sign in sign up
<<
Home , Law enforcement, Police, Police accountability, Police officer

U.S. Department of Office of Justice Programs National Institute of Justice National Institute of Justice R e s e a r c h i n B r i e f July 2001

Issues and Findings Early Warning Systems: Responding Discussed in this Brief: A system- atic study of early warning systems to the Problem Officer designed to identify officers who may be having problems on the job by Samuel Walker, Geoffrey P. Alpert, and Dennis J. Kenney and to provide those officers with the appropriate counseling or train- ing. The findings are based on a It has become a truism among police providing counseling or training to help survey of 832 local enforce- chiefs that 10 percent of their officers them change their problematic behavior. ment agencies and site visits to cause 90 percent of the problems. Inves- three departments with established By 1999, 39 percent of all municipal and early warning systems. tigative journalists have documented departments in which as few as 2 percent agencies that Key issues: A growing body of evi- of all officers are responsible for 50 per- serve greater than 50,000 dence indicates that in any police 1 people either had an early warning sys- department a small percentage of cent of all citizen complaints. The phe- officers are responsible for a dispro- nomenon of the “problem officer” was tem in place or were planning to imple- portionate share of citizen com- identified in the 1970s: Herman Goldstein ment one. The growing popularity of plaints. Early warning systems help noted that problem officers “are well these systems as a remedy for police supervisors identify these officers, misconduct raises questions about their intervene with them, and monitor known to their supervisors, to the top their subsequent performance. administrators, to their peers, and to effectiveness and about the various pro- the residents of the areas in which they gram elements that are associated with Even though early warning systems effectiveness. To date, however, little has are becoming more popular among work,” but that “little is done to alter their 4 law enforcement agencies, little conduct.”2 In 1981, the U.S. Commission been written on the subject. This Brief research has addressed the effec- on Civil Rights recommended that all reports on the first indepth investigation tiveness of such programs. This police departments create an early warn- of early warning systems. The investiga- Brief reports on a study that estab- tion combined the results of a national lishes a baseline description of early ing system to identify problem officers, warning system programs and asks those “who are frequently the subject of survey of law enforcement agencies with some fundamental questions: complaints or who demonstrate identifi- the findings of case studies of three Are early warning systems effec- able patterns of inappropriate behavior.”3 agencies with established systems. tive in reducing mis- conduct? An early warning system is a data-based How prevalent are early Are some types of early warning police management tool designed to iden- warning systems? systems more effective than others? tify officers whose behavior is problemat- ic and provide a form of intervention to As part of the national evaluation of What impact do early warning early warning systems, the Police systems have on the departments correct that performance. As an early in which they operate? response, a department intervenes before Research Forum—funded by such an officer is in a situation that war- the National Institute of Justice and the Do early warning systems have Office of Community Oriented Policing unintended and undesirable rants formal disciplinary action. The effects? system alerts the department to these Services—surveyed 832 sheriffs’ offices and municipal and depart- Key findings: Twenty-seven individuals and warns the officers while percent of local law enforcement ments serving populations of 50,000 or agencies serving populations of at least 50,000 had an early warning

continued… Support for this research was provided through a transfer of funds to NIJ from

COMMUNITY ORIENTED POLICING SERVICES the Office of Community Oriented Policing Services. U.S. DEPARTMENT OF JUSTICE R e s e a r c h i n B r i e f

