U.S. flag

An official website of the United States government, Department of Justice.

NCJRS Virtual Library

The Virtual Library houses over 235,000 criminal justice resources, including all known OJP works.
Click here to search the NCJRS Virtual Library

Baltimore Community Policing Experiment

NCJ Number
A M Pate; S O Annan
Date Published
161 pages
This report evaluates an effort by the Baltimore Police Department to implement two types of community policing -- foot patrol and ombudsman policing -- in an attempt to bridge the distance between citizens and the police.
A multi-stage selection process was used to ensure that the experimental neighborhoods were comparable to each other and representative of a broad range of socioeconomic conditions. Three areas were selected in each cluster for the experiment and were randomly assigned to receive either foot patrol, ombudsman policing, or no new police programs at all. The elements of sampling procedures included a real listing and household selection, respondent selection within the household, supervisor and interviewer training, contacts at sample households, call-back procedures, in-field editing, validation, and response rates. The report summarizes the outcome measures used, such as recalled program awareness; evaluation of police service; perceived area social disorder and crime problems; perceived likelihood of area crime, and perceived area safety. Also discussed are fear of crime, crime avoidance behaviors used, utilization of crime prevention devices, familiarity with neighbors, cohesiveness of neighborhood, satisfaction with area, victimization in area, and knowledge of others' victimization. Calls for service and crime data for the six experimental areas were collected, categorized, and analyzed. Both types of community policing were implemented in July 1987; the report describes how they were implemented, the level of activity dedicated to the programs, and the public response to these efforts. Four different types of analyses were performed: multivariate regression analyses to provide statistical indicators of overall program effects, regression analysis to test for differences in program effects across the two selected neighborhoods, regression analyses to test for possible subgroup-specific program effects, and interrupted time series analyses of calls for service and recorded crime data to determine if trends or levels were affected by program implementation. The most significant result of the experiment was that ombudsman policing, when practiced full-time, produced great improvements in evaluations of police effectiveness and behavior, reduced perceptions of disorder, increased feelings of safety, and reduced awareness of victimization in the neighborhood. When implemented with a part-time staff, this type of policing only improved evaluations of police effectiveness. Foot patrol had no significant effect on evaluations of police and had mixed effects on perceived crime and disorder. For a summarization, see Baltimore Community Policing Experiment: Summary Report, NCJ-121574.