The Effect of Race on Crime: A Multilevel Analysis

Many studies have been carried out to examine the sources of racial disparities in crime. However, there are some limitations in most of those studies. One limitation is that the majority focus on black-white comparisons. Another limitation is that many primarily examine violent offending. In addition, most studies have solely relied on either contextual level or individual level explanations. My dissertation attempts to address these limitations in previous literature by using data from the first wave of the National Longitudinal Study of Adolescent Health to examine racial disparities in different types of offenses among non-Latino whites, non-Latino blacks, non-Latino Asians, and Latinos. I use a multilevel linear method to model self-reported violent, property, and drug offenses when controlling both contextual level and individual level covariates that are drawn from social disorganization, anomie/strain, social learning, and social bond theory, simultaneously. Different from previous studies, I also use item response modeling to construct measures for my dependent variables such as violent offense. Findings from my dissertation show that there are some disparities in different types of offending between white and non-white adolescents. Furthermore, the gaps in different types of offending for Asian-white, black-white, and Latino-white comparisons are affected by different explanatory covariates. Demographic background such as immigration status, school bonds, and grades (social bond theory) seem to explain Asians’ lower level of reported violence in relation to whites. The gaps in violence between whites and blacks seem to stem from multiple sources. All aforementioned four theories seem to provide some explanations for this group comparison, but none of these theories explain away the differences in violent offending between them. For Latino-white gaps in violence, contextual effects such as concentration disadvantages account for the differences between these two groups. As for property and drug offending, the social learning predictor peer delinquency seems to explain the gaps between whites and other races. Additionally, different covariates such as peer delinquency, school bond, and grades also have differential effects on different types of offense across different neighborhoods.