Updating Human Capital Decisions: Evidence from SAT Score Shocks and College Applications

Abstract

We estimate whether students update their college application portfolios in response to large, unanticipated information shocks generated by the release of SAT scores -a primary component of admissions decisions. Exploiting new population data on the timing of college selection and a policy that induces students to choose colleges prior to taking exam, we find that the release of scores causes students to update their portfolios in terms of selectivity, tuition, and sector. However, the magnitude of updating is too modest to significantly reduce unexplained variation across students, suggesting that non-academic factors may be the dominant determinants of college choice.

High Schools and Students' Initial Colleges and Majors

“…First, they point toward the value of interventions that inform students of the educational options for which they are academically qualified, which can better align students from low‐SES high schools with universities (Hoxby and Turner ). Heterogeneity in student preferences ensures that under‐ and over‐placements to universities will occur, especially since nonacademic factors also play an important role in determining the college match (Bond et al Forthcoming). However, the systematic relationship between under‐placement and student SES we document is disconcerting in light of evidence that more‐selective institutions, as measured by the academic qualifications of entering students, improve educational outcomes (Arcidiacono and Koedel ; Cohodes and Goodman ; Hoekstra ; Melguizo ) .…”

Take Two! SAT Retaking and College Enrollment Gaps

“…10 We use data from from Chetty et al (2017) to characterize the expected income of each college's initial enrollees at ages 32-34. 11 We also observe all colleges to which a student ultimately sends SAT scores, a decent proxy for college application behavior given that many colleges require standardized test scores as part of their admissions process (Pallais, 2015;Bond et al, 2018;Smith, 2018). Table 1 shows mean characteristics of the full sample, the subset who first took the SAT by November of senior year and thus comprise our regression discontinuity sample, and three subsamples of interest: lower scoring students, defined as those whose first SAT score is closest to or below the median threshold of 1500; low income students, defined as those with self-reported family income below $50,000; and underrepresented minority (URM) students, defined as those who report their race/ethnicity as Black, Hispanic or Native American.…”

Can Better Peers Signal Less Success? The Disruptive Effect of Perceived Rank on Career Investment

“…Although candidates are unaware of the cutoff between classes when they apply to the university, they could decline the offer once the test scores and class order are revealed (Bond et al, 2017). To verify this type of selection bias, we run the McCrary's (2008) test and find no evidence of missing students on one side of the threshold.…”