The purpose of this study was to examine the feasibility and costs of implementing performance indicators as a new way of measuring police performance as a method of improving accountability and police services.
Four sites participated in the study including a large metropolitan police department (Dallas, Texas), a medium sized police department (Knoxville, Tennessee), a small police department (Kettering, Ohio), and a large sheriff's office (Broward County, Florida). Project staff designed, planned, and conducted community surveys, officer surveys, and surveys with persons who had had recent contact -- either voluntary or involuntary -- with one of the four law enforcement agencies. The methods used to conduct the surveys varied.
Community surveys (Part 1) were conducted by telephone between August 12, 2008 and October 27, 2008 by the Schaeffer Center for Public Policy at the University of Baltimore. Calling was done through the Schaeffer Center's CATI system. The Schaeffer Center completed 458 telephone surveys with randomly-selected residents of Knoxville, Broward County, and Kettering. The brief surveys included questions on opinions of police effectiveness and professionalism. It also queried respondents about neighborhood crime and disorder problems and victimization.
A total of 312 officer surveys (Part 2) were conducted in Knoxville, Broward County, and Kettering in the fall of 2008 and spring of 2009. The officer surveys had three parts: job satisfaction items, procedural knowledge items, and questions about the culture of integrity. The surveys were administered differently in participating sites. In Broward County, the surveys were web-based questionnaires that officers took online. In Kettering and Knoxville, the surveys were administered at roll call sessions for all shifts by project staff. For both modes of administration, officers completed the surveys anonymously.
Brief telephone surveys were administered between December 2008 and March 2009 to persons who had a recent contact with the law enforcement officers in Dallas, Knoxville, Broward County, and Kettering. The voluntary contact surveys (Part 3) were completed by 804 individuals who voluntarily contacted the police department to report a crime while the involuntary contact surveys (Part 4) were completed by 761 adults who were stopped by police in any of the four jurisdictions. Surveys in all four jurisdictions were conducted using lists of persons who had had a recent law enforcement encounter supplied by the participating law enforcement agencies. The voluntary and involuntary contact surveys were conducted by RAND adjunct staff in all areas but Dallas. Due to the large number of respondents in Dallas, the telephone surveys were conducted by the Survey Center at George Mason University.
The community survey (Part 1) sample is comprised of 458 individuals including 150 from Kettering, 150 from Knoxville, and 158 from Broward County. Specifically, sampling for the Knoxville and Broward community surveys was based on a random digit dialing methodology with a screening question to ensure that potential respondents lived within the geographic bounds of the jurisdiction being sampled. The Kettering Ohio sample utilized directory assisted numbers since the geographic area of Kettering is much smaller than that of Knoxville or Broward County. The directory listed numbers, purchased from Survey Sampling, Inc. precisely pre-identified respondents who lived in Kettering. Calling was done through the Schaeffer Center's CATI system, which controlled the number of attempts and eliminated calls to each town until the 150 completion target was reached. Up to six attempts were made to contact a household before it was removed from the working sample. Any adult in the household who was contacted by the Schaefer Center's was considered an eligible respondent for the purpose of this survey. The original sample size for each city was 1,500. In Kettering, the target of 150 was reached after 1,277 of the initial 1,500 sample households were contacted. In Knoxville, the target of 150 was reached after contacting 2,898 households. In Broward County, the target of 150 was reached after contacting 3,201 households.
The officer survey (Part 2) sample is comprised of 312 individuals including 83 from Knoxville, 20 from Kettering, and 209 from Broward County.
The Part 3 sample includes 804 respondents who had a recent voluntary contact with law enforcement (reported a crime). The Part 4 sample includes 761 respondents who had a recent involuntary contact (issued a summons or misdemeanor arrest). Specifically, the Dallas samples consisted of persons who reported property crimes and traffic and class C summons cases. In Kettering, the voluntary contact sample was composed of persons who reported a crime (other than sexual assault) and traffic summonses. The Knoxville samples consisted of persons who reported crimes and traffic summonses. In Broward, the voluntary contact sample consisted of persons who reported a crime (other than serious violent offenses) and misdemeanor arrests.
All individuals living within the jurisdiction of the Knoxville, Tennessee and Kettering, Ohio police departments and the Broward County, Florida Sheriff's Office with telephone access in the spring and summer of 2008 (Part 1). All officers who were employed by the Knoxville, Tennessee and Kettering, Ohio police departments and the Broward County, Florida Sheriff's Office in fall of 2008 and spring of 2009 (Part 2). All individuals with telephone access between December 2008 and February 2009 who had voluntary contact (Part 3) or involuntary contact (Part 4) with officers in the Knoxville, Tennessee, Kettering, Ohio, and Dallas, Texas, police departments and the Broward County, Florida Sheriff's Office.
Community surveys (Part 1)
Officer surveys (Part 2)
Voluntary contact surveys (Part 3)
Involuntary contact surveys (Part 4)
computer-assisted telephone interview (CATI)
paper and pencil interview (PAPI)
Part 1 contains 55 variables including police effectiveness, police professionalism, neighborhood problems, and victimization variables. Part 1 also contains reverse-coded variables, demographic variables, as well as police effectiveness, police professionalism, and neighborhood problem scales.
Part 2 contains 55 variables pertaining to organizational strength, leadership, organizational knowledge, and organizational integrity. Other variables include years of service and the officer's assignment type.
Part 3 contains 19 variables and Part 4 contains 19 variables pertaining to satisfaction with the way the police officer or deputy sheriff handled the encounter. Demographic variables (age, race, gender) are also included.
Response rates for the Officer Survey (Part 2) ranged from 14 percent to 25 percent. Response rates for the community surveys (Part 1), voluntary contact surveys (Part 3) and involuntary contact surveys (Part 4) are not available.
Several Likert-type scales were used.