Compstat and Organizational Change in the Lowell Police Department Challenges and Opportunities

1201 Connecticut Avenue, N.W. Washington, D.C. 20036-2636 (202) 833-1460 Fax: (202) 659-9149 E-mail: pfi[email protected] www.policefoundation.org Compstat and Organizational Change in the Lowell Police Department Challenges and Opportunities

James J. Willis Stephen D. Mastrofski David Weisburd Rosann Greenspan

WASHINGTON, DC

i The Police Foundation is a private, independent, nonprofit organization dedicated to supporting innovation and improvement in policing. Established in 1970, the foundation has conducted seminal research in police behavior, policy, and procedure, and works to transfer to local agencies the best information about practices for dealing effectively with a range of important police operational and administrative concerns. Motivating all of the foundation’s efforts is the goal of efficient, humane policing that operates within the framework of democratic principles and the highest ideals of the nation. The Police Foundation’s research findings are published as an information service.

This project was supported by Grant Number 98-IJ-CX-007 by the National Institute of Justice, Office of Justice Programs, U.S. Department of Justice. Points of view or opinions contained in this docu- ment are those of the authors and do not necessarily represent the official position or policies of the U.S. Department of Justice or the Police Foundation.

Additional reports are forthcoming from the larger project, Compstat and Organizational Change, from which this report on the Lowell, Massachusetts, Police Department’s implementation of Compstat is derived. A Police Foundation Report, The Growth of Compstat in American Policing, will describe the national survey that assessed the number of American police agencies using Compstat and measured the degree to which the elements of Compstat were part of their routine and structure. A third report will describe intensive examinations of Compstat’s implementation in three police departments–Newark, New Jersey, Minneapolis, Minnesota, and Lowell, Massachusetts.

©2003 by the Police Foundation. All rights, including translation into other languages, reserved under the Universal Copyright Convention, the Berne Convention for the Protection of Literary and Artistic Works and the International and Pan American Copyright Conventions. Permission to quote readily granted.

Library of Congress Cataloging-in-Publication Data

CompStat and organizational change in the Lowell Police Department : challenges and opportunities / James J. Willis ... [et al.]. p. cm. ISBN 1-884614-19-1 (alk. paper) 1. Lowell (Mass). Police Dept. 2. Police administration—Massachusetts—Lowell. 3. Police administration—Data processing. 4. Law enforcement—Data processing. 5. Crime analysis—Data processing. I. Willis, James J.

HV8148.L85C66 2004 363.2'0285--dc22 2004001900

ISBN 1-884614-19-1

1201 Connecticut Avenue, NW Washington, DC 20036-2636 (202) 833-1460 • Fax: (202) 659-9149 Email: [email protected] www.policefoundation.org ii Table of Contents Foreword ...... v Acknowledgments ...... vi

I. Introduction ...... 1 Background on Lowell and its police ...... 2

II. Overview of Compstat at Lowell ...... 3

III. Origins and Development of Compstat at Lowell ...... 5 The influence of the NYPD ...... 5 The role of the superintendent ...... 5 Government support—City Hall and the Office of Community Oriented Policing Services . . .6 Early Compstats—fuzzy memories and change ...... 7

IV. Research Methods ...... 8

V. Analysis of Lowell’s Experience with Compstat ...... 9 Mission Clarification ...... 9 Internal Accountability ...... 13 Geographic Organization of Operational Command ...... 18 Organizational structure ...... 19 Decision making ...... 19 Coordination issues ...... 22 Geographic versus temporal organization ...... 23 Organizational Flexibility ...... 24 Manpower ...... 26 City politics ...... 27 Rivalry between sectors ...... 28 Data-Driven Problem Identification and Assessment ...... 30 A history of Compstat—overcoming technical and learning obstacles ...... 31 What kinds of data are used? ...... 33 Visibility of crimes at Compstat ...... 33 How are the data collected, processed, and analyzed? ...... 34 How are the data used? ...... 35 Quality of the data—timeliness and accuracy ...... 38 The Crime Analysis Unit (CAU)—autonomy, time constraints, and frustrations ...... 39 Innovative Problem-Solving Tactics ...... 41 Problem solving and brainstorming during Compstat meetings ...... 42 Follow-up ...... 43 Traditional police responses ...... 44 Traditional and innovative police responses ...... 45 Innovative police responses ...... 46 External Information Exchange ...... 48

iii VI. Discussion ...... 51 Mission Clarification ...... 52 Internal Accountability ...... 53 Geographic Organization of Operational Command ...... 54 Organizational Flexibility ...... 55 Data-Driven Problem Identification and Assessment ...... 55 Innovative Problem-Solving Tactics ...... 56 External Information Exchange ...... 57

VII. Conclusions ...... 58

Appendices ...... 61 Appendix I: Lowell Police Department Organizational Chart ...... 63 Appendix II: Map of the City of Lowell and City Neighborhoods ...... 67 Appendix III: Patrol Officer Survey ...... 71 Appendix IV: Sample Compstat Prep Sheet ...... 81 References ...... 87 Authors ...... 89

iv Foreword This report provides another challenge to Compstat’s proponents by showing the program The birth of Compstat dates back to 1994, when to be a tool whose potential is unfulfilled. Accord- former NYPD police commissioner William Brat- ing to Bratton, Compstat meetings created a sense ton asked a team of officers to create a simple data- of immediate accountability that galvanized New base with information about the major crimes that York’s local commands, fostered innovative prob- cities must report to the FBI. A significant change lem solving, and guided the department in ration- in police practice ensued when the simple data- ally allocating resources to precincts that most base became an elaborate program where police needed them. Implementation of the program in entered crime reports into a computer system that the much smaller Lowell Police Department re- sorted them by type. With the continuing evolu- veals, however, that a gap divides the theory and tion of the program, officers began scrutinizing practice of Compstat. the statistics it generated to create maps and charts Lowell, Massachusetts, like New York City, showing notable changes and emerging problem stoked Compstat’s reputation for working miracles spots. Meanwhile, department heads convened in crime-ridden streets. Lowell’s crime rate began regular meetings to discuss crime trends, to ques- to decline in 1994 and continued to drop after the tion district commanders on their responses to department implemented Compstat. Like New crime, and to work out future strategies. York, Lowell conducts biweekly Compstat meet- The systematic use of hard data and height- ings where the department’s leaders question sec- ened accountability to reduce crime has been her- tor commanders on problems and crime spikes. alded as a seminal innovation in police manage- While many cities that use Compstat only call ment. Compstat’s many advocates claim that it has meetings when a particular sector needs attention, spurred the development of innovative, local, Lowell holds regular meetings where command- crime-fighting strategies and improved public ers present statistics on their sectors and face an safety. These perceived successes have caused an unnerving grilling if crime has increased. exuberant Compstat movement to rapidly sweep The theory of Compstat notwithstanding, the nation. A Police Foundation survey found that Lowell’s program was subject to internal conflicts a third of the country’s 515 largest police depart- that made it deviate from New York’s prototype. ments had implemented a Compstat-like program Scarce resources and a veiled sense of competi- by 2000 (Weisburd et al. 2001). tion made commanders reluctant to share re- The aura that surrounds Compstat in polic- sources with sectors that were hardest hit by crime. ing circles stems from the marvels that it report- Lack of training in data analysis and general ex- edly worked in New York, where crime plummeted clusion of rank-and-file officers from the Compstat in the 1990s. Faith in the program is not univer- process bred indifference toward the program sal, however, and a number of detractors have among many department members. Conservative arisen to contest the Compstat dogma. They have attitudes toward crime fighting led to continued argued that crime dropped in New York as a re- reliance on traditional police responses rather than sult of factors such as demographic shifts, the end the innovative, problem-solving strategies that are of the crack epidemic, and a strong economy. They central to the Compstat process. The absence of have pointed to other U.S. cities that saw crime systematic follow-up at Lowell’s Compstat meet- similarly decline in the 1990s though they lacked ings often caused the department to plot strategy the benefit of Compstat. They have also challenged on the basis of officers’ impressions of what had Compstat by questioning the ability of police work previously worked, not on the basis of the data. to significantly affect crime trends that reflect fac- Moreover, the hefty burden of accountability car- tors beyond the control of the police. Bratton ried by sector commanders may have made them launched Compstat in the conviction that police reluctant to try new approaches to problems, can manage for better outcomes, but skeptics have though Bratton had seen accountability as a cata- contended that police reforms, including Comp- lyst that would energize police to attack crime. stat, make little dent on the economic trends and The prospect of being publicly criticized by the social pathologies that spawn crime. superintendent may have made more impact on

