Program Profile: Procedural Justice Training Program (Seattle Police Department)

This program was designed to “slow down” police officers’ thought processes during encounters with citizens. The program is rated No Effects. At the 6-week follow-up, treatment group officers were statistically significantly less likely to be involved in an incident in which physical force was used, but there were no statistically significant differences on the fraction of incidents that resulted in an arrest, the number of citizen complaints filed against the officer, and other outcomes.

A No Effects rating implies that implementing the program is unlikely to result in the intended outcome(s) and may result in a negative outcome(s).

This program's rating is based on evidence that includes at least one high-quality randomized controlled trial.

Program Goals

The procedural justice training program for the Seattle Police Department (SPD) was designed to affect the way officers gather and process information in the field. The goal was to help police officers “slow down” their thought processes during citizen encounters, even while performing mundane aspects of the job. The training was intended to reduce the frequency with which officers engage in behavior that could be perceived by the public as unjust. The training involved short supervisory meetings with officers who were identified as being at higher risk of having negative encounters with the public.

Target Population/Program Components

The program was designed for SPD officers who were identified as being at higher risk for experiencing negative encounters with the public. Specifically, these were officers who worked in hot spots, or places within the city characterized by consistently high crime (which may place officers at higher risk for being in encounters that could escalate rapidly). Additionally, these officers had been involved in an incident in which they were injured, used force, or a complaint had been filed about them or the incident in which they had been involved.

The program had two components: notification and engagement. During the notification process, randomly selected officers and their supervisory sergeants were notified of their selection for the program. Next, the officer, his or her supervisory sergeant, and the lieutenant worked together to identify a citizen encounter that the officer had recently been involved in. The encounter had to meet at least one of the following criteria: 1) the incident was initially classified as “priority one,” meaning a report had been filed about the incident; or 2) the incident involved an arrest. Incidents that involved an active complaint about the officer or use of force, or that were subject to disciplinary review for another reason were not eligible. Once the incident was selected, a supervisory training meeting, or engagement, was scheduled.

During the engagement, the supervisory sergeant explained how enforcing the law in a procedurally just manner was a central component of the SPD’s mission and that the goal of the meeting was to discuss how the officer used procedural justice during the incident in question. Officers were asked to remember the incident, including how they learned about the incident, what they thought they would see on arrival, and what they observed once they arrived at the scene. This prompted the officers to reflect on their thought processes during parts of the job in which automatic responses tend to take over and officers may tend not to think very carefully.

Sergeants were instructed to ask open-ended questions (such as “What did you observe when you arrived?”). This allowed the officers to determine the timing and pace of the meeting. This was also intended to remind officers that authority does not always perfectly coincide with total control (i.e., allowing officers to speak freely does not mean that the sergeant is not in charge of the meeting; thus, allowing citizens to speak does not diminish the officers’ control over a situation).

After reviewing the encounters, the sergeants asked the officers how they felt about the outcome, and if there was anything they would have done differently. Sergeants also asked officers to provide feedback about the sergeant’s own performance during the meeting. This allowed lower-ranking officers to discuss how they felt about the sergeants’ performance, which was intended to model the desired behavior for officers (i.e., that they consider how the citizens they encounter feel about their performance on the job). Officers were also given an anonymous comment card, to be filled out after the meeting, with their responses shielded from view.

See ‘Implementation Information’ for additional information on the sergeant script for the supervisory meeting.

Program Theory

The program is based on the principles of the procedural justice model of policing (Sunshine and Tyler 2003;Tyler 1988), which suggests that in order for officers to be perceived as legitimate and trustworthy, citizens must perceive the following: 1) they are being treated fairly relative to others (motivation, honesty, ethicality); 2) they are given the chance to explain or defend their behavior (opportunities for representation and error correction); and 3) their explanation is taken into account before any police action is taken (perceived officer bias and decision quality). In Washington State, procedural justice is referred to as LEED, which stands for “listen and explain with equity and dignity.”

The program is also based on psychological research findings suggesting that as the brain familiarizes itself with a task, a person can “automate” his or her behavior, paying increasingly less active attention to small details and taking cognitive “shortcuts” to reach the same result, which increases efficiency (Augoustinos and Innes 1990; Fiske and Taylor 1991; Tversky and Kahneman 1974). Police officers who automate their behavior do not consciously process all new information provided to them by involved citizens and may be less conscious of the words they choose or actions they take when interacting with civilians, which is a necessary component for procedural justice.

Study 1

Owens and colleagues (2018) used a randomized controlled design to evaluate the effect of a procedural justice training program for police officers in Seattle, Wash. Eligible officers were those who patrolled in crime hot spots, which placed them at a higher risk for negative citizen encounters. The researchers used data from computer-aided dispatch (CAD) incidents recorded from January to December 2010, to calculate risk scores for officers who had been involved in one or more incidents in which they were injured, used force, or had a complaint filed about them or the incident. The officers with the highest risk scores were then randomly assigned to either the treatment group (n = 221) or the control group (n = 1,213). The mean predicted risk score for treatment officers (m = 0.116) was similar to that of the control officers (m = 0.120). The selection of officers with the highest risk scores was based on the idea that officers who worked in hot spots might be more likely, to benefit from additional procedural justice training than officers who worked in areas with lower levels of predicted risk.

Officers assigned to the treatment group participated in a supervisory meeting focused on a recent citizen encounter selected jointly by the officer, the supervisory sergeant, and the lieutenant. Control group officers did not participate in any supervisory meetings. Officers in the treatment group were instructed by email (the notification) to set up the supervisory meeting (the engagement). For the evaluation, the preintervention period included each officer’s activity prior to being notified about the meeting. The postintervention period included officer activity for 6 weeks after the meeting.

The experiment took place between May and November 2013. Officer activity and arrest rates were recorded for the 1- and 6-week periods before and after treatment. The three primary outcomes were 1) fraction of incidents that resulted in an arrest, 2) number of times an officer was involved in an incident in which physical force was used, and 3) the number of citizen complaints filed against the officer. The researchers used a first-differences model and a lagged dependent variable model to test the effect of the intervention. The first-differences model captures time invariant differences in officer behavior, or behaviors that do not change over time (e.g., some officers always use a lot of force). The lagged dependent variable model captures more dynamic sources of behavioral variation between officers (e.g., officers who often use force may be reassigned to beats where they will inherently use less force). Subgroup analyses were not conducted.

The procedural justice training program was implemented through a collaboration consisting of the research team, the Seattle Police Department, and the city of Seattle, Wash. The program is a low-cost, low-intensity supervisory program based on principles of procedural justice and research evidence from psychology and economics.

Information on the procedural justice training toolkit and sergeant script for the procedural justice supervisory meeting can be found in Appendix A of the study (Owens et al. 2016).

These sources were used in the development of the program profile:

These sources were used in the development of the program profile: