|
Abstract
|
PURPOSE In the present paper, we test whether academic factors differentially predict violent offending, based on "differential etiology of violence" hypothesis as proposed by Savage and Wozniak (Thugs and Thieves: The Differential Etiology of Violence, [58]).
METHODSWe use data from the National Longitudinal Survey of Adolescent Health. We used a three-tier statistical approach to test the differential etiology of violence thesis. First, we compared the slopes of coefficients generated by ordinary least squares regression models. Second, we examined negative binomial regression partial coefficients estimating the association between academic achievement and violent offending, controlling for nonviolent offending. Finally we used binary logistic regression to compare "any violent" to "nonviolent-only" offenders.
RESULTSThe findings suggest that academic achievement, but not school attachment, is differentially associated with violent behavior in longitudinal, conservative models. Grade average is lower and several indicators of school problems are higher among violent offenders than nonviolent-only offenders, and GPA distinguishes violent from chronic, nonviolent-only offenders in models that control for a host of social factors as well as intelligence (measured with a picture vocabulary test), ADHD, alcohol, and drug use.
CONCLUSIONSThe findings are consistent with the differential etiology of violence thesis, challenging the adequacy of general theories of offending and calling into question the assumption that the causes of violent offending are identical to those of nonviolent offending. Policy implications of these findings include the possibility that academic measures and programs might be used to target and address violent behavior problems; theoretical implications include a call for criminological theories specific to violence and appropriate tests of those theories.
Language: en |