Study
McEwen (2009a) used a pre- and post-test quasi-experimental design to study the effect of adding four crime scene specialists to two of the four investigative squads in the Phoenix Police Department?s homicide unit. The specialists were assigned to the two squads in the experimental group (the Homicide Clearance Project), while the other two squads functioned as normal (comparison group). Data was collected at baseline, July 1, 2003 to June 30, 2004, and during the test period, September 1, 2004 to June 30, 2005. During baseline, the two experimental squads handled 98 cases and 105 victims, while the comparison squad handled 97 cases and 104 victims. Throughout the test period, the experimental squads handled 83 cases with 93 victims; the comparison squads had 84 cases and 90 victims. Police shootings were excluded from homicide statistics for both squads at baseline and during the test period.
The Phoenix Police Department?s homicide unit (and four squads within the unit) covered the entire geographical area of Phoenix, meaning there were no discernible patterns to indicate certain squads operated in certain areas of the city. However, to ensure there were no differences in the geographic patterns of homicides investigated by the experimental and comparison squads, the bivariate K-function statistical method was used. Two tests with the cross-K statistic were conducted: the first compared the spatial pattern of homicides covered by the experimental and comparison group during baseline; the second compared the spatial pattern of homicides covered by groups during the test period. No significant differences were found related to spatial distance between areas covered by the experimental group and areas covered by the comparison group. Further, chi-square tests were conducted to determine if there were differences in characteristics of victims across squads and time periods. No significant differences were found in victim characteristics from baseline to the test period.
Data was drawn from PACE, the Phoenix Police Department?s record management system, and the Laboratory Information Management System (LIMS). The initial (original) homicide report and supplemental reports are recorded in PACE. The report includes basic descriptions of the circumstances surrounding the homicide, witnesses, victim information, vehicle descriptions, suspect descriptions, and other related information. Supplemental reports were prepared by other investigators involved in the case, scene agents, lab personnel, and patrol officers. Patrol officers, investigators, crime scene specialists, latent print examiners, and forensic scientists all have access to PACE. Cases included in the evaluation had an average of 50 supplemental reports. Information was coded from PACE for all 362 cases, and each case was assigned a non-identifying number. LIMS contained detailed information on the results of a forensic evidence analysis conducted by those in the crime laboratory. Forensic evidence analysis included entry of latent prints into the state identification system so victims, suspects, and witnesses could be identified; analysis of firearm casings; analysis of firearms to determine whether they may have been the weapon used in the homicide; analysis of gunshot residue; and the development and comparison of DNA profiles.
To evaluate the impact of the Phoenix Homicide Clearance Project on homicide clearance rates, the following outcomes were measured: case clearances by arrest, victim clearances by arrest, which takes into account that some cases have multiple victims, case clearances due to quick action and whodunit arrests, victim clearances due to quick action and whodunit arrests, and quality of scene reports. Quick action includes clearances when the suspect is not immediately known, but due to quick action by the officers and investigators the suspect is identified and arrested in a relatively short period of time. Whodunit refers to clearances that require a substantial amount of investigative effort for an arrest to be made (McEwen 2009a; McEwen 2009b).
No subgroup analyses were conducted.