Issues and Findings more.5 Usable responses were received Intervening with the officer. The pri- …continued from 571 agencies, a response rate of mary goal of early warning systems is to 69 percent. The response rate was signifi- change the behavior of individual officers system in 1999; another 12 per- cent were planning to establish cantly higher for municipal agencies than who have been identified as having prob- such a program. for ’s departments. lematic performance records. The basic intervention strategy involves a combina- Larger agencies were more likely Approximately one-fourth (27 percent) of than smaller agencies to use an tion of and education. The early warning system. Among the surveyed agencies had an early warn- theory of simple deterrence assumes that agencies with 1,000 or more ing system in 1999. One-half of these officers who are subject to intervention sworn officers, 79 percent had or systems had been created since 1994, will change their behavior in response to a planned to have an early warning and slightly more than one-third had been 7 system; only 56 percent of agen- perceived threat of . General cies with between 500 and 999 created since 1996. These data, combined deterrence assumes that officers not sub- sworn officers had or planned to with the number of agencies indicating ject to the system will also change their have such a program. that a system was being planned (another behavior to avoid potential punishment. No standards have been estab- 12 percent), suggest that such systems will Early warning systems also operate on lished for identifying which offi- spread rapidly in the next few years. the assumption that training, as part of cers should participate in early warning programs, but there is Early warning systems are more preva- the intervention, can help officers general agreement that a number lent among municipal law enforcement improve their performance. of factors can help identify prob- lem officers: citizen complaints, agencies than among county sheriffs’ In most systems (62 percent), the initial firearm-discharge reports, use- departments. intervention generally consists of a review of-force reports, civil litigation, by the officer’s immediate supervisor. resisting- incidents, and How does an early warning pursuits and vehicular accidents. Almost half of the responding agencies system work? (45 percent) involve other command offi- Data from the three case-study agencies (in Miami, , Early warning systems have three basic cers in counseling the officer. Also, these and New Orleans) indicate the phases: selection, intervention, and systems frequently include a training class following: postintervention monitoring. for groups of officers identified by the sys- In spite of considerable differ- tem (45 percent of survey respondents). ences among the programs, each Selecting officers for the program. program appeared to reduce No standards have been established Monitoring the officer’s subsequent problem behaviors significantly. for identifying officers for early warning performance. Nearly all (90 percent) the Early warning systems encour- programs, but there is general agreement agencies that have an early warning sys- age changes in the behavior of about the criteria that should influence tem in place report that they monitor an supervisors, as well as of the officer’s performance after the initial identified officers. their selection. Performance indicators that can help identify officers with prob- intervention. Such monitoring is generally Early warning systems are lematic behavior include citizen com- informal and conducted by the officer’s high-maintenance programs that immediate supervisor, but some depart- require ongoing administrative plaints, firearm-discharge and use-of-force attention. reports, civil litigation, resisting-arrest ments have developed a formal process of observation, evaluation, and reporting. A caveat is in order about the find- incidents, and high-speed pursuits and 6 Almost half of the agencies (47 percent) ings reported here. The research vehicular damage. design was limited in a number of monitor the officer’s performance for 36 ways, and each of the early warn- Although a few departments rely only on months after the initial intervention. Half ing systems studied operates in the citizen complaints to select officers for of the agencies indicate that the followup context of a department’s larger intervention, most use a combination of period is not specified and that officers commitment to increased account- ability. It is impossible to disentan- performance indicators. Among systems are monitored either continuously or on gle the effect of the department’s that factor in citizen complaints, most a case-by-case basis. culture of accountability from that (67 percent) require three complaints in of the early warning program. a given timeframe (76 percent specify a Target audience: and local 12-month period) to identify an officer. law enforcement administrators, planners, and policymakers; researchers; and educators. 2 R e s e a r c h i n B r i e f

