This research examines the extent to which, individual and neighbourhood characteristics contribute to the risk of two different types of adolescent aggressive behaviour: aggression and delinquency. In addition, it explores potential mechanisms explaining the influence of neighbourhood conditions on adolescent aggressive behaviour. Data regarding adolescent behaviour is taken from self-reported surveys, applied to 1,686 Colombian adolescents, residing in 103 neighbourhoods. Data regarding neighbourhoods is taken from official government datasets, as well as two community surveys that are independent of the individual aggression survey. A range of statistical approaches is used to develop reliable valid measures of both adolescent aggressive behaviour and neighbourhood characteristics: multilevel Rasch models, multilevel factor analysis, ecometrics, spatial multiple membership models, Geographic Information Systems and hierarchical Bayes procedures. For the analysis, the research develops an explicit conceptual framework and uses multi level modelling and multilevel structural equation modelling to obtain unbiased estimates of overall effects, cross-level interactions, direct and indirect effects. Results indicate that individual and neighbourhood-level factors are not only directly and indirectly associated with adolescent aggressive behaviour, but also interact with one another to shape adolescent behaviour. By identifying the processes through which neighbourhoods constrain, enhance or modify adolescent behaviours, these results may be used to inform community based programs, aiming to reduce adolescent aggressive behaviour.
|Date of Award||2012|