v some sector commanders than Compstat’s ven- Acknowledgments turesome ideal of crime fighting. This report suggests that we should temper We are deeply indebted to Superintendent Edward our enthusiasm for Compstat, but it also acknowl- F. Davis III for granting us access to his police de- edges the valuable impact that the program has partment and for his unconditional backing of this made on the Lowell Police Department. The de- project. We would also like to thank Deputy Super- partment’s decision makers have become more intendent Kenneth Lavallee for all his assistance familiar with the use of data and better informed during our time at Lowell, especially when it came about what is taking place in their areas. Sector to administering the officer survey. A note of ap- captains feel more accountable for identifying and preciation also goes to Deputy Superintendent addressing crime problems, and there has been Dennis Cormier and Captain William Taylor for some successful use of innovative, crime-fighting their support. We would also like to acknowledge strategies. At the same time, the endurance of tra- Jill Casey, Stephanie Hunter, and Suzannah Hacker ditional practices and structures appears to have for all their administrative help. In addition, we inhibited Compstat’s potential for innovation. would like to thank Officer Mark Trudel for his Additional training is necessary if police are to put encouragement. Thanks also to Sergeant David more faith in Compstat’s data-driven approach Abbott for explaining the intricacies of the detail. than in time-honored responses to crime, and This case study would have been impossible other members of the department must shoulder without the unfailing support of the sector cap- some of the burden of accountability that weighs tains, Captain Arthur Ryan Jr., Captain Susan so heavily on sector commanders. The most ad- Siopes, and Captain Robert DeMoura. Their can- vanced technology is pointless unless the police did and intelligent remarks contributed mightily themselves understand its value and have the to the overall quality of this project. Moreover, training to use it. By exploring both Lowell’s fail- their thoughtful observations provided invaluable ures and successes in carrying out Bratton’s vision, insights into the complexities, rewards, and chal- this report reveals the fallible, human dimension lenges of modern police work. of the Compstat process. We would like to thank Carol Fitzgerald and Brenda Bond for their immeasurable help in ex- Hubert Williams plaining how Compstat works in the Lowell Po- President lice Department. We very much appreciated their time and effort in answering our queries and pro- viding us with relevant documents. Thanks also to the rest of the Crime Analysis Unit: Officer Tom Lombard, Dawn Reeby, Derek Desrochers, and Robin Smith. Their comments and patience con- tributed significantly to our understanding of the overall Compstat process. We would especially like to thank the dozens of precinct personnel who we interviewed and observed in meetings, as well as those patrol officers who completed our survey. We would also like to thank Ann Marie McNally, a former research associate at the Police Foundation, for her help with the data entry process. Finally, we wish to express thanks to Dr. Lisa Yarkony, Police Foundation research associate and editor of this report, and to Police Foundation Communications Director, Mary Malina, manag- ing editor, who oversaw its production. vi Compstat and Organizational Change in the Lowell Police Department

I. Introduction A National Assessment,” we identified seven core elements of Compstat: (1) mission clarification; What the late Jack Maple, formerly deputy police (2) internal accountability; (3) geographic orga- commissioner of the New York Police Department, nization of operational command; (4) organiza- pithily called “putting cops on dots” has rapidly tional flexibility; (5) data-driven identification of become a fixture in many large police departments problems and assessment of the department’s prob- across the country. Between 1994, when the NYPD lem-solving efforts; (6) innovative problem-solv- first implemented Compstat,1 and 2001, when we ing tactics; and (7) external information exchange completed our national survey, a third of depart- (Weisburd et al. 2001). Using these elements as a ments with one hundred or more sworn officers general framework, we analyzed how Compstat had implemented a Compstat-like program and is being implemented across the country. The 26 percent were planning to do so.2 Widely vaunted project’s initial stage consisted of a national sur- and even referred to as an “emerging paradigm” vey that assessed the number of local police de- in law enforcement, Compstat is an information partments that were using Compstat and measured and management tool that maps crime statistics the degree to which these Compstat elements were and holds command staff more accountable for part of a department’s structure and routine. The the level of crime in their beats (Walsh 2001, 347). second stage involved sixteen short site visits to Its sweeping popularity among police and policy identify emerging patterns and differences among makers has been fueled by a flurry of national Compstat programs across police agencies. Finally, publicity that attributes the recent plummet in we selected three police departments, Newark, New York’s crime rate to Compstat’s innovative New Jersey; Minneapolis, Minnesota; and Lowell, use of geographic information systems technol- Massachusetts, as sites suitable for lengthier and ogy and cutting-edge management principles. more intensive research and sent a researcher to Much of the literature on Compstat consists each department for a period ranging from two to of brief studies that rely heavily on anecdotal evi- eight months. He or she was responsible for gath- dence or concentrate on the nation’s largest and ering detailed information that explained how most exceptional police department, the NYPD.3 Compstat worked at all levels of the police orga- To date, there has been little systematic analysis nization. of the elements of Compstat and their implemen- There are three primary reasons for selecting tation in smaller departments. This report, which the Lowell Police Department as a case study: (1) is an in-depth evaluation of how Compstat works Its high score on our national survey indicated in a much smaller agency, the Lowell Police De- that it had fully implemented Compstat; (2) It had partment (LPD), is part of a project funded by the received considerable publicity as an innovative National Institute of Justice and conducted by the department under Davis’ leadership;4 and (3) It Police Foundation. In an earlier component of the was relatively small compared to most other de- project, “Compstat and Organizational Change: partments that had received publicity for their

1. There is some disagreement about what the acronym “Compstat” actually means. Former NYPD police commissioner William Bratton suggests that it stands for “computer-statistics meetings” (Bratton 1998, 233), but Silverman attributes its name to “compare Stats,” a computer filename (Silverman 1999, 98). Some commentators have collapsed these interpretations and argue that Compstat refers to “computer comparison statistics” (U.S. National Agricultural Library 1998, http://www.nalusda.gov/pavnet/iag/cecompst.htm). 2. Forty-two percent had not implemented a Compstat-like program. For the first detailed analysis on the extent of Compstat’s implementation by police departments across the country, see Compstat and Organizational Change: Findings from a National Survey, Weisburd et al., Police Foundation (2001). 3. James L. Heskett, “NYPD New,” Harvard Business School Report no. N9–396–29 (April 1996); Eli Silverman, NYPD Battles Crime, Northeastern University Press (1999); Phyllis McDonald, Sheldon Greenberg, and William J. Bratton, Managing Police Operations: Implementing the NYPD Crime Control Model Using COMPSTAT, Wadsworth Publishing Co. (2001); Vincent E. Henry, The COMPSTAT Paradigm: Management Accountability in Policing, Business, and the Public Sector, foreword by William J. Bratton, LooseLeaf Law Publications (2002). 4. In May 2002, Superintendent Davis was awarded one of only two Leadership Awards by the Police Executive Research Forum (PERF) for his role “in bringing major changes in the department that greatly benefited the city.” In this context, Compstat was recognized as “an innovative crime-tracking program” (Skruck, 2002).