Limitations of the survey Also, the three systems differ from Characteristics of officers identified findings one another in terms of structure and by early warning systems. Demo- administrative history, and the three graphically, officers identified by the The responses from the national survey departments differ in their history systems do not differ significantly from should be viewed with some caution. of police officer and the control group in terms of race or Some law enforcement agencies may accountability (see “Three , ethnicity. Males, are somewhat overrep- have claimed to have an early warning three stories”). resented and females are underrepre- system when such a system is not actu- sented. One disturbing finding was a ally functioning. Several police depart- One goal of the case studies was to slight tendency of early warning offi- ments created systems in the 1970s, evaluate the impact of early warning cers to be promoted at higher rates than but none of those appears to have sur- systems on the officers involved. In control officers. This issue should be 8 vived as a permanent program. New Orleans, citizen complaints about the subject of future research, which officers in the early warning program should to identify more pre- Findings from three were analyzed for 2-year periods cisely whether some departments tend case studies before and after the initial intervention. to reward through promotion the kind of Officers subject to early warning inter- The research strategy for the case active (and possibly aggressive) behav- vention participate in a Professional studies was modeled after the birth ior that is likely to cause officers to be Performance Enhancement Program cohort study of juvenile delinquency identified by an early warning system. (PPEP) class; their critiques of the conducted by Wolfgang and col- class were analyzed and a 2-day class The impact of early warning sys- leagues.9 They found that a small was observed to determine both the tems on officers’ performance. group within the entire cohort (6.3 per- content of the intervention and officer Early warning systems appear to have cent of the total) were “chronic delin- responses to various components. a dramatic effect on reducing citizen quents” and were responsible for half complaints and other indicators of prob- of all the serious committed by Demographic and performance data lematic police performance among the entire cohort. The early warning were collected in Miami–Dade and those officers subject to intervention. concept rests on the assumption that Minneapolis on a cohort of all officers In Minneapolis, the average number of within any cohort of police officers, a hired in certain years—whether or not citizen complaints received by officers small percentage will have substan- they were identified by the early warn- subject to early intervention dropped by tially worse performance records than ing systems. The performance data 67 percent 1 year after the intervention. their peers and, consequently, will included citizen complaints, use-of- In New Orleans, that number dropped merit departmental intervention. The force reports, reprimands, suspen- by 62 percent 1 year after intervention research was designed to confirm or sions, terminations, commendations, (exhibit 1). In Miami–Dade, only 4 per- refute the assumption. and promotions. Other data were col- cent of the early warning cohort had lected as available in each site. Three police departments were cho- zero use-of-force reports prior to inter- sen for the case study investigation: These records were sorted into two vention; following intervention, 50 per- Miami–Dade County, Minneapolis, and groups: officers identified by the early cent had zero use-of-force reports. New Orleans. The three sites represent warning system and officers not iden- Data from New Orleans indicate that large urban areas, but the size of each tified, with the latter serving as a con- officers respond positively to early police force varies considerably: At trol group. The performance records of warning intervention. In the time of the study, Miami–Dade had the early warning group were analyzed evaluations of the PPEP classes, 2,920 sworn officers, New Orleans had for the 2-year periods before and after officers gave it an average rating of 7 1,576 sworn officers, and Minneapolis the intervention to determine the on a scale of 1 to 10. All of the officers had 890 sworn officers. impact of the intervention on the offi- made at least one positive comment cers’ behavior. The analysis controlled The three sites were chosen for sever- about the class, and some made specific for assignment to patrol duty on the al reasons. Each has an early warning comments about how it had helped assumption that citizen complaints and system that had been operating for at them. Officers in the PPEP class that use-of-force incidents are infrequently least 4 years at the time of the study. was directly observed were actively generated in other assignments.

3 R e s e a r c h i n B r i e f engaged in those components they issues, training in specific law enforce- significant effects on supervisors. The perceived to be related to the practical ment techniques, stern warning about existence of an intervention system problems of police work, particularly possible discipline in the future) or communicates to supervisors their incidents that often generate com- whether certain aspects are more effec- responsibility to monitor officers who plaints or other problems. Officers were tive for certain types of officers. have been identified by the program. disengaged, however, in components The New Orleans program requires that they perceived to be abstract, The impact of early warning sys- supervisors to monitor identified moralistic, or otherwise unrelated to tems on supervisors. The original officers under their command for 6 practical aspects of police work. design of this study did not include months and to complete signed evalua- evaluating the impact of these systems tions of the officers’ performance every This study could not determine the on supervisors. Nonetheless, the quali- 2 weeks. Officials in Miami–Dade most effective aspects of intervention tative component of the research found think that their system helps ensure (e.g., counseling regarding personal that these systems have potentially

