This thesis consists of three interrelated yet self-contained empirical essays that use large-scale individual level survey data to study the problem of crime among children and youth in Canada and the United States. The first essay is intended to identify from a large set of potential explanatory factors important correlates of youth criminal and educational outcomes, accounting for unobserved correlations among different youth outcomes. The second essay tries to address an empirical puzzle, that is, American teenagers on average are three times as likely to engage in fights as their Canadian peers and this cross-country violence gap has opened up among children as young as 4-5 years old. The third essay analyzes the impact on youth crime of a nation-wide policy reform in the Canadian youth criminal justice system, i.e. the superseding of the Young Offenders Act by the Youth Criminal Justice Act in 2003.