@PNASNews published an RCT of Procedural Justice training for police, to foster fair/respectful treatment of citizens. Quick take: High-quality RCT found no effect on community perception of police legitimacy (primary outcome) but suggestive effects on some other outcomes.

Program & Study Design:

  • The study randomly assigned 28 police officers in 3 cities (Tucson, Cambridge, Houston) to PJ training vs control, and randomly assigned 120 crime hot spots in the cities to be policed by trained vs control officers over a 9-month period. PJ training was an intensive, 5-day course.

Findings:

  • The study found no significant effect on the study's primary pre-specified outcome - residents' perception of policy legitimacy (e.g., trust in police, obligation to obey police), as measured with community surveys.

  • The study found positive effects on some secondary outcomes (e.g., residents' perceptions of police misbehavior such as harrassment, observer measures of police behavior during ride-alongs) but not others (e.g., citizen calls for service - the main pre-specified measure of crime).

  • The positive effects are encouraging but I think best viewed as preliminary because (i) study measured many outcomes (which can cause false-positives), & (2) T officers might temporarily display expected PJ behaviors for observers during ride-alongs (social desirability bias).

  • Also, the 28 police officers in the sample had all volunteered for the study, so we don't yet know PJ training's effects on other police who don't volunteer.

Comment:

  • Given the urgency of police reform, we need more high-quality RCTs like this to identify effective police training programs. Disclosure: My former employer, Arnold Ventures, funded this study. Here's the full study report.

We use cookies to improve your experience and to help us understand how you use our site. Please refer to our cookie notice and privacy statement for more information regarding cookies and other third-party tracking that may be enabled.

Intuit Mailchimp logo