Three cities, three stories T he three early warning systems reports. Monthly reports list employees monitoring of those whose performance in the sites selected for the case studies who received two or more complaints is problematic. have different administrative histories and during the previous 60 days, regardless Minneapolis. When the study began, the program structures, and the three police of disposition. Minneapolis Police Department (MPD) had departments have different histories with The reports are disseminated through a mixed reputation and was in transition regard to police officer use of force and the chain of command to the supervisors under the leadership of a relatively new accountability. of each officer identified. As one official chief. MPD has long had a national repu- Miami–Dade County. The Miami–Dade described the system, supervisors use tation as a police department receptive Police Department (MDPD) currently the reports “as a resource to determine to research. At the same time, however, enjoys a reputation for high standards if job stress or performance problems MPD had a troubled local reputation with of professionalism and accountability to exist.”a The information is intended to respect to the use of force by its officers. reforms instituted following controversial help supervisors evaluate and guide an This reputation eventually brought a num- racial incidents in the late 1970s and early employee’s job performance and conduct ber of important political and administra- 1980s. in conjunction with other information. tive changes in the 1990s. The mayor declined to reappoint the incumbent As a result of the real and perceived prob- The intervention phase of EIS consists pri- police chief, who had failed to discipline lems between police and citizens, the marily of an informal counseling session the police officers. The new police chief Dade County Commission enacted legisla- between the supervisor and the officer. began raising standards of accountability; tion that opened to the public the internal The supervisor is expected to discuss the among other reforms, he instituted a ver- investigations conducted by MDPD. In report with the officer and determine sion of the COMPSTAT process. These addition, an employee profile system (EPS) whether further action is needed. Such changes have had direct implications for was created to track all complaints, use- actions may include making referrals to the system of accountability within the of-force incidents, commendations, disci- employee assistance programs inside or MPD and complicate any attempt to eval- plinary actions, and dispositions of all outside the department, such as psycho- uate the impact of MPD’s early warning internal investigations. As an offshoot logical services, stress abatement pro- system. of the EPS, MDPD created the Early grams, or specialized training programs. Identification System (EIS) under the The program was established in the early Postintervention monitoring of officers in supervision of the Internal Review Bureau. 1990s and has undergone a number of the early warning system is informal and significant administrative changes, includ- MDPD’s EIS began operating in 1981. conducted by supervisors. Review of offi- ing a period of slightly more than 1 year Quarterly reports list all officers who cers’ performance records is designed to in the mid-1990s when the system receive two or more citizen complaints identify officers who continue to exhibit ceased functioning altogether. After the that were investigated and closed or who patterns of misconduct and to make the data collection period for this study, a were involved in three or more use-of- officers aware that their performance is new procedure was instituted that calls force incidents during the previous 3 being closely scrutinized. Additionally, the for reviewing all reports of potentially months. Annual reports list officers who program puts supervisors on notice that problematic officer performance every were identified in two or more quarterly their responsibilities include the close

4 R e s e a r c h i n B r i e f that supervisors will attend to potential The impact of early warning sys- a sophisticated research design. The problem officers under their command. tems on the rest of the depart- qualitative findings suggest that an In this respect, the systems mandate ment. The original design of this study effective early intervention program or encourage changes in supervisor did not include evaluating the impact of depends on a general commitment to behavior that could potentially affect these systems on the departments in accountability within an organization. the standards of supervision of all which they operate. Nonetheless, the Such a program is unlikely to create or officers, not just those subject to early qualitative component identified a foster a climate of accountability where intervention. Furthermore, the system’s number of important issues for future that commitment does not already exist. database can give supervisors relevant research. The extent to which a system information about officers newly changes the climate of accountability The data developed as a part of an assigned to them and about whom within a is not early warning system can be used to they know very little. known, and identifying it would require effect changes in , procedures, or training. Presumably, such changes

Three cities, three stories (continued)

