We construct a time series of college attendance patterns for the United States and document reversal: family background was better predictor before World War II, but academic ability afterward. model choice that explains this reversal. The model’s central mechanism is an exogenous surge leads colleges to be oversubscribed, institute selective admissions, raise their quality relative peers, as ...

We use the introduction of Texas Top Ten Percent rule to estimate effect access a selective college on graduation and earnings outcomes for two groups students. For highly ranked students at more disadvantaged high schools, who gained under policy, enrollment increased. Less advantaged tended lose access, shifted toward less-selective colleges but did not see declines in overall enrollment, gra...

We document that college completion rates have increased since the 1990s, after declining in 1970s and 1980s. find most of increase graduation can be explained by grade inflation other factors, such as changing student characteristics institutional resources, play little or no role. This is because GPA strongly predicts graduation, GPAs been rising 1990s. finding holds national survey data reco...

This paper examines if students’ college outcomes are sensitive to access to college admissions tests. I construct a dataset of every test center location and district policy in the United States linked to the universe of individual testing records and a large sample of college enrollment records. I find evidence that SAT taking is responsive to the opening or closing of a testing center at a s...

We study the causes of sustained participation in political movements. To identify persistent effect protest participation, we randomly indirectly incentivize Hong Kong university students into an antiauthoritarian protest. role social networks, randomize this treatment’s intensity across major-cohort cells. find that incentives to attend one within a movement increase subsequent attendance but...

About 21% of U.S. college students are overweight. However, aside from the 1995 National College Health Risk Behavior Survey (NCHRBS), the prevalence of health-risk behavior among college students has not been well characterized. The objective of this study is to investigate how college student’s body fruit and vegetable intake is affected by (1) demographics, (2) dietary habits, and (3) lifest...

Endowment payouts have become an increasingly important component of universities’ revenues in recent decades. We study how universities respond to financial shocks to endowments and thus shed light on a number of existing models of endowment behavior. Endowments actively reduce payouts relative to their stated payout policies following negative, but not positive, shocks. This asymmetric behavi...

Although in-person college access programs can be effective, less is known about whether low-cost and scalable virtual interventions achieve the same benefits. We evaluate two variants of a counseling program. Students randomly assigned to program felt more supported applying applied broadly four-year colleges, but were not likely accepted or enroll. analyze rich extensive survey data explore m...

This paper provides new evidence on gender bias in performance evaluations of university teachers. We exploit a quasi-experimental dataset on 19,962 teaching evaluations, where students are randomly allocated to female or male teachers. Despite the fact that neither students’ grades nor self-study hours are affected by the teacher’s gender, we find that in particular male students evaluate fema...

Rising college enrollment over the last quarter century has not been met with a proportional increase in college completion. Comparing the high school classes of 1972 and 1992, we show declines in college completion rates have been most pronounced for men who first enroll in less selective public universities and community colleges. We decompose the decline into the components due to changes in...