Stressed About GPA? Uncertainty and Limited Academic Exploration in College (JMP)
Presented at: 2024 Midwest Economic Association, 2024 SOLE, Federal Reserve Bank of Chicago, 2025 AEFP

One third of the U.S. undergraduate students switch their major within the first three years of their college experience, which suggests a large uncertainty in ability students face when they arrive at college. This paper examines the impact of uncertainty in academic ability and the incentive to maintain a high cumulative GPA on college courses and major choices. The taste for maintaining a high GPA can prevent students from taking courses that maximize the amount of information obtained about their uncertain academic ability. To identify students' taste for maintaining a high GPA, I use a policy variation that expanded the pass/fail grading option during the COVID-19 pandemic at lage public universities in Texas. I document that students were 1.2% - 5.6% more likely to enroll in STEM courses when their school continued to offer the expanded pass/fail grading option, which reduced the risk of disclosing bad grades for the students. Based on the empirical facts, I develop a dynamic model of college course and major choices, incorporating factors such as taste for maintaining a high GPA and ability learning. I carefully model grading equations using the statistical relationship between the Binomial and Generalized Beta distributions to capture how grade inflation weakens grade signals about students' ability. The model is estimated to match the educational decisions of students at a large flagship state university in Texas. Counterfactual analysis highlights the implications of grading policies aimed at reducing uncertainty in academic ability and equalizing grading standards across departments, which can lower dropout rates and increase STEM graduation rate up to 22%.

The Effect of Bilingual Education: Long-run Student Outcomes and Spillover Effects (with Kevin Hunt)
Presented at: 2024 Midwest Economic Association (SOLE Session)

A large and growing share of students arrive at school as English language learners (ELL). In Texas, schools provide bilingual education or an English as a second language programs for these students. Using a regression discontinuity design based on a quasi-experimental policy variation in Texas, this paper examines the effect of exposure to bilingual education in 1st grade on students' test scores and college enrollment. We find no effect of bilingual education in first grade on elementary school standardized test scores but large significant effects on college enrollment for both ELL and non-ELL students: bilingual education increases four-year university enrollment for ELL and non-ELL students by 6.4 and 6.8 p.p., respectively. The positive effects for both student groups in the long-run suggests that bilingual education has positive effects that are not well measured by short-run test scores for ELL students, and these effects spill-over to the outcomes on non-ELL students.

Equity-Efficiency Implications of Output-Based College Resource Allocation (with Chao Fu and Jesse Gregory)