2 weeks. This procedure substantially while under investigation by the depart- The PPEP class consists of an overview heightens the intensity of the level of ment and 105 officers were either arrest- and explanation of the program and units supervision. Thus, the findings reported ed or issued a citation for a on human behavior, stress management, here do not reflect current practices in violation. These are extremely high figures conflict management, complaint avoid- the department. compared with police departments of ance, sensitivity training, “extraneous similar size.b contributors to conflict” (such as sub- The only selection criterion for the system stance abuse), and techniques and assess- is citizen complaints. The formal selection The officials associated with NOPD’s ment (which includes training related to criteria have changed over the years, Professional Performance Enhancement such police activities as tactical stops, situ- however. Currently, a quarterly report lists Program (PPEP) have a strong sense of ation assessment, handcuffing, and custo- all officers with two or more citizen com- identification with the program and are dial security). Each class includes a private plaints, whether sustained or unsustained. committed to maintaining and improving counseling session with the instructor, it. The department also conducts random The intervention phase in Minneapolis during which the officer’s record is integrity “stings” to identify possible cor- consists of only an informal counseling reviewed and the reasons for being select- rupt activities by officers. Furthermore, session between the officer and his or her ed for the program are explained. PPEP does not limit its focus to individual immediate supervisor. In the early years, officers, but also examines training, proce- Immediate supervisors are required to supervisors were required to document dures, and supervision.”c monitor each officer for a period of 6 their counseling session in the form of a months after the intervention. During memorandum to the . There is As in Minneapolis, changes in the pro- that period, the supervisor is required to currently no documentation requirement, gram occurred after the data collection observe the officer interacting with citi- and MPD’s program does not include any period. It is likely that the administration zens while on duty and to complete a formal postintervention monitoring. Apart of the program has weakened somewhat, bi-weekly evaluation of the officer’s from the routine supervision applied to all due largely to the retirement or departure performance. officers, officers who are subject to inter- of key individuals. Thus, the findings vention are not subject to formal monitor- reported here do not reflect current ing and no special data are collected on practices in the department. a. Charette, Bernard, “Early Identification of their performance. and Misconduct,” Miami: Officers are selected for the program on Metro-Dade Police Department, n.d., p. 5. New Orleans. In the mid-1990s, the New the basis of three categories of perform- b. “Disciplinary Action Breakdown,” New Orleans Police Department (NOPD) had a ance indicators: incidents involving con- Orleans Police Department, February 9, 1999. national reputation for both corruption flict in arrest and nonarrest situations and c. New Orleans Police Department, Public and use of force by its officers. Between referrals from supervisors. However, inter- Integrity Division, “To Whom It May 1995 and 1998, NOPD terminated an vention is not automatic; Concern,” May 5, 1998. average of slightly more than 18 officers review performance records and exercise per year and imposed an average of more discretion in selecting officers. than 100 suspensions per year. At the same time, 97 officers resigned or retired

5 R e s e a r c h i n B r i e f help reduce existing problems and help Exhibit 1. Annual average number of complaints against officers, before the department maintain and raise its and after intervention standards of accountability. Thus, these systems can be an important tool for Number of complaints organizational development and human 2 resource management.10

The nature of early warning 1.5 systems. A second goal of the case studies was to describe the systems themselves. In all three sites, qualita- 1 tive data gathered from official docu- ments and interviews with key stake- holders yielded a description and 0.5 assessment of the formal structure and administrative history of each 0 program, along with an assessment Minneapolis New Orleans of its place in the larger processes of accountability in the department. Before intervention After intervention

In addition to finding that the early warning systems in the three sites vary should be viewed with caution. As the the intervention delivered was not considerably in terms of their formal first-ever study of such systems, this consistent for the period studied. program elements, the study docu- project encountered a number of Significant changes also occurred in mented that an effective system unanticipated problems with the data. two sites immediately following the requires considerable investment of First, it was not possible to collect ret- data collection period. In one instance, resources and administrative atten- rospectively systematic data on posi- the system was substantially strength- tion. Miami–Dade’s program, for tive police officer performance (e.g., ened. In the other, it is likely that the example, is part of a sophisticated incidents when an officer avoided administration of the system has dete- data system on officers and their per- using force or citizens felt they had riorated significantly; this deteriora- formance. The New Orleans program been treated fairly and respectfully). tion may have begun during the study, involves several staff members, Thus, it is not known whether early affecting the data that were collected. including one full-time data analyst intervention had a deterrent effect on and two other full-time employees who desirable officer behavior. Policing strategies and legal spend part of their time entering data. considerations Second, the early warning systems in Early warning systems should not be each site studied operate in the context Early warning systems and policing considered alarm clocks—they are not of a larger commitment to increased strategies. These intervention strategies mechanical devices that can be pro- accountability on the part of the police are compatible with both community- grammed to automatically sound an department. Given the original research oriented and problem-oriented policing. alarm. Rather, they are extremely com- design, it is impossible to disentangle Community-oriented policing seeks to plex, high-maintenance administrative the effect of this general climate of ris- establish closer relations between the operations that require close and ongo- ing standards of accountability on offi- police and the communities they serve. ing human attention. Without this cer performance from the effect of the Insofar as the systems seek to reduce attention, the systems are likely to intervention program itself. citizen complaints and other forms of falter or fail. problematic behavior, they are fully Finally, the early warning systems in consistent with these goals.11 Limitations of the case study two of the three sites experienced sig- findings. The findings regarding the nificant changes during the years for Problem-oriented policing focuses on impact of early warning intervention which data were collected. Thus, identifying specific police problems

