Evaluation of the Use of Computers in Patrol Cars by the San Francisco Police Department, 1999-2000

The two systematic methods of data collection used for this research project were pencil-and-paper surveys of officers' attitudes toward computers and community policing (Part 1) and direct observation of the behavior of officers on patrol, including measurements of time to complete reports and time engaged in police activities (Part 2). Of the 1,091 sworn members of the San Francisco Police Department, 819 participated in the survey and ride-along observational activities across the three time periods: 163 at Time 1, 205 at Time 2, and 451 at Time 3. The 819 patrol officers were asked to complete surveys on computer attitudes and community policing (Part 1) and a subset of 463 officers participated in the observational activities: 96 at Time 1, 185 at Time 2, and 182 at Time 3 (Part 2). Patrol officers participated in the observational study if they were assigned to one of the shift and sector combinations that was randomly selected for a ride-along. The simplest quasi-experimental design would have been to observe the behavior of patrol officers in five stations that had received MCTs compared to officers in five stations that had not yet received them. However, the extraneous effects of observation potentially contaminate a design that simply compares an experimental group to a control group. Therefore, the researchers used the Solomon four-group design to estimate and control for the effects of observation. The basic unit of measurement and analysis was the ride-along. Random combinations of shift halves and sectors were used to develop the ride-along schedule. The shift half was a five-hour period that occurred either during the first half or the second half of the shift. During their shifts, officers were assigned to a district sector, and they patrolled that sector for the entire time. The shift halves and sectors that were randomly sampled at Time 1 were sampled again at the successive measurement periods. Time 1 was the pre-test period before the MCTs were installed. Time 2 was the post-test period during which some groups had MCTs and some groups did not, allowing the researchers to control for the effects of observation. Time 3 was the second post-test period during which all groups had MCTs. Due to the longitudinal nature of the study, it was not possible to measure individual patrol officers as the unit of analysis. Officers were rotated on a regular basis from one district station to another. The nature of police work differed significantly between district stations, so measuring the same officer at two different stations would introduce a confounding factor into the study. Within a district station, shift and sector provided a level of analysis that was consistent across the different measurement periods, and thus was less likely to introduce extraneous factors into the study. Prior to data collection activities at each district station, members of the research team attended roll call at all of the shifts in order to provide information about the study to patrol officers. At the beginning of a given ride-along, the study was explained to the patrol officer, and his or her informed consent was obtained. The officer was told that participation was voluntary and that he or she could withdraw from the study at any time. During the five-hour ride-along, the observer collected information on the amount of time the officer spent on various secondary activities. At the end of the ride-along, the officer was asked to complete the Computer Attitude and Community Policing Attitude Surveys. Survey responses and observational data were anonymous and could not be linked to the officers. Only the ride number identified the observations and surveys. Officers who did not participate in a ride-along were asked to complete the surveys and place them in a drop box at the district station. These surveys were anonymous.