The effect of violence on educational outcomes

This dissertation consists of three essays that explore the relationship between neighborhood crime and educational attainment. The first essay, entitled "The Effect of High School Shootings on Schools and Student Performance," analyzes how fatal shootings in high schools affect schools and students using data from shooting databases, school report cards, and the Common Core of Data. I examine schools' test scores, enrollment and number of teachers, as well as graduation, attendance and suspension rates at schools that experienced a shooting, employing a difference-in-differences strategy that uses other high schools in the same district as the comparison group. The findings suggest that homicidal shootings significantly decrease the enrollment of students in grade 9 (the high-school entrance grade) and test scores in math and English standardized tests. Using student-level data from California, I confirm that shootings lower test results for students that remain enrolled. I find no statistically significant effect of suicidal shootings on all outcome variables of interest and this suggests that the traumatic impact of homicidal shootings play a key role in explaining our results. The second chapter, "The Effect of Violent Crime on Educational Attainment," focuses on violent crime as an important determinant of educational attainment. I use two different models, a fixed effects model and an instrumental variable model to identify the causal effect of violent crime on educational attainment. First, a city fixed effects model with city-specific linear trends exploits different trends in violent crime among 229 major cities. The results indicate that the association between violent crime and grade 12 enrollment is significant while the same association with grade 9 enrollment is not. This difference suggests that student drop-outs drive the decline in enrollment associated with a high violent crime rate. Second, this study uses the implementation of the Electronic Benefit Transfer (EBT) program as an instrument to estimate the effect of violent crime on educational attainment. The EBT program implementation changed the method of payment for welfare recipients to debit cards from paper checks. This change lowered local crime rates by reducing cash flow and opportunities for crime in disadvantaged neighborhoods. Using the 1997 National Longitudinal Survey of Youth, two-stage least squares estimates show that exposure to violent crimes lowers educational attainment. These results provide evidence that non-educational policies can affect educational outcomes and shed light on the importance of coordinating educational and non-educational policies to enhance student educational outcomes. The final chapter, "The Effect of the Safe Schools/Healthy Students Initiative on Educational Outcomes," studies the effect of the Safe Schools/Healthy Students (SS/HS) Initiative, a community-level intervention, on educational outcomes. From 1999 to 2009, 365 local education agencies received the SS/HS Initiative grant for their proposed school-community partnership plans. I use two different approaches, a difference-in-difference model and a propensity score matching model, to identify the effect of the school-community partnerships on educational outcomes. First, the results from a difference-in-difference model indicate that the SS/HS Initiative increases grade 12 enrollment by 1.9 percent, respectively. Second, I use a propensity score matching model to match grantee local education agencies with similar, non-grantee local education agencies. The results from this model show that grantee school districts have at least 3.3 percent higher enrollment in all high school grades. These results provide evidence that school and community partnerships improve educational outcomes