1 Compstat and Organizational Change in the Lowell Police Department

Compstat efforts, notably, New York City, Balti- Background on Lowell more, New Orleans, Newark, and Philadelphia. and its police The last reason was the most compelling because an examination of Lowell had potential to pro- The City of Lowell, Massachusetts, is located thirty vide insights into the special challenges and op- miles northwest of Boston on the banks of the portunities that arise when small departments try Merrimack and Concord Rivers. Formerly a pow- to institute a program of organizational change erful center for U.S. textile production, Lowell that originated in much larger agencies. On the began to decline during the Great Depression. The one hand, smaller departments typically have city’s long slump came to a halt, however, when fewer resources to mobilize for new tasks, pro- the late Senator Paul Tsongas spearheaded the grams, and structures. On the other hand, they launch of the Lowell Plan in the early eighties. may find it easier to overcome some of the inter- The plan provides for a partnership between pub- nal resistance that is so endemic to large police lic and private sectors to encourage and guide organizations (Mastrofski, Ritti, and Hoffmaster business development in the city. Its ultimate goals 1987). What could Lowell’s experience with are to strike a balance between manufacturing and Compstat teach us and other researchers? knowledge-based industries and to create an at- This report serves three purposes: (1) to pro- tractive downtown that encourages use of public vide a detailed description of Lowell’s Compstat spaces for civic, cultural, and recreational activi- program that should interest police chiefs and ties (Bluestone and Stevenson 2000, 54–56, 72). other police personnel who are curious about According to the 2000 census, Lowell had Compstat; (2) to explain the benefits and chal- 105,167 residents within its fourteen square miles. lenges of implementing the various key elements The city is divided into several neighborhoods and of Compstat; and (3) to use our knowledge of is moderately diverse: 62.5 percent white, 16.5 Lowell to provide some insights into Compstat’s percent Asian, 14.0 percent Latino, 3.5 percent future in law enforcement. black, 0.2 percent American Indian and Alaskan Compstat’s primary goal is to make police or- Native, and 3.3 percent other race or two or more ganizations more rational and responsive to man- races (U.S. Census Bureau 2002). Income levels agement’s direction. The seven elements of Comp- for 2000 are currently unavailable, but in 1990 stat had been discussed in the organizational Lowell’s median family income was $29,351, with development literature and used by numerous 18 percent of families living below the poverty line. police departments for many years before the The Lowell Plan envisions a thriving and livable NYPD launched its Compstat program. The city for the twenty-first century, and an important NYPD’s contribution was to assemble these ele- element of this vision is a strong emphasis on con- ments into a coherent package (Weisburd et al. trolling crime and disorder. Since Edward F. Davis 2001). According to Compstat’s principal propo- III became superintendent in 1995, Lowell has nents, Compstat’s various elements interweave to increased the number of patrol officers from 159 form a program with its own logical integrity and to 210, a 32 percent increase, and implemented a to make Compstat work like a well-oiled machine. nationally recognized community-policing pro- Indeed, Compstat as a crime-fighting tool is intu- gram. As part of a joint city government and com- itively appealing, with its use of sophisticated tech- nology for the timely identification of crime prob- lems and practice of holding middle managers accountable for reducing them (Bratton 1998, Compstat’s primary goal is to 233–39; Maple 1999, 31–33; Kelling and Sousa make police organizations more 2001, 2–3; Silverman 1999, 97–124). Our exami- nation of Lowell’s Compstat, however, challenged rational and responsive to the program’s accepted image as a smoothly func- management’s direction. tioning machine by revealing numerous paradoxes and incompatibilities among its various elements.

2 Compstat and Organizational Change in the Lowell Police Department

munity initiative, the department has established Lowell’s poor now reside, and the East, which cov- seven precinct stations and formed fourteen neigh- ered Lowell’s downtown and contained a large eld- borhood groups that meet regularly with police erly population. All of Lowell’s patrol officers were officers to identify their most pressing concerns assigned to a sector, aside from the handful that and offer suggestions for their resolution. A com- covered various housing developments through- bination of efforts by city government, commu- out the city and were assigned to specialty posi- nity members, and local police led Lowell to win tions. As of August 2000, there were approxi- the prestigious “1999 All-America City Award” mately the same number of patrol officers in each from the National Civic League in recognition of section, with forty-eight assigned to the North, “exemplary community problem solving.” fifty-one to the East, and fifty-two to the West. There were 260 sworn officers and approxi- In addition to the Operations Bureau, the de- mately eighty civilians in the Lowell Police De- partment was divided between an Investigative and partment when we visited in 2000.5 In terms of its Prevention Bureau, also headed by a deputy, and organization, the department consisted of an In- an Administrative Division that answered directly vestigative and Prevention Bureau, an Operations to Davis. The Crime Prevention Division, Evidence Bureau, and an Administrative Division (see Ap- Response Division, Special Investigation Section, pendix I). The bureaus were each headed by a Criminal Investigation Section, and Legal Division deputy superintendent, and the Administrative fell under the command of the deputy in charge Division was supervised directly by the superin- of the Investigative and Prevention Bureau. Finally, tendent. The city’s neighborhoods were divided the Administrative Division consisted of the Ac- into three sectors determined by both census block creditation Section, Budget and Finance Section, population and the presence of physical bound- Communications Section, Detail Section, Informa- aries, such as rivers and roads: North (Pawtucket- tion Technology Section, Intelligence Crime Analy- ville, Centralville, and Belvidere), East (Back Cen- sis Section, Professional Standards Section, Train- tral, Downtown, and South Lowell), and West ing Section, and Employees Assistance Section. (The Highlands and The Acre) (see Appendix II). These sectors or “service divisions” were under the command of a sector captain and contained II. Overview of Compstat within the Operations Bureau, along with the at Lowell Traffic Division, Headquarters Division, and Com- munity Response Division. The Merrimack and This section provides an overview of how Comp- Concord Rivers provided clear, natural bound- stat has operated in recent years. It introduces its aries, with a major road further subdividing the major features and provides a reference point from East and West Sectors. The different sectors cov- which to assess how much Compstat has evolved ered approximately the same area, but their dif- since its inception. An in-depth discussion of the ferent demographic groups and socioeconomic major features of Lowell’s Compstat program will structures presented each sector captain with a appear in subsequent sections. unique set of problems. A department report, Superintendent Davis first implemented the based on the 1990 census, indicated that the North existing format for a Compstat presentation on Sector was the most populous with a population February 22, 2001. It calls for one of the three of 40,635, followed closely by the West with a sector captains to be present at each biweekly population of 40,442, and finally the East with a Compstat meeting and for data to be presented population of 22,617. In general, the North Sec- for all sectors over a six-week period. Meetings tor had the highest median family income of about take place on alternate Thursdays, when approxi- $35,000, compared to the West, where half of mately twenty to thirty members of the depart-

5. In addition to the superintendent and the 210 line officers, there were two deputy superintendents, eight captains, thirteen lieutenants, and twenty-six sergeants.

