interestingly, this improvement in cooperation is not attributed to GPT's superior messaging skills. Instead, it appears that when the trustee has GPT's assistance, it encourages the trustor to scrutinize the trustee's message more closely, understanding that it could be genuinely crafted, a mixture of personal input and GPT suggestions, or solely generated by GPT. The detailed scrutiny by the trustor, and potentially the trustee's awareness of this scrutiny, aligns the beliefs of the trustor with those trustees who send either genuine or mixed messages, thereby fostering an environment that encourages the development of trust.