SPI Findings

Process and outcome evaluation of the randomized control design

Offender-focused strategy outperformed foot patrol and problem-solving strategies.

Compared to the control areas, the treatment areas that received the offender-focused strategy experienced a 22% decrease in violent crime, and a 31% decrease in violent street felonies.

Process and outcome evaluation, pre- and post- deployment

From pre- to post-deployment, officers with BWCs experienced a 22.5 percent decline in officially recorded complaints, while across all other precincts there was a 45.1 percent increase in complaints.

The research team randomly selected officers to wear a BWC from the pool of 467 officers who provided consent. The evaluation relied on BWC metadata automatically generated by camera activation, official police computer-aided dispatch (CAD) data, official arrest data, official use of force reports, and citizen complaints reported to the PPD.

• Officers wearing BWCs activated their camera in 45% of incidents at the beginning of the study, however, after PPD adopted a more restrictive activation policy, activations increased to roughly 75% of incidents.
• While BWCs had little to no impact on officer perceptions, handing of dispatched calls for service, response time, likelihood of arrest, and use of force, they were related to a decline in the rate of complaints for officers who were mandated to wear a BWC and an increase in the rate of complaints for officers who voluntarily wore BWCs.
• Officers mandated to wear BWCs were less likely to agree that BWCs improve officer efficacy after being assigned a BWC.

This project evaluation included process evaluation methods that were designed to describe the activities of and individuals served by the Pinellas County Sherriff’s Office (PCSO) Mental Health Unit (MHU), and it also included outcome evaluation methods designed to assess the effectiveness of the MHU.

• On average, participants significantly reduced their average number of involuntary commitments following their first contact with the MHU. Whereas these 17 individuals averaged 3.29 involuntary commitments in the nine months prior to their first MHU involvement, they averaged only 0.82 involuntary commitments over the nine months following their initiation with the MHU.

Ninety high crime areas were randomly assigned to receive none, two, or four 15-minute supplemental CEPs a day for a three month period. Offense reports, CAD data, resident surveys, officer focus groups, and officer surveys were used to conduct process and outcome evaluations of the NI-Loc program. 

• Results indicate that the NI-Loc intervention did not affect crime, calls for service, or public perceptions of police, in treatment areas compared to controls.

An experimental intervention testing the efficacy of a police-mental health provider co-responder model for addressing mental health-related calls for service.

• There were no significant reductions in the numbers of calls for service and mental health calls between the treatment and control groups during the treatment and post-treatment periods
• However, those who completed some treatment appeared to be high-utilizers of police resources before the intervention. Those who completed some treatment show decreased patterns in general calls for service and mental health calls compared to the dropouts.

Examination of descriptive statistics, correlation coefficients, and logistic regression models.

The evaluation validated the risk assessment tool as a predictive device: for every one-point increase in the assessment tool score that a dispute received, the odds of violence occurring in that dispute over time increase by 29%.

The intervention was intentioned to result in a series of staggered short-term and intermediate outcomes that would over time contribute to increase safety and collective efficacy in the community around issues of homelessness (the long-term impacts). The evaluation also conducted a process evaluation on implementation of HOT.

• There was some evidence that calls for service dropped in response to the SPI-HOT team’s work, and that calls for service did not simply increase in adjacent areas.
• The HOT initiative clearly achieved the goal of developing stronger relationships between the Sacramento Sheriff’s Office and public and non-profit organizations who work with the same homeless population.
• There was limited evidence that activities by HOT have resulted in substantial institutional and cultural change across the Sherriff’s Office and its other patrols.

Pre-post impact analysis, assessment of service provision

16% reduction in violent crime at the Central Precinct (target area).

Exploratory analyses found that providing services could reduce the likelihood of future crimes.

Interviews concluded that SIP was effective because of the multi-agency collaboration and ability to balance treatment and enforcement approaches.

A quasiexperimental design based on propensity score matching (PSM) methods in which RADAR response plan recipients would be matched with similarly-situated people in a comparison jurisdiction to compare outcomes accompanied by a process evaluation. 

• RADAR had a positive influence on Shoreline Police Department’s culture in terms of changing attitudes and responses to people with BH/DD.
• The evaluation did not find significant effects on rates of mental health-related calls for service or incidents.
• Use of force was lower after RADAR was implemented, but this was not statistically significant.