6 R e s e a r c h i n B r i e f and developing carefully tailored ance is unsatisfactory, and has a pro- varies among systems. Some systems responses.12 Early warning systems gram in place to correct that behavior.14 rely on informal monitoring of the approach the problem officer as the subject officers; others employ a concern to be addressed, and the concerns and areas formal mechanism of observation and intervention is the response tailored to for further research documentation by supervisors. The change the behavior that leads to indi- relative impact of different postinter- cators of unsatisfactory performance. Each of an early warning system’s vention monitoring systems on three phases involves a number of individual officers, supervisors, and Early warning systems and traffic-stop complex policy issues. departments requires further research. data. The issue of by Selection. Although the selection cri- police has recently emerged as a nation- One tool among many al controversy. In response to this con- teria for most early warning systems troversy, a number of law enforcement consider a range of performance indi- Early warning systems have emerged agencies have begun to collect data on cators, some rely solely on citizen as a popular remedy for police miscon- the race and ethnicity of drivers stopped complaints. A number of problems duct. This study suggests that these by their officers. related to official data on citizen com- systems can reduce citizen complaints plaints, including underreporting, and other problematic police behavior. An officer who makes a disproportion- have been documented.15 Using a Officers in the three departments inves- ate number of traffic stops of racial or broader range of indicators is more tigated as case studies were involved ethnic minorities (relative to other offi- likely to identify officers whose behav- in substantially fewer citizen com- cers with the same assignment) may ior requires departmental intervention. plaints and use-of-force incidents after be a problem officer who warrants the the intervention than before. In these Intervention. In most early warning attention of the department. Traffic-stop three departments, however, the sys- systems, intervention consists of an information can be readily incorporated tems were part of larger efforts to raise informal counseling session between into the database and used to identify standards of accountability. The effec- the officer and his or her immediate possible racial disparities (as well as tiveness of such a system is reinforced supervisor. Some systems require no other potential problems, such as dis- by (and probably dependent on) other documentation of the content of that proportionate stops of female drivers or policies and procedures that enforce session, which raises concerns about unacceptably low levels of activity). standards of discipline and create a whether supervisors deliver the climate of accountability. Legal considerations of these systems. intended content of the intervention. Some law enforcement agencies may It is possible that a supervisor may An effective early warning system is a resist creating an early warning system minimize the importance of the inter- complex, high-maintenance operation for fear that a plaintiff’s attorney may vention by telling an officer “not to that requires a significant investment subpoena the database’s information on worry about it,” thus reinforcing the of administrative resources. Some sys- officer misconduct and use that informa- officer’s behavior. Involving higher tems appear to be essentially symbolic tion against the agency in alleg- ranking command officers is likely to gestures with little substantive con- ing excessive use of force.13 Several ensure that the intervention serves tent, and it is unlikely that an inter- experts argue, however, that in the cur- the intended goals. Further research vention program can be effective in a rent legal environment, an early warning is needed on the most effective forms law enforcement agency that has no system is more likely to shield an of intervention and whether it is pos- serious commitment to accountability. agency against liability for deliberate sible to tailor certain forms of inter- It can be an effective management indifference regarding police use of vention to particular categories of tool, but it should be seen as only one force. Such a system demonstrates that officers. of many tools needed to raise stan- the agency has a clear policy regarding dards of performance and improve Postintervention monitoring. The misconduct, has made a good faith effort the quality of police services. to identify employees whose perform- nature of postintervention monitoring

