Direct Tests

A few studies have reported comparisons closely tied to our research question. Land, McCall, and Cohen (1991) compared models across crime type and concluded that poverty is most strongly related to homicide and assault rates and least strongly associated with robbery, larceny, and auto theft. Krivo and Peterson (1996) found that extreme levels of neighborhood disadvantage produce exceptionally high rates of violent crime but not substantially higher rates of property crime, a pattern that they largely replicate with their more extensive National Neighborhood Crime Study data (Peterson & Krivo, 2010). All of these findings are in the direction we expect, though we acknowledge that authors may not emphasize nonsignificant differences when they do find them. In our own recent analysis, we directly tested whether poverty is a differential predictor of violent behavior. Using data from the National Longitudinal Survey of Adolescent Health (Add Health), we first regressed a dummy variable differentiating any violent and nonviolent-only offenders (self-report) on a series of control variables and family income. We found that violent adolescents had significantly lower family income than nonviolent-only offenders, controlling for age, minority status, sex, community disorder, parent education, intact family, peer delinquency, alcohol use, drug use, and, importantly, frequency of nonviolent offending. When we compared any violent to frequent nonviolent- only offenders, the association was still negative, in the predicted direction, but only marginally statistically significant. We did not find that arbitrary cut-offs, comparing those “in poverty” vs. those not in poverty, distinguished between the two groups of offenders.

< Prev   CONTENTS   Source   Next >