The Effects of Local Violent Crime on High-Stakes Tests

Random shocks to cognitive performance on high-stakes standardized tests have long-lasting consequences, particularly when test results are used as the sole mechanism to determine school admissions. This study considers the effects of exposure to local violent crime on high-stakes standardized test outcomes in the context of Mexico City’s centralized high school admission system. To do so, we exploit within-school variation in exposure to local violent crime over time. Our results show that exposure to violent crime reduces test scores for female students but not for males, leading to a gender biased high school placement. That is, female students’ test scores decrease by 11 percent of a standard deviation after the exposure to violent crime occurring within 0.1 miles of their school during the week before the test, and approximately 19 percent of those students are assigned to less-preferred high schools than the ones to which they would have been assigned otherwise. The effect is highly localized both in time and geographic proximity, suggesting that temporary psychological harm is one of the main mechanisms through which exposure to violent crime affects cognitive performance.