3 Compstat and Organizational Change in the Lowell Police Department

ment’s top brass, plus a handful of invited guests, many agencies across the country, merely con- sergeants, and patrol officers, file into a large room ducted an annual review of local Part I crime rates at Lowell’s Cross Point Training Facility. Davis sits collected in the FBI’s Uniform Crime Reports at one end of several tables that form an orderly (UCR). The purpose of this brief examination was rectangle and is flanked on either side by mem- to provide the department with a general indica- bers of his command staff. He or one his deputies, tion of its overall success in controlling crime in his absence, begins the meeting at 9 a.m. with during the previous year. In contrast to this rela- some introductory comments. Then the lights dim tively narrow focus, Lowell’s Compstat program to focus attention on the lone figure of the sector plays a continuous and critical role in the depart- captain who stands in the front of the room. Mem- ment-wide process of identifying specific crimes bers of the Crime Analysis Unit (CAU) use laptop as soon as they emerge, driving decision making, computers to project crime data and maps onto a and facilitating problem-solving strategies. nearby screen, and over the course of the next two- Lowell’s CAU now inputs data on a daily basis and-a-half hours the sector captain, who is en- for a wide variety of crimes ranging from aggra- tirely responsible for the policing of his or her area vated assaults to traffic accidents. Some of these of the city, reports on the sector’s crime incidents, data are made available via the mainframe to all trends, and tactical responses. The sector captain department personnel, as well as through roll-call also faces questions, suggestions, and comments announcements and a daily newsletter. The CAU from audience members. Typical remarks might also uses these data to prepare maps, spreadsheets, include: “What are you doing about motor vehicle and descriptive statistics, which are given to sec- breaks on East Street? They seem to be up from tor captains on the Monday before the Compstat the last Compstat period;” “I have always felt that meeting. The presenting captain is then respon- traffic stops are useful for identifying potential sible for examining the detailed analysis for his or suspects;” “Has anyone got any suggestions about her sector—a process that may take a period of how we should deal with this latest outbreak of several hours over the next few days—in order to graffiti in the downtown area?” prepare fully for Thursday’s meeting. In addition The multiple goals of this Compstat meeting to preparing for Compstat, Lowell’s sector captains include eliciting collective input on crime patterns and their executive officers are responsible for and problem-solving strategies; encouraging in- accessing and reviewing all daily police reports formation sharing on crime locations, victims, and from their sectors and for responding to crime suspects; and facilitating the deployment of de- problems. partment resources. In addition, the forum acts as Before Compstat, the use of timely crime data a mechanism for holding the sector captain ac- for the implementation of crime-reduction strate- countable for crime in his or her beat. Even though gies had no place in the organization and opera- conversation is shared around the room, the pri- tion of the LPD. The following section will exam- mary focus of audience members remains on the ine the factors that led to the formation of Lowell’s sector captain. Any failure to provide a satisfac- program. It will also show that departments can tory response to the various inquiries may lead to implement and adopt Compstat with a modest a rebuke from Davis. outlay of resources and can readily adapt it to work The Compstat presentation at Lowell is the within their existing organizational structures. end product of a lengthy process that begins when Hence, Compstat is not a program that is prohibi- an individual patrol officer files an incident re- tively expensive for small departments, nor does port. Once this report is scanned into the depart- it require a great deal of organizational change. ment’s mainframe, the members of the CAU are Indeed, Compstat’s low cost and flexibility con- responsible for inputting relevant crime data into tributed to its rapid development within Lowell’s a database and using these data to aggregate, ana- relatively small police department. lyze, and map crime incidents. Prior to 1996, when Compstat came to Lowell, crime analysis barely existed, since Lowell’s administration, like that of

4 Compstat and Organizational Change in the Lowell Police Department

principles are most evident at the department’s III. Origins and Development of regular, twice-weekly meetings where precinct Compstat at Lowell commanders tell top brass about ongoing crime problems and their efforts to address them. In its Our examination of the implementation and de- use of these basic principles, Compstat represents velopment of Lowell’s Compstat program reveals a transformation in police operations and man- several key points that are worth highlighting: (1) agement, as well as in attitudes toward the capac- Lowell’s program drew heavily on the NYPD ex- ity of law enforcement to influence crime rates perience; (2) Nearly all of the impetus for its (McDonald et al. 2001). implementation came from Davis; (3) The super- According to our national survey, the NYPD’s intendent’s innovative reforms received strong experience has had a powerful impact on depart- political support from city hall, while external ments across the country, since about 70 percent government grants enabled the initial formation of police departments with Compstat programs and rapid growth of the CAU; and (4) The Comp- reported attending a Compstat session in New stat format has changed significantly in the few York City (Weisburd et al. 2001). The NYPD also years since its inception. influenced Lowell’s implementation of Compstat as a result of a casual conversation that took place The influence of the NYPD between Davis and Bratton at a promotion cer- emony in New York during 1996. Bratton, as Davis Davis and his command staff tailored Compstat recalled, described Compstat as a useful way of to suit their own department’s priorities but also bringing “a private sector mentality to the public borrowed heavily from the NYPD’s program. Some sector,” and this remark made him eager to try background on the NYPD experience, therefore, out the program. Davis, as one officer remarked, will help frame our understanding of Lowell’s ef- “came back from one of his visits with Bratton forts. When William Bratton became commis- and said, ‘Let’s have a Compstat here.’” Shortly sioner of the NYPD, he sought to transform a slug- thereafter, he followed Bratton’s basic model, or gish, bureaucratic organization with demoralized “took stuff from their [New York’s] game plan personnel into an outfit that responded keenly and book,” as another officer put it, and brought effectively to crime problems. He did so by adopt- Compstat to Lowell. ing management principles advanced by organi- zational development experts, such as using data The role of the superintendent to make informed decisions, giving priority to operational rather than administrative concerns, Davis’ brief encounter with Bratton in 1996 con- and holding key personnel accountable. Commis- vinced him to implement Compstat, but he has sioner Bratton, as a result, managed to “turn remarked that two other factors contributed to its around” the NYPD and reduce crime throughout development. Around the same time, he read the city (Bratton 1998). James Heskett’s case study of the NYPD (1996) as The NYPD’s Compstat is a “strategic control a participant in the Senior Executives in State and system” that identifies and disseminates informa- Local Government Seminar at Harvard’s Kennedy tion on crime problems and tracks efforts to ad- School of Government. Davis explained that this dress them by implementing four basic principles: informative article gave him some ideas for (1) accurate and timely intelligence about crime Lowell’s Compstat program by emphasizing the made available to all levels in the organization; relationship between crime strategies and crime (2) the selection of the most effective tactics for statistics, as well as the value of breaking cities down specific problems; (3) rapid deployment of people into sectors. He also recalled how a meeting with and resources to implement those tactics; and (4) Frank Hartmann at Harvard’s Senior Management “relentless” follow-up and assessment to learn Institute for Police underscored the value of data for what happened and make subsequent tactical ad- driving the decision-making process, a perspective justments as necessary (Bratton 1998, 224). These that mirrored his own organizational philosophy.

