NBER posted an RCT of Sit-D, a police training program teaching the value of developing multiple perspectives on a situation. Quick take: High-quality RCT finds suggestive evidence of a 12% reduction in police use of force over 1 yr (findings are moderately encouraging but not as strong as presented in study).

​

Program:

​

• Situational Decision-making Training (Sit-D) is a 4-session training that teaches officers that “to respond effectively to ambiguous situations, it is critical to go beyond one’s first impression and develop additional possible explanations for what is occurring.”

 

Study Design:

​

• The study randomly assigned 2,070 police officers in Chicago Police Department to treatment (Sit-D training) vs control. Based on careful review, this was a high-quality RCT (e.g., baseline balance, minimal attrition, prespecified analyses).

 

Findings:

​

• The study unfortunately doesn't report on all prespecified primary outcomes over the 1 yr post-training follow-up, & highlights effects on other outcomes that could be chance findings given the large number of measured outcomes.

• But I was able to find the impacts for many (not all) prespecified primary outcomes in the report or appendix, & show them in the table below. There are some sizable effects (e.g., reduced use of force) but most are not statistically significant & they seem to fade a bit over the 1 yr follow-up.

​

​

​

​

​

​

​

​

​

​

 

 

 

 

Comment:

​

• Bottom line: I think that the findings are moderately encouraging but not yet reliable (due to lack of statistical significance), & that Sit-D would be a strong candidate for future RCTs, perhaps with training enhancements or booster sessions to try to strengthen the impact.

• Given the urgency of police reform, we need more high-quality RCTs like this to identify truly effective police training programs. The vast majority of police training programs currently in use are untested & many are likely ineffective.