The broken windows thesis has had a profound impact on policing strategies around the world. The thesis suggests that police can most effectively fight crime by focusing their efforts on targeting disorder--minor crimes and nuisance behaviors such as loitering, public drinking and vandalism, as well as dilapidated physical conditions in a community. The strategy was most prominently used in New York City in the 1990s, and has been often credited for the crime drop observed in the city over that decade. Despite the widespread influence of the broken windows thesis, there has been relatively little rigorous empirical research which has sought to test the validity of its theoretical propositions.

This dissertation aimed to address this shortcoming by using structural equation modeling to test the relationships between perceived disorder, fear of crime, collective efficacy and perceptions of crime suggested by the broken windows thesis using survey data collected during a randomized, experimental evaluation of broken windows policing in three cities in California.  The results are supportive of the broken windows thesis, but also raise some challenges.  Perceptions of disorder were found to increase fear of crime, reduce collective efficacy and lead to crime as suggested.  However, fear of crime was not significantly related to collective efficacy as suggested, and the direct effect of perceived social disorder on perceptions of crime was the strongest effect in every model.  Nevertheless, the findings do suggest that a reduction of disorder in a community may have positive effects in the form of reducing fear and promoting collective efficacy, and suggest the limitations of studies which only test for direct effects of disorder on crime and/or do not examine the variables at the perceptual level.  Future research needs to further examine the broken windows thesis, ideally involving a prospective longitudinal study of crime at place.