Abstract
Only half of SAT-takers retake the exam, with even lower retake rates among low-income students and underrepresented minority (URM) students. We exploit discontinuous jumps in retake probabilities at multiples of 100, driven by left-digit bias, to estimate retaking’s causal effects. Retaking substantially improves SAT scores and increases four-year college enrollment rates, particularly for low-income and URM students. Eliminating disparities in retake rates could close up to 10 percent of the income-based gap and up to 7 percent of the race-based gap in four-year college enrollment rates of high school graduates. (JEL I21, I23, I24, J15)
