Sobre desorganização social e vitimização no Brasil : investigando eficácia coletiva com uso da primeira Pesquisa Nacional de Vitimização

Abstract

This dissertation investigates the Robert Sampson's concept of collective efficacy through quantitative analysis of data from the First National Victimization Survey (PNV). The proposal is to discuss and evaluate the foundations of the "Latin American Paradox of Neighborhood Effects" hypothesis, which suggests that in countries of that region the social ties density, does not turn into lower crime rates, especially in areas with high economic deprivation and high violent crime. The empirical analysis was built by adjusting statistical binary regression models with fixed and mixed effects to estimate the chances of victimization for crimes against property (theft or robbery) and against person (simple or aggravated assaults) and measures perceived Social Cohesion and Expectancy of Informal Social Control as variables of interest. The results partially confirm the hypothesis of the Latin American paradox of neighborhood effects, suggesting the importance of the institutional dimension as a moderator of the effects of the perception of social ties on victimization. On the other hand, they raised the possibility that exercise of Informal Social Control in a direct way is a concept weakness that seems poorly adapted to conditions of socioeconomic and community mobilization present in Brazilian communities.