A Mixed-Methods Evaluation of the Phoenix Crime Gun Liaison Program: Leveraging Patrol Officers for Investigations

Abstract

Despite considerable attention to gun violence, crime gun investigations have received limited empirical focus. Research suggests that reliance on specialized investigators and long periods of time between gun crimes, evidence collection, and evidence processing are barriers to timely intelligence. The Phoenix Police Department Crime Gun Liaison Program (CGLP) intended to close this gap by increasing the capacity of patrol officers to investigate gun-related offenses. To examine the impact of the CGLP, a randomized control trial was used to select 16 treatment and 16 matched control officers. We supplement quantitative analyses with qualitative results from focus groups. Treatment officers were referred to and responded to significantly more firearm-related incidents, which resulted in collecting significantly more casings, firearms, fingerprints, and DNA samples during the six-month study period than control officers. Treatment officers also generated a significantly higher number of NIBIN leads and arrests. However, incident-level analyses provided less consistent results, though incidents involving treatment officers remained more likely to result in arrest. Qualitative results support these findings, with officers finding value in the program and reporting improved ability to respond to gun crime and reduced time spent processing scenes. Leveraging patrol officers can improve police agency capacity for responding to gun violence.

Keywords:

• Firearm
• investigations
• ballistic evidence
• gun crime
• police

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 Officer pairs were not perfectly matched on gender. Three female officers were randomly assigned to the treatment group and two were randomly assigned to the control group. However, the results in Table 1 indicate that this did not result in meaningful effect size differences between groups (g=-0.17).

2 Firearm-related incidents were identified by the crime analysis unit based on call type information. These calls included discharging a firearm within city limits, prohibited possession, and shots fired, as well as aggravated assaults, domestic violence, and burglary incidents that mentioned the presence of a firearm. All incidents flagged with a CGLP officer response were also provided. It is important to note that before the CGLP program, many of these incidents would not have resulted in a detective response (e.g., in cases of shots fired without a victim), or would have been responded to by a patrol officer, which frequently resulted in the attainment of low-quality evidence. CGIU personnel reported that detectives often had to go back out to crime scenes, collect additional evidence, re-interview witnesses, and were solely responsible for writing search warrants. As such, this program was meant to enable evidence collection in cases that previously would not have received a specialized response and to improve the quality and timeliness of evidence collected by patrol officers responding to gun crimes. Unfortunately, the specific incident type was not provided at the individual incident level. As a result, we are not able to examine differences across the types of incidents CGLP and control officers responded to, or whether those differences could drive differential case outcomes across groups. This limitation that should be considered in future research examining similar interventions.

3 As an alternative to chi-square tests, we also examined the data using Fishers Exact tests to address the small sample size and imbalance among some incident-level variables. The findings were substantively the same as the results of the chi-square tests.

4 As an alternative to binary logistic regression, we also examined Exact Logistic Regression models to address the imbalance between the number of CGLP and control incidents. The findings were substantively the same using both approaches.

5 Recall that some incidents involved both CGLP officers and control officers (n = 44). Per PPD informational bulletins about the program, any officer responding to a firearm-related incident is supposed to request a CGLP officer response. As such, the CGLP implementation inherently resulted in some level of treatment contamination, although we attempted to minimize this concern by intentionally selecting paired treatment and control officers who work different days of the week. To assess whether this contamination influenced our findings, we conducted supplemental analyses restricted to those incidents that only involved CGLP officers (n = 195) compared to those incidents that only involved control officers (n = 23). There were no differences in the direction or significance of the results using this alternative approach. We chose to present the comparisons between incidents involving any CGLP response to more accurately reflect the program in practice. Supplemental results are available on request.

6 We examined supplemental multivariate models that also control for month based on the differences in months between incidents involving CGLP officers and those involving control officers at the bivariate level. The results did not differ substantially in terms of significance or magnitude. We present the models without controlling for monthly effects to preserve statistical power and to provide more parsimonious results.

Additional information

Funding

This work was supported by the Bureau of Justice Assistance (BJA), a branch of the U.S. Department of Justice (DOJ), Office of Justice Programs (OJP) under grant number 2019-WY-BX-0002. Views expressed here do not necessarily represent the official position of the BJA, DOJ, or the Phoenix Police Department. The authors would like to thank Lieutenant James Hester, Assistant Crime Lab Administrator Benjamin Swanholm, and CGIU/NIBIN Supervisor Jessica Ellefritz from the Phoenix Police Department and Dr. Natalie Kroovand Hipple from Indiana University Bloomington. Study procedures were reviewed and approved by the Institutional Review Board housed within The Office of Research Integrity and Assurance at Arizona State University.