Essays on College Student Outcomes and Inequality

Abstract

This dissertation consists of three essays that estimate the impact of policies and practices on college student outcomes and inequality in higher education. The first essay shows that salient and accessible signals about academic qualifications affect college enrollment choices. After highly selective four-year colleges published GPAs that could guarantee transfer admissions, high-GPA community college graduates increased flagship enrollment by 32 percent. Increased transfers to flagships over less selective regional and private colleges also accompanied higher graduation rates and lower student debt. Gains were largest for students with historically lower transfer rates. These findings suggest that publishing transparent college admissions standards can be a low- to no-cost strategy for increasing access to selective colleges. The second essay estimates the impact of faculty on college students’ long-term achievement. I use statewide data from Texas and Virginia to study whether students assigned to more effective faculty in developmental courses are more likely to graduate on time and have higher earnings. Using residual pass rates as a measure of faculty value-added, I find that replacing a community college instructor in the bottom 5 percent with an average instructor boosts academic persistence, college completion, and earnings by 0.02-0.13 SD. Impacts on long-run outcomes are mediated by short-run academic indicators such as reduced course withdrawal, improved follow-up course performance, and increased credit accumulation. Faculty value-added is decreasing in the number of courses and students taught each semester, and is less related to demographic characteristics such as faculty gender, race, education, and age. These results suggest that efforts to improve faculty hiring, assignments, or working conditions may affect student outcomes. The third essay evaluates a program to reduce faculty’s biased behaviors and identifies potential channels through which faculty can affect student outcomes and gaps. Although more colleges have invested in unconscious or implicit bias training for college faculty to reduce achievement gaps by race and gender, little is known about their impacts on faculty choices and student performance. Through a field experiment with over one thousand faculty members at four colleges, we estimate the impact of de-biasing tool that introduces faculty to the concept of implicit bias, offers faculty a chance to understand their own biases through the implicit association test, and understand evidence-based teaching and advising practices to implement. This paper presents pilot evidence from one community college during Fall 2020, and shows that the program increased the use of blind grading among faculty. Second, faculty’s average gap in course grades between white and non-white students does not change significantly in the immediate semester, but then decreases by half in the following year. Finally, there is suggestive evidence that persistence and completion increase. This is the first study to our knowledge that estimates the impact of de-biasing teaching tools on faculty teaching choices and student outcomes.