The research then integrated census data, police records, and an independent survey of more than 3,500 residents to test a theory of collective efficacy and structural constraints. Collective efficacy, which involves cohesion among residents combined with shared expectations for the social control of public space, explained lower rates of crime and observed disorder after controlling for neighborhood structural characteristics. Collective efficacy also was linked to lower rates of violent crime, after accounting for disorder and the reciprocal effects of violence. However, contrary to the broken-windows theory, the relationship between public disorder and crime was spurious, except perhaps for robbery. Findings contradicted the strong version of the broken-windows thesis, but they did not imply the theoretical irrelevance of disorder. Nevertheless, findings indicated that the current fascination in policy circles on cleaning up disorder through law enforcement techniques appears simplistic and largely misplaced, at least in terms of directly reducing crime. Tables, figure, footnotes, and 78 references (Author abstract modified)