This afternoon I attended a seminar by Patrick Sharkey, associate professor at NYU. The name of his talk was “the effects of community violence on child cognitive skills and academic performance”.
His research measures the impact of neighborhood violence on children academic performance. he uses”coincidence” estimation. Within each neighborhood, he argues that the relative timing of homicides and interviews produces exogenous variation of timing. He interviews a child four days after a homicide happened in the neighborhood, and sees if her test score changes relative to children who are unaffected by homicides. He measures “neighborhood” using block group, census tract, and neighborhood cluster. Note that this method only picks up short run effects. Sharkey emphasized on the importance of evaluating how kids are functioning on a DAY-TO-DAY basis (e.g. mindset) because it is those incremental changes that lead to differential personal development in the future. His methodology is novel in two aspects: First, he employs the different timing of violence as exogenous variation and measures its impact on the test scores of school children. Second, he targets on specific incidents, e.g. homicide in the Chicago data, instead of the aggregate crime rates. This breaks down the impact of violence by category so that we can get a better idea of whether a particular type of violence affects children’s cognitive ability and extrapolate the mechanism of the impact.
However, this method is not free from flaws. African-American children are overrepresented (nine times the number of white children), and this weakens his later argument on the comparative statistics between white and black children. Furthermore, there may be a systematic difference between treatment and control group children since an interviewer might not want to go to the same neighborhood until some days after the interview is done.
The results are consistent with his hypothesis that neighborhood violence does affect a child’s academic performance. His paper on New York City public schools also found that black students are affected more by the violence in the community. Students who have experienced violence in the neighborhood are 1.13% less likely to pass the exam, and the reduction is 3% for black students. These results imply that violence has a long reach. It also suggests tha violence may be a central mechanism by which neighborhood disadvantage affects child development (cognitive skills and academic outcomes) and other health outcomes in adulthood (e.g. the feeling of safety).
There were a lot of insightful comments from the audience in the Q&A session. One professor suggested that parenting strategy might play a role in determining the level to which children are affected by local violence. Someone suggested finding news coverage on local violence and assess whether the widely covered events affect people only in the neighborhood or the whole city/state. It might also be useful to look at whether the victim is in the network of the child and control for this variable when running the regression.