5 Compstat and Organizational Change in the Lowell Police Department

Convinced that Compstat would be a useful attributed the ultimate decision to a strategic plan- addition to the department, Davis ran the idea by ning process involving all his command staff. He his civilian staff and commanding officers during also noted that there was generally a lot of diffi- one of the department’s annual strategic planning culty getting officers to think geographically, de- retreats in late 1996. The retreat’s purpose was to spite command’s support for the change. Officers continue exploring the possibility of decentraliz- had mixed responses to the change, according to ing the department geographically as part of a the recollection of one Lowell lieutenant: major transformation toward community polic- ing. The initial impetus for this move to commu- You know, you got a variety of responses at all nity policing had emerged from a 1994 strategic levels. Some were kind of interested. A lot were on the fence. And there were a few adamantly planning session. The department was still imple- opposed to any kind of change whatsoever, menting the strategic plan in 1996 when Davis who feel it’s just constitutionally wrong to suggested that they implement Compstat. change (quoted in Thacher 1998, 35). Proponents of community policing argue that police reform requires geographic decentralization Ultimately, the department moved toward a and devolution of decision making down the chain system that established three sector captains, or of command (Eck and Maguire 2000, 218). In “sector bosses” in department vernacular, each re- keeping with this school of thought, these factors, sponsible for a separate area of the city. In con- rather than the implementation of Compstat, trast to the decision to decentralize, no respon- drove Davis’ decision to reorganize the department dents, including Davis, remembered any resistance under the twenty-four-hour supervision of the within the department to the decision to imple- sector captain (Thacher 1998, 36). The creation ment Compstat. It seems likely that officers re- of this sector structure was certainly consistent garded the implementation of Compstat as a rela- with Compstat’s requirements, but the department tively minor event when compared to the large- went much further. Stating that one of the basic scale, structural transformation implied in the tenets of community policing was “one officer, one move to community policing. Since management neighborhood,” Davis explained that assigning is structured geographically under both Compstat officers to specific sectors encouraged them to get and community policing, Lowell’s Compstat pro- to know the residents on their beats and to be re- gram was easily adapted to the department’s pre- sponsive to community problems. He highlighted existing community-policing model. None of the importance of line officer decision making by those interviewed suggested that Compstat was commenting that he wanted his patrol officers to revolutionary, and our overall impression was that recognize their responsibility for “their beat” and most remembered it as somewhat of a novelty. by quoting former LAPD chief Edward M. Davis, who talked about the significance of “territorial Government support—City Hall imperative.” Ironically enough, as we shall see, and the Office of Community community policing and Compstat operate at Oriented Policing Services cross-purposes in relation to the decentralization of command. Community policing delegates deci- In NYPD Battles Crime, Eli Silverman (1999, 181) sion-making authority as far down the chain of argues that a major factor in Compstat’s success command as possible, while Compstat concen- in New York was the strong “external backing” it trates decision-making power among middle man- received from Mayor Rudolph Giuliani. Similarly, agers and holds them directly accountable to the Davis was appointed superintendent with the top brass. Since district commanders are prima- staunch support of City Manager Richard Johnson, rily responsible for identifying and solving prob- the head and arguably most powerful member of lems, the capacity of the rank and file to exercise city hall. In contrast to New York, this political discretion is necessarily constrained. backing was based upon Davis’ ardent commit- Davis may have been the catalyst for decen- ment to community policing. Davis’ appointment tralizing the organization geographically, but he in 1994 corresponded with federal passage of the

6 Compstat and Organizational Change in the Lowell Police Department

Violent Crime Control and Law Enforcement Act. patterns in the city by listing the location, address, The act provided for over $8 billion over a six- and time of calls to the police. year period to deploy 100,000 additional com- Since its first appointments in 1996, the CAU munity-policing officers and also created the COPS has grown to five full-time members, three of Office to oversee the process and administer ex- whom have master’s degrees in criminal justice, tensive funding for nationwide implementation of and several student interns from local universi- community-policing programs (U.S. Department ties. Further financial support for full-time posi- of Justice 1994). Davis was able to use the new tions and for computer hardware and software resources earmarked for community policing to came from additional local, state, and federal implement Compstat. The years 1994 and 1995, grants that were made available for community therefore, represented a “unique” opportunity for policing and problem-solving policing (Thacher him, as he recalled, since he had both the power- 1998, 51). Clear indications that Compstat does ful support of city government and the sudden not require significant new resources are evident availability of considerable federal and state funds. in Davis’ decision to implement Compstat by In addition, Davis and the department were gain- mobilizing the department’s existing personnel ing popularity with local residents who were and his use of available community-policing grants thrilled when the city established its first com- to expand the CAU rapidly. Davis noted, “The cost munity-policing precinct in Centralville, a tough is mainly people” and estimated that it took only neighborhood in the North sector. $100,000 to hire four additional employees for the Davis reminisced fondly about this unique first year of the program (Anderson 2001, 5). political environment. He remembered watching Compstat’s relatively low cost and adaptability are Senator John Kerry at 3 a.m. on C-SPAN talking likely to contribute to its speedy adoption by about the funds available through the 1994 crime smaller police departments across the nation, as bill, and he commented that any motion dealing our survey indicated. At the time of our survey, in with crime that came before the city council 1999, only 11 percent of departments with be- around this time would pass by a wide margin of tween fifty and ninety-nine sworn officers had either seven to two or eight to one in the depart- implemented Compstat. The popularity of the ment’s favor. As a further illustration of this po- program can be seen, however, by the fact that 30 litical support, he noted that the department’s percent were planning to adopt it (Weisburd et al. budget had doubled in the five years between 1996 2001). and 2000 from $10 million to $20 million. This support transferred into the “considerable leeway” Early Compstats—fuzzy memories that the city granted Davis in the hiring process. and change The COPS Office contributed significantly to the emergence and development of Compstat in Memories fade with time, and after the passage of the Lowell Police Department. To get Compstat several years, it is hardly surprising that individual off the ground, Davis reassigned an input clerk, accounts of the first Compstat sessions at Lowell who was already working in Records, and a patrol differ. Many people gave conflicting statements on officer familiar with databases to form a new Crime whether Compstat was actually implemented Analysis Unit. The creation of the first civilian sometime in 1996 or early in 1997 and how the crime analyst position was supported through a first meetings were managed. Several of those in- COPS MORE grant (Thacher 1998, 51).6 Using terviewed remembered the early Compstats as Computer-Aided-Dispatch (CAD) data, the being confrontational “pressure cookers” where fledgling CAU prepared and disseminated simple Davis played the role of a “hard man” and put crime sheets that helped identify crime trends and command staff “on the spot” by peppering them

6. COPS MORE (Making Officer Redeployment Effective) grants were to help increase the amount of time officers could spend on community policing by covering up to 75 percent of the total cost of technology, equipment, or civilian salaries for one year.

