Although past research has found links between violent environments and student-level academic outcomes (Henrich et al., 2003; Margolin & Gordis, 2000), relatively little work has examined how community violence and the schools embedded within these neighborhood contexts might reciprocally interact over time.  The present study uses a longitudinal, cross-lagged panel analysis to explore the ways that neighborhood crime, school climate, and school academic achievement relate to one another across time within a sample of predominantly low-income, public elementary schools in Chicago.  Understanding the ways that these processes influence one another is critical for informing a growing body of policy and intervention efforts that aim to improve student outcomes through affecting higher-order school and neighborhood functioning.

For the present study, data on all public elementary schools  in Chicago (n=500) were collected from the Chicago Public Schools Office of Performance Research database for the 2007, 2008, and 2009 school years.  School climate was measured using the Student Connection Survey, which assessed students’ perspectives of their schools’ safe and respectful climate (SRC), socioemotional learning environment (SEL), and academic rigor (AR; Osher, Kendziora, & Chinen, 2008).  School-level academic achievement was represented by the percentage of students from each school who met or exceeded standards on Illinois Standards Achievement Tests (ISAT) in math and reading for each academic year.  Neighborhood crime was operationalized using spatial identifiers (latitude and longitude) for individual “index” crimes collected from the Chicago Police Department. These were geocoded using ArcGIS software and spatially aggregated within census tracts (n=400) to form estimates of school neighborhood crime for each year. 

To understand the relationships between neighborhood crime, school climate, and school academic achievement, a set of cross-lagged panel analyses was conducted within a structural equation modeling framework using Mplus.  All models controlled for school poverty, accounted for schools’ nesting in neighborhoods, and included within-year residual correlations.  A model testing the direct relationships between neighborhood crime and school academic achievement revealed that higher levels of neighborhood crime predict significant decreases in school academic achievement across time, yet the reverse was not true.  Subsequent models that included school climate variables showed that higher neighborhood crime predicted marginally significant reductions in SRC, statistically significant reductions in SEL, and no differences in AR over time.  In turn, higher levels of SRC, SEL, and AR were all found to predict increases in academic achievement over time. Tests of indirect pathways indicated preliminary support for SEL (but not SRC or AR) as a mediator of the relationship between neighborhood crime and decreases in academic achievement. 

Results suggest that community crime is directly related to changes in school academic achievement, a process that appears to be primarily unidirectional but may be partially explained by changes in schools’ levels of SEL.  These results suggest several opportunities for policy and intervention, and support the use of school-level socioemotional learning initiatives for promoting academic resilience in high-crime contexts.  More information on policy implications and robustness checks (including separate analyses of violent vs. property crime) will be included in the final presentation.