SPI Findings

2022 SPI Map and Findings graphic

Learn more about SPI site successes as it relates to gun violence, and violence reduction through our SPI Successes One-Pagers or using the below table.


SPI Findings

SiteImplementationResearch Design & FindingsTopic
Boston, MA2009

The Boston SPI addressed violent crime, especially robberies and assaults committed with guns, using community policing and problem-oriented policing Safe Streets Teams (SSTs).

  • Completed 28-year longitudinal analysis of violent crime to identify 13 chronic hot spots.
  • Targeted SSTs, consisting of a sergeant and six patrol officers, in hot spots.
  • Deployed nearly 400 different situational/environmental, enforcement, and community/social service techniques to identify and address recurring problems.
Research Design

A quasi-experimental design with propensity score matching (statistically based matching of target sites to control/comparison sites).


More than 17% reduction in violent crime.

More than 15% reduction in aggravated assaults.

More than 19% reduction in robberies.

Gun Violence, Violence Reduction
Boston, MA2011

The Boston SPI implemented the Boston Homicide Clearance Intervention to improve homicide clearance rates and investigations.

  • Conducted spatial analysis on 314 homicide victimizations.
  • Convened a Homicide Advisory Committee.
  • Expanded the police department homicide unit by 36% to include an additional detective, a crime analyst, and two Victim-Witness Resource Officers on each squad.
  • Updated 40-hour crime scene response and investigation in-service training.
Research Design

Process evaluation of homicide protocols, impact evaluation during the implementation period, overall outcome evaluation of clearance rates.


9.8% increase in homicide clearance rates.

18.4% increase in adjusted homicide clearance rates (including cases awaiting a grand jury).

Violence Reduction
Brooklyn Park, MN2013

The Brooklyn Park, MN SPI - the Assets Coming Together to Take Action (ACT) Initiative —aimed to reduce crime at hot spots by garnering greater cooperation and collaboration with the public in addressing crime problems. The intervention was also designed to increase collective efficacy among citizens residing in the hot spots, and improve their evaluations of police legitimacy. A key innovation of ACT was the involvement of the entire patrol force in delivering the program: all patrol officers were expected to undertake the activities described below during their discretionary time when they were not responding to 911 calls, taking breaks, or report writing.

Research Design

The research partner evaluated ACT using a partially block-randomized controlled trial in which 42 identified hot spots were matched into groups according to their crime rates and overall land use of the site (e.g., commercial, residential). The hot spots were then randomly allocated into treatment (ACT) or control (policing as usual) conditions within their matched groups.

  • Sufficiently staffed police department, unallocated patrol time can be used to to develop a concentrated community based problem solving efforts at crime hot spots.
  • Officers implemented the intervention with fidelity. 
  • The program’s emphasis on community collaboration with the police led to increased reporting of crime.
Community Engagement, Violence Reduction
Cambridge, MA2011

The Cambridge SPI, a collaborative effort with the cities of Everett and Somerville, called RASOR (Regional Analytics for the Safety of Our Residents), focused on preventing victimization and social harm.

  • Created partnerships with case managers, social service providers, other criminal justice agencies, and research staff to collectively cultivate a focused deterrence strategy.
  • Examined violent criminal and arrest histories as well as social harm to identify impact players.
  • Developed detailed case profiles on potential candidates and assigned each a case manager.
  • Conducted notification meetings and delivered resources.
Research Design

Outcome evaluation and process evaluation of the randomized experimental design.


No statistically significant differences in time to arraignment between RASOR and control.

The more time invested by case management and services, the longer the survival time of individuals participating in RASOR (i.e., did not reoffend as quickly as the control group).

Violence Reduction
Chicago, IL2015

The Chicago, IL SPI addressed violent crime through implementation and evaluation of the police department’s Strategic Decision Support Centers (SDSC). Chicago’s SDSCs bring together police officers and analysts from the University of Chicago Crime Lab to integrate crime intelligence, data analysis, and technology in daily, real-time approaches to violent crime reduction and prevention. The Chicago SPI expanded SDSCs into six districts in the city, and evaluated the impact of this expansion on violent crime in the target districts and citywide.

Research Design

The SPI conducted a process evaluation of SDSCs for developmental and operational lessons learned. The site also employed a difference-in-difference models to estimate crime reduction effects resulting from adding an SDSC to a Chicago police district. 

  • Estimated crime reductions, depending on crimes and times modeled, varied between 3 percent and 17 percent. 
  • SDSCs estimated to reduce shootings by more than 40 percent in one district. 
  • Some SDSCs were fully integrated into their district’s full suite of operations; others could be characterized as being more than independent service units but not yet achieving a partial level of integration into operations.
Gun Violence, Violence Reduction
Chula Vista, CA2013

The Chula Vista, CA SPI developed the Domestic Abuse Response Team (DART) and provided assigned officers a protocol to follow at each Domestic Violence (DV) call involving Intimate Partners (IP).The DV protocol developed was carried out by patrol officers in one sector of the city along with their regular call-handling responsibilities.