7 Compstat and Organizational Change in the Lowell Police Department

ing meaningful crime changes because the “qual- . . . Lowell’s Compstat program, which ity of analysis was lacking.” Ten days later the department implemented Lowell’s existing Comp- required little change to the existing stat with only one sector captain present at each organizational structure, has evolved meeting. considerably since its inception. IV. Research Methods with questions about crime rates in their sectors. Between October 2000 and June 2001, we ob- One lieutenant recalled that what made Compstat served eight biweekly Compstat meetings and particularly tough was “people did not know seven weekly operations meetings in Lowell. We where they were going with it.” also conducted thirty-one formal interviews with Officers who attended the first Compstats also city and police department personnel including: gave varying accounts on the primary focus of the the mayor, city manager, superintendent, middle meetings. Davis commented that initial Compstats managers or sector captains, civilian staff, captains, centered on “crime issues,” but other attendees lieutenants, detectives, first-line supervisors or remembered spending a significant amount of time sergeants, and patrol officers. We tried to gain the dealing with administrative concerns. One officer trust of department members by guaranteeing commented that Compstat meetings often in- interviewees anonymity, whenever possible, and cluded discussions on the utility of decentraliz- by ensuring confidentiality through our uncondi- ing the entire department, including the Crimi- tional refusal to act as a conduit for information nal Investigation Section. Another recalled that the within the department. Despite some initial sus- early Compstats were “very administrative and picion, most of those interviewed felt comfortable largely concerned questions about the work de- enough to engage in lengthy and candid discus- tail and problems with manpower, while crime sions about Compstat. On average the interviews only occupied about 20 percent of the ‘talk time.’” lasted one-and-a-half to two hours, with many These accounts clearly indicate that Lowell’s running over the allotted time. Compstat program, which required little change We conducted six post-Compstat debriefings, to the existing organizational structure, has each lasting about fifteen to twenty minutes, in evolved considerably since its inception. When we order to help us identify the main crime problems first arrived in October 2000, several captains and in each sector and track responses to these prob- members of the administrative staff had been dis- lems over time. We debriefed Davis or the deputy cussing the possibility of changing a Compstat superintendent who ran Thursday’s biweekly format that had not been altered for two years. At Compstat meeting immediately after the meeting the time, the three sector captains all presented and usually talked to the presenting sector cap- crime data on a biweekly basis. Unfortunately, by tains on the following Monday or Tuesday. We the time the third presenter walked to the podium, distributed surveys to 124 patrol officers in which two to three hours had passed and members of we asked them to describe their involvement in the audience were weary, having difficulty con- Compstat and their views of the program; and we centrating, and less capable of providing useful collected ninety-seven completed surveys, yield- feedback on crime problems (the NYPD had ex- ing a response rate of 78 percent (See Appendix perienced a similar scheduling problem). On Feb- III). We also collected documents that could fur- ruary 12, 2001, Davis met with several of his com- ther our understanding of Compstat, including all mand staff and senior members of the CAU to the Compstat maps, spreadsheets, and crime discuss some changes. He was concerned, as he analyses that were provided to sector captains; told them, that the sector captain who presented internal department memos; research grants; ar- last received short shrift and that the two-week ticles on Lowell; community handouts; and copies reporting period might be too short for identify- of the department’s newsletter, the Daily Bulletin.

8 Compstat and Organizational Change in the Lowell Police Department

We promised respondents that we would do clude specific terms, such as reducing crime by our best to conceal their identities. It was obviously 10 percent in a year, for which the organization impossible to guard against the identification of the and its leaders could be held accountable (Bratton superintendent of police, city manager and mayor, 1998, 252). The establishment of a mission state- as we made clear. It was also challenging in such ment, therefore, helps police agencies to function a small police organization to protect the confi- more effectively by encouraging leaders and line dentiality of those who occupied the few special- officers to commit to a clearly defined goal, like ized and mid/upper-level management positions, crime reduction, that is highly valued by the especially the sector captains. We consequently department’s leadership. Despite these seemingly decided to omit identifying characteristics, such obvious benefits, our analysis of Lowell’s Compstat as a respondent’s ethnicity or number of years in program suggests that mission statements might the department, in our initial draft of this report. resonate differently with the public than they do We also made a concerted effort in our final revi- with those inside the department. Furthermore, sion to identify and amend any text that could mission statements might present a set of chal- possibly breach a respondent’s confidentiality. lenges to police agencies with potentially negative outcomes. A mission statement that is inappro- priate, for instance, or exceeds the organization’s capacities might contribute to organizational dys- V. Analysis of Lowell’s function and ultimately undermine the police Experience with Compstat chief’s credibility if the agency fails to meet its stated goal. Using the seven key components we identified as Our general survey showed that 92 percent of Compstat’s general framework, we compared large departments that had reported implement- Lowell’s Compstat program with data from our ing a Compstat program had also reported that national survey to help assess how typical Lowell’s they “set specific objectives in terms that could program was of programs in other departments. be precisely measured.” In other words, a general We also used our qualitative data to assess the statement that clarified a department’s overall mis- dosage or amount of each element within the de- sion was closely associated with implementing a partment to determine the extent to which each Compstat program, and Lowell, in this sense, was of Compstat’s key components had been institu- typical of other Compstat departments. The over- tionalized throughout the organization. We tried, all mission that it clearly promulgated in much of finally, to elucidate some of the problems associ- its literature and on its new Web site was: “To work ated with Compstat by examining the challenges with the community to reduce crime, the fear of the department faced. This gave us some insights crime, and improve the quality of life in the City into Compstat’s ability to operate as a coherent pro- of Lowell.” gram and a clearer sense of its long-term prospects. Compstat demands that departments establish a clear and specific organizational mission rather Mission Clarification than a general commitment to a broad set of ob- jectives. When our national survey asked, “In the The first element of Compstat is mission clarifica- last twelve months has your agency publicly an- tion. Compstat assumes that police agencies must nounced a goal of reducing crime or some other have a clearly defined organizational mission in problem by a specific number?,” only 49 percent order to function effectively. When Bratton assumed of departments responded in the affirmative, and command of the NYPD, one of top management’s almost a third of these departments reported fo- first tasks was to clarify a mission statement that cusing on “many different goals” (Weisburd et al. embodied the organization’s fundamental reason 2001). Since Lowell had announced a clearer and for existing. In order to convey a clear sense of more specific goal—that of making Lowell the the department’s commitment, top management safest city of its size in the United States—it was reasoned that the mission statement should in- atypical of other Compstat departments.

9 Compstat and Organizational Change in the Lowell Police Department

When Davis was first appointed superinten- he believed that the “articulation” of this goal gave dent, he occasionally met with the late Senator momentum to the Davis’ overall plan to change Paul Tsongas, a resident of Lowell, who initially the department. He told us, “The safest city phrase suggested that, “they come up with a vision for was what brought it all together.” the city.” He did not recall having a particularly Lowell’s mission statement differed from that lengthy discussion with Tsongas on this topic but of the NYPD by not committing the department noted that these meetings enabled them to create to reduce crime by a specific percentage. It also the goal of making Lowell “the safest city of its diverged from the NYPD by primarily targeting size in the United States.” According to Davis, city residents. Commissioner Bratton used the Tsongas believed it was important for them to ar- department’s mission statement to motivate po- ticulate a goal that would be clear to everyone and lice personnel and hold them accountable, while help Lowell with its steep crime rate. Around 1995 this appeared to be a less important consideration to 1996, Davis told us earnestly, “Lowell was get- in Lowell. Davis was more concerned with the ting beaten down,” and coming up with the goal mission statement’s appeal to external constitu- was about “more than just trying to make people ents than to department members, as another re- safer . . . the city’s future was hanging in the bal- spondent recalled: “It was really used for the ben- ance . . . [and] the goal gave people in the city efit of those outside of the department . . . it was some hope.” Not only were city residents in crisis never used within the department . . . it was not in the early 1990s, so was the Lowell Police De- like the department rallied around it . . . the state- partment. Public confidence in the department had ment is not part of the guys’ [line officers] daily... deteriorated to the point that “the association of you know . . . what they talk about.” downtown businesses voted to hire private secu- Since Lowell’s broad mission statement was rity to patrol Lowell’s rapidly deteriorating com- designed to appeal primarily to city residents, its mercial district” (Thacher 1998, 9). implementation did not resonate quite so strongly Given that the mission statement’s intention within the police department. In contrast to a was to rally public support for a department that crime-reduction goal defined by a modest percent- seemed incapable of stemming crime in a danger- age over a finite period (a year, for example), the ous city, it is not surprising that Davis and Tsongas’ adoption of such an ambitious and enduring goal proclamation was unburdened with technical de- as becoming “the safest city of its size in the United tails and emboldened with powerful symbolism. States,” may have further mitigated its impact Similar to the NYPD, crime reduction lay at the within the police organization—it was just too heart of Davis’ mission for the department—a goal ambitious for police officers to incorporate within that he associated closely with Compstat. An im- their daily operations. One officer characterized portant point to note: Davis stated that Compstat the statement as more of a slogan than a mission allowed one to examine data before making any because it was “not realistic,” and he embellished decisions. “If you approach problems any other his claim by comparing it to the New England way,” he commented, “you are allowing the ‘cause Patriot’s recent win in the 2002 Super Bowl: “Just du jour’ to set your priorities when you should be like the Patriots winning the Super Bowl, sure, dealing with the crime rate” (emphasis added). they might have won . . . but now we expect them However, in contrast to the NYPD, the resolution to do it again . . . it is not a practical statement.” to make Lowell “the safest city of its size in the In using this analogy, he was drawing attention to United States” did not contain potentially confus- the unrealistic expectations conjured up by the ing percentage reductions. Furthermore, it ap- “safest city” analogy and the fact that “crime can- pealed directly to residents by creating an attrac- not continue to drop forever.” The ambitious na- tive vision of Lowell as a pleasant or desirable place ture of the mission statement helped explain why to live (again). One respondent stated that the the few times he heard reference to it within the mission statement was a “big deal” and remem- department was when an officer at a crime scene bered that it was popular in the newspaper and joked: “Uh-huh . . . another murder in the safest among community members. More importantly, city in America.”

