The NPO program treated the police officer as a community social worker rather than a law enforcement officer. A community effect analysis, which also includes an assessment of attitudes toward the community, suggests that research and control neighborhoods differed significantly on victimization before and after the program. Changes in level of victimization could not be attributed to the policing intervention. Awareness of the program, especially by residents in the research neighborhoods, increased dramatically over the life of the project. Improvements in police and community relations were much more indirect than first anticipated. Research neighborhoods had a lower self-conception than did their counterparts in the control neighborhoods. The same research neighborhoods showed a positive shift in attitude toward their community while the comparison neighborhoods showed a decline and restabilization in attitudes toward the community. There was a general decline in attitudes toward the police even when attitudes toward the community were improving. 4 tables, 2 notes, and about 45 references. (Author abstract modified)