Purpose: The purpose of this paper is to evaluate an experiment to improve residents’ opinions of the police in Portland, Oregon. Officers conducted community engagement patrols (CEPs) in 60 high-crime areas. The CEPs prioritized non-investigative contacts with community members to build trust and promote positive police–community interactions in designated high-crime locations. It is hypothesized that community members living in/near intervention sites would report greater exposure to officers, more positive interactions and feel more positively about police than residents in control areas.

Design/methodology/approach: In total, 90 crime hot spots were identified using crime reports and calls for service. Locations were randomized into three groups: 2 CEPs/day (n=30), 4 CEPs/day (n=30), and control (i.e. no supplemental patrols, n=30). Officers were dispatched to treatment locations via the computer-aided dispatch system for 90 consecutive days, resulting in 16,200 scheduled CEPs. Surveys were mailed to 11,760 households immediately after the intervention ended and 1,537 were returned (13.1 percent).

Findings: Residents from intervention areas reported a higher number of positive police contacts, whereas contacts that residents perceived as negative did not differ between the three conditions. Community attitudes, including perceived police legitimacy, were generally unaffected by CEP dosage.

Originality/value: This paper documents the outcomes of a large-scale field experiment seeking to improve public attitudes toward police using directed CEPs in crime hot spots. Whereas the intervention succeeded in providing more opportunities for positive contact with police, attitude change may necessitate longer-term strategies.