10 Compstat and Organizational Change in the Lowell Police Department

Some department personnel might then have cially in comparison to more ambiguous strate- regarded the mission statement as unrealistic, al- gies or programs that appear to challenge the ca- though our observations and survey data sug- nonical crime-fighting role of the police. Commu- gested that patrol officers did recognize and ac- nity policing, for example, has been the focus of a cept the relationship between Compstat and the prodigious amount of scholarship over the last department’s approach to fighting crime. Even if twenty years, but its goal and key elements are most doubted the practicality of the department’s still subject to much debate among police practi- “safest city” mission, they endorsed the focus of tioners and academics. What police department, the effort. One patrol officer described Compstat’s however, would not want to adopt a program explicit focus on the identification of crime pat- whose clear purpose is to reduce crime through terns by saying, “It enables the department to give the implementation of a well-defined set of tech- a focused effort on policing as opposed to hap- nologies and procedures? The appeal of Compstat’s hazardly driving around in circles . . . it allows the crime-fighting goal to the police increases the like- department to focus on a specific area.” Results lihood that it will endure. from our patrol officer survey further supported There is an implicit and important consider- this observation that the rank and file clearly as- ation contained within the preceding comments about Compstat’s objective. What measures or benchmarks will the department use to evaluate What police department . . . would not its progress toward a specific goal? This became want to adopt a program whose clear an issue every October or November, when the FBI published its annual Uniform Crime Reports, purpose is to reduce crime through the and Lowell’s crime analysts examined the sixty- implementation of a well-defined set of two U.S. cities with populations between 95,000 technologies and procedures? and 100,000. Lowell’s analysts listed these cities alphabetically and used them to create a table, which they sent to Davis but not to the rest of the sociated the goal of crime reduction with Comp- department. Since the rankings were not dissemi- stat. Approximately 92 percent of those surveyed nated more widely, and we did not observe any responded that “reducing violent crime in the city” specific reference to the “safest city” goal during and “improving the quality of life in the city” were our stay, it appeared that the table served prima- very or somewhat important to the department’s rily to give Davis an annual impression of how Compstat strategy. The power of Compstat’s im- Lowell was doing in relation to other cities of its age as a crime-fighting tool is further reinforced size. In short, the “safest city” goal remained an when we consider that: (1) Davis did not devise a implicit rather than a highly visible element of mission statement that explicitly defined Comp- daily operations. This does not mean that the goal stat’s goal but incorporated it within the depart- was merely symbolic, particularly since Davis de- ment’s goal of making Lowell the safest city of its scribed worrying about the extent to which the size in the nation; and (2) The implementation of department was meeting its goal at every biweekly Compstat was not accompanied by any depart- Compstat meeting. It may suggest, however, that ment-wide training. Despite the absence of these the six-year-old “safest city” imagery had become means of fostering a shared understanding of so commonplace as to no longer provoke much Compstat’s purpose, there was still a general con- interest within the department. sensus that Compstat was a means of refocusing Underlying the simplicity of the Compstat the department’s energy on reducing crime. mission is a more complex set of challenges: What It seems likely that officers had a broad un- happens to Compstat when a police department derstanding that the fundamental objective of fails to meet the crime-reduction goal of its mis- Compstat was to control crime. The simplicity and sion? Will failure to meet the goal lead to cyni- long tradition of this goal in police departments cism both within and outside the department? Will helps to explain the program’s popularity, espe- the end result be the termination of Compstat, as

11 Compstat and Organizational Change in the Lowell Police Department

city residents and police officers question the value hiring of fifty-one new police officers, and it is of the entire program? this increase that might have contributed to the Tsongas and Davis were aware of the danger overall reduction in crime. Furthermore, Lowell’s of establishing a mission statement that set an recent crime increase appears to lend credence to unattainable benchmark for success. Davis com- a more cautious standpoint in attributing crime mented that in coming up with a vision for the reductions to the police in general or, more spe- department, he remembered thinking that the “saf- cifically, to Compstat. In Lowell, index crimes for est city” statement was “kind of reaching.” He 2001 increased 12.7 percent over 2000 (4,507 in- noted that even though he recognized that the dex crimes compared to 3,999). As a result of its “safest city” goal was ambitious, he felt it was tan- increasing crime rate, Lowell fell to the twenty- gible. They both believed, he said, “Lowell was a eighth safest city of its size in 2000.7 After a pe- place that they could get their hands around . . . riod of rapid decline, this reversal has been met that there was real potential for serious gains.” The with some disquiet. In 2000, the local newspaper department has indeed had considerable success reported, “If the past year is any indication, Supt. in achieving its goal, notwithstanding the risks Ed Davis and the Lowell Police Department will involved in setting such an ambitious benchmark have their work cut out for them in the next 12 for success. For cities with populations between months” (Lowell Sun 2000). 90,000 and 110,000, Lowell ranked the forty-fifth Whatever the causes for mounting crime rates, safest in 1993 and jumped to the fifteenth safest some evidence suggests that Davis was feeling in 1997. This drop in crime is among the reasons apprehensive as the local press reported on the why many police administrators and scholars con- slight upturn in crime. In a November 2000 inter- tinue to pay close attention to the department’s view he expressed concern that crime was rising achievements (Lehrer 2001). for the first time in six years, and by December Davis, who was reluctant to attribute the de- 2000 we overheard a comment that Davis was now cline in crime to Compstat alone, cited three ad- paying for previous statements he had made in ditional causes: an improving economy, an increase which he claimed credit for Lowell’s declining in the length of jail sentences, and the efforts of crime rate. One respondent noted that the first the police. Public opinion, however, focused on response to any news of an increasing crime rate the impact of Compstat, as shown in a newspaper was “damage control,” and he expressed disap- headline from October 1999 which reported that pointment that the department was not looking “Crime in Lowell Continues to Plummet” and at- at the 10 percent increase more critically and ask- tributed “much of the success in combating crime” ing questions such as, “What’s different from last to “targeted policing through the Compstat pro- year; what’s happening nationally?” Two recent gram” (Iven 1999). This success, notwithstand- articles in the local paper also suggested that criti- ing, there is reason to be cautious about Compstat’s cism of Davis was becoming more acute. One re- ability to reduce crime. Silverman, like Kelling and marked that Davis’ reluctance to attend city coun- Sousa, has argued that the NYPD’s success in re- cil meetings appeared to reflect a lack of focus on ducing crime was a direct result of its Compstat public safety issues (Scott 2002), while the other program (Silverman 1999, 125–177; Kelling and expressed disappointment with the upward crime Sousa 2001, 2), but many criminologists remain trend given the department’s $20 million budget unconvinced. Crime, they argue, is too complex a (Lowell Sun Online, April 4, 2002). phenomenon to be mitigated by any single ap- These comments illustrate the conflicting pres- proach (Bouza 1997; Eck and Maguire 2000; sures that Compstat imposes on police chiefs and Harcourt 2002). In Lowell’s case, the implemen- their departments. Compstat requires chiefs to tation of Compstat roughly corresponded with the formulate highly visible, public mission statements

7. According to the 2000 Uniform Crime Reports, Lowell experienced 3,803 Part I crimes per 100,000 people. In comparison, the crime rate for Simi Valley, CA, America’s safest city with a population between 90,000 and 110,000, was 1,441 per 100,000 people (US Department of Justice 2001).