7 R e s e a r c h i n B r i e f

Notes 7. Zimring, Franklin, and Gordon Hawkins, 13. Reiter, Lou, Law Enforcement Deterrence, Chicago: University of Chicago Administrative Investigations, chapter 18. 1. “Kansas Police Go After Their ‘Bad Press, 1973. Boys,’” New York Time, September 10, 1991; 14. Gallagher, G. Patrick, “The Liability and “Waves of Abuse Laid to a Few Officers,” 8. Milton, Catherine H., Jeanne Wahl Halleck, Shield: From Policy to Internal Affairs,” in Globe, October 4, 1992. James Lardner, and Gary L. Albrecht, Police Reiter, Lou, Law Enforcement Administrative Use of , Washington: The Police Investigations, chapter 20; and Beh, Hazel 2. Herman Goldstein, Policing a Free Society, Foundation, 1977: 94–110. Glenn, “Municipal Liability for Failure To Cambridge, MA: Ballinger, 1977: 171. Investigate Citizen Complaints Against 9. Wolfgang, Marvin E., Robert M. Figlio, and Police,” Fordham Urban Law Journal XXV 3. Who is Guarding the Guardians? Washington, Thorsten Sellin, Delinquency in a Birth Cohort, (2) 1998: 209–254. DC: U.S. Commission on Civil Rights, 1981: 81. Chicago: University of Chicago Press, 1972. 15. Walker, Samuel, : 4. Kappeler, Victor, Richard Sluder, and 10. Mathis, Robert L., and John H. Jackson, The Role of Citizen Oversight, Belmont, CA: Geoffrey Alpert, Forces of Deviance: eds., Human Resource Management: Essential Wadsworth Thompson, 2001. Understanding the Dark Side of Policing, Perspectives, Cincinnati: Southwestern College Prospect Heights, IL: Waveland Press, 1998. Publishing, 1999: 98–102; and Poole, Michael, and Malcolm Warner, The IEBM Handbook of Findings and conclusions of the research 5. The first wave of the survey occurred in Human Resource Management, : Inter- reported here are those of the authors and do August 1998, with a second wave in October national Thomson Business Press, 1998: 93. not necessarily reflect the official position or 1998 and followup in February 1999. policies of the U.S. Department of Justice. 11. Alpert, Geoffrey, and Mark H. Moore, 6. For discussions of recommended perform- “Measuring Performance in the New Paradigm ance categories, see International Association of Policing,” in Performance Measures for the The National Institute of Justice is a of Chiefs of Police, Building Integrity and System, Washington, DC: component of the Office of Justice Pro- Reducing Drug Corruption in Police U.S. Department of Justice, Bureau of Justice grams, which also includes the Bureau of Departments, Washington, DC: U.S. Statistics, 1993: 109–142. Justice Assistance, the Bureau of Justice Department of Justice, Bureau of Justice Statistics, the Office of Juvenile Justice and Assistance, 1989: 80; and Reiter, Lou, Law 12. Eck, John E., and William Spelman, Delinquency Prevention, and the Office for Enforcement Administrative Investigations: A Problem-Solving: Problem-Oriented Policing Victims of Crime. Manual Guide, 2nd ed., Tallahassee, FL: Lou in Newport News, Washington, DC: U.S. Reiter and Associates, 1998: 18.2. Department of Justice, National Institute of Justice, 1987. This and other NIJ publications can be This study was conducted by Samuel Walker, Ph.D., Professor, University of found at and downloaded from the NIJ Nebraska at Omaha; Geoffrey P. Alpert, Ph.D., Washington State University; and Web site (http://www.ojp.usdoj.gov/nij). Dennis J. Kenney, Ph.D., Rutgers University. Support for the study was provided by NIJ grant number 98–IJ–CX–0002 through a transfer of funds from the Office of Community Oriented Policing Services. NCJ 188565

U.S. Department of Justice Office of Justice Programs National Institute of Justice

Washington, DC 20531 Official Business Penalty for Private Use $300