  • Officers delivered a face-to-face message to subjects in IP-DV calls that were verbal disputes. A different message was delivered to both IP-DV victims and suspects in crimes, while offenders who were arrested received a highly structured warning at the jail from officers.
  • Handouts reinforced the verbal messages.
  • Both IP-DV victims and suspects were to receive a face-to-face follow-up contact by officers after three days, while subjects involved in IP-DV verbal disputes were to receive a follow-up text after 30 days. Chronic IP-DV subjects, offenders and victims – those who persisted in verbal disputes or crimes despite the messaging – received a customized intervention.
Research Design

The study used a quasi-experimental design to evaluate the impact on crimes, calls, and repeats, assessing whether dyads or locations
– had a subsequent call or crime within 12 months of the initial triggering event.

  • Although the DART initiative was initially designed to reduce repeat IP-DV calls, the initiative had a significant impact on crime – resulting in a 24% drop in DV crimes after one year while the volume of DV calls remained mostly steady.
  • While the DART initiative relied on patrol officer time, it was not resource intensive; calls for service (CFS) data indicated that DART officers spent an extra seven to eight minutes on DV CFS.
Violence Reduction
Glendale, AZ2009

The Glendale SPI sought to reduce crime and disorder pertaining to convenience store thefts in the southeast quadrant of the city. 

  • Provided officers with 20 hours of POP training using the SARA (Scanning, Analysis, Response, and Assessment) model to identify and reduce crime.
  • Analyzed crime to find that crime centered at six local convenience stores.
  • Implemented a three-pronged response that included engagement with store leadership, prevention strategies (Crime Prevention through Environmental Design or CPTED), and suppression strategies.
Research Design

Pre/post analysis with multiple units and multiple control groups, interrupted time series analysis of convenience stores.


Decline in calls for service in 5 of 6 target stores.

18% reduction in direct costs for officer response and estimated decrease of more than $1.9 million in victimization costs.

Sustained reduction in 4 of 6 target stores (nearly 20% in calls for service overall, two years after the intervention).

Violence Reduction
Glendale, AZ2011

The Glendale SPI continued the efforts of its prior project by focusing POP efforts on problem offenders and organizational retail theft.

  • Conducted intensive problem analysis of calls for service data, resident surveys, and social and physical disorder surveys.
  • Used new data collection protocols and social network analysis.
  • Implemented traditional crime prevention strategies, as well as CPTED, outreach and education, counter insurgency, and informal focused deterrence.
Research Design

Bivariate analysis and interrupted time series analysis.


Identified and targeted prolific offenders, and generated short-term, notable declines in several micro hot spots.

27% reduction in calls for service in a target area (large mall) and short-term 15% reduction in calls for service in a target apartment complex.

Violence Reduction
Kansas City, MO2012

The Kansas City SPI addressed violent crime, particularly gun violence, using a model called the Kansas City No Violence Alliance (KC NoVA).

  • Used social network analysis and group audits to identify violent offender groups (and their frequent associates).
  • Implemented a full range of interventions (e.g., offender notification sessions, enhanced monitoring by patrol officers, strategic federal prosecution, intensive probation and parole supervision, social services).
  • Collaborated with numerous criminal justice stakeholders.
Research Design

Bivariate analysis and interrupted time series analysis to assess the impact of the focused deterrence pulling levers strategy.


40% reduction in homicide.

19% reduction in gun-related aggravated assaults.

Gun Violence, Violence Reduction
Kansas City, KS2016

The Kansas City, MO SPI approach employed evidence-based strategies incorporating both place-based and person-based violence prevention methods in a micro hot spot (MHS) network experiment. The SPI implemented two interventions to examine the effectiveness of each in the East Patrol Division: 

  • Deployed Saturation Patrol in Five Areas: A place-based approach that added 15 minutes of police presence during high-crime periods.
  • Deployed Network-Based Intervention in Five Areas: An approach using social network analysis to identify persons central to violent crime for individualized police attention (diversion/services or enforcement actions).
Research Design

The sites, with buffer zones, were randomly assigned as either treatment or control areas: five for saturation patrol; five for network-based intervention (NBI); and six control areas. Comparable pre- and post-intervention periods were established to measure results and changes over time.

  • Saturation patrol resulted in fewer high priority calls for service
  • Saturation patrol resulted in fewer crimes, although the impact was modest
  • Network-based interventions resulted in fewer high priority calls for service during the first year of implementation but increases in the second year. 
  • Crime in network-based intervention experienced few changes in overall crime, and violent crime in these MHSs increased relative to control areas
Violence Reduction