12 Compstat and Organizational Change in the Lowell Police Department

that set tangible organizational goals for reducing ment’s failure to live up to its goal—modify the crime, as well as hold them and their departments mission statement. Davis suggested that this was accountable for meeting these standards. Increased currently happening at Lowell. He acknowledged expectations for lower crime rates, therefore, put that the department had recently experienced police chiefs under considerable pressure to claim “some drawbacks with the crime rate,” and con- some responsibility and generate positive press for tinued to say, “Some people have come in and any successes. Absent convincing evidence that given a qualifier” to amend the mission statement. police departments possess the capacity to reduce He noted that it was important to take into ac- crime and that managers have the will and skill to count Lowell’s socioeconomic and demographic mobilize that capacity, setting a specific crime-re- characteristics in comparison to other cities, and duction goal is like a batter of untested capabili- he added that they were now referring to Lowell ties pointing to center field each time he comes to as the “safest city of its size and type in the United bat. This works only if he delivers a home run States” (emphasis added). Of course, the depart- more often than not. ment could decide to drop, rather than merely Mission statements can do more than create refine, its vision entirely, but this seemed very unrealistic expectations; they can also be dysfunc- unlikely at Lowell. Davis said firmly that Lowell’s tional in other ways. An increasing crime rate is “safest city” statement was still “vitally important likely to foster a great deal of public scrutiny and today.” concern over a department’s failure to fulfill its goals. The pressure that this places on a chief and Internal Accountability his organization may provoke a knee-jerk reac- tion from the police (“damage control”) rather For a department’s mission statement to be effec- than a more thorough investigation of the crime tive, workers need to be held responsible for meet- increase. A chief will probably respond to this pres- ing the goals that the department espouses. sure by exhorting his officers to work harder. This, Compstat does this by holding operational com- in turn, might alienate managers and rank-and- manders accountable for knowing their command, file officers who feel the chief is blaming them for being well acquainted with its problems, and mea- crime problems that stem from factors beyond surably reducing them—or at least demonstrat- their control, such as poverty, drugs, and unem- ing a diligent effort to learn from the experience. ployment. Compstat, in short, makes someone responsible for Typical of such officers was one respondent tackling and reducing crime and imposes adverse who specifically noted the impact of broader struc- career consequences, such as removal from com- tural factors to support his earlier comments that mand, on those who fail to comply. In conducting the department’s mission was not attainable in a our fieldwork, we discovered that accountability practical sense: “There are cities of the same size, was experienced most intensely by district com- say in California, where people have much higher manders and far less so by those further down the incomes,” and it is consequently easier to control chain of command. In addition, our research re- crime. The department’s failure to achieve its goal, vealed the paradox that holding officers to a very as we have seen, threatened to breed this type of high standard of accountability inhibited two other cynicism among civilians and department mem- Compstat components: Compstat’s ability to fa- bers who accused Davis, as the easiest target, of cilitate innovative problem solving through brain- disingenuously claiming responsibility for previ- storming and its capacity to reallocate resources ous successes. In short, depending on the depart- to crime problems that most needed them, a com- ment’s capacity to meet its objectives, mission ponent we address more fully under the section statements might motivate the organization to “Organizational Flexibility.” Finally, we discovered succeed or exacerbate its continuing failure to that there are two challenges to the potency of meet those same objectives. accountability’s ultimate threat to replace district These observations suggest an obvious means commanders for poor performance: (1) There may of reducing the dissonance caused by a depart- only be a small pool of suitable replacements who

13 Compstat and Organizational Change in the Lowell Police Department

are willing or able to do the job of district com- sector captain responded that it prevented “slack- mander (this is particularly the case in smaller ing off.” Another sector captain noted that Comp- departments); and (2) Union and civil service re- stat was a way of “keeping them honest” since quirements make it exceedingly difficult to remove “having things up there on a map can show you officers for poor performance. how bad things are, and you cannot say, ‘Ooh, I Responses to our national survey and inten- missed those reports; I did not see them.’” sive site interviews suggested that departments These sector captains clearly recognized and that have implemented a Compstat-like program accepted that Compstat held them accountable for consider internal accountability to be a very im- all that occurred in their respective beats. Their portant feature of Compstat. Almost seven in ten comments also brought out the central role that of these departments told us that a district com- Compstat meetings played in fostering account- mander would be “somewhat” or “very likely” to ability by allowing Davis to visibly assert his lead- be replaced if he or she did not “know about the ership. Since all the command staff attended the crime patterns” in his or her district. Almost eight biweekly meetings, Compstat provided Davis with in ten of these departments told us, in turn, that a an ideal opportunity to display his authority and commander of a specialized unit would be “some- hold his sector captains publicly responsible. what” or “very likely” to be replaced if he or she Compstat may provide a suitable venue, but our regularly failed to fulfill requests for cooperation research suggests that the accountability mecha- from district commanders. A much smaller pro- nism also relies upon the leadership style of the portion of these departments reported that a dis- individual who runs Compstat. trict commander would be replaced simply if crime Davis’ leadership style, as observed at Lowell’s continued to rise in a district. Few departments Compstat meetings, was to constantly ask ques- take this extreme position because Compstat gen- tions and make suggestions. Davis remarked that erally requires commanders to be familiar with he was seeking to foster “data-driven decision problems and develop solutions to them but does making in a learning organization” by interrogat- not hold them too accountable for achieving out- ing his command staff about their responses to comes that may be unresponsive to well-planned various crime problems and encouraging others police interventions (Weisburd et al. 2001). to promulgate helpful solutions. In an interview Internal accountability was an integral part of with David Thacher (1998, 37), a researcher who Lowell’s Compstat program, as it was in many visited Lowell in 1997 as part of a national COPS other Compstat programs examined in our na- evaluation, Davis explained, “You have to be 90 tional survey. In fact, Davis explicitly recognized percent a teacher when you have this job and that’s the importance of this feature when he defined what I do.” He hopes that by asking people, “What Compstat as a means “to manage the police de- they’re working on and how they’ve come to this partment in a timely manner with an eye toward conclusion . . . in front of people . . . eventually accountability.” He was not alone in acknowledg- they’ll get the idea of it—that it’s their responsi- ing the value of this element, as the comments of bility.” other department members reveal. When asked In addition to promoting an information and what was particularly useful about Compstat, one data-driven environment, a chief can use Compstat as an arena to reward or punish his command staff in order to convey his expectations about accept- able performance. Holding command staff ac- Internal accountability was an countable for crime in their beats was a contro- integral part of Lowell’s . . . versial element of the NYPD Compstat program. program, as it was in many other There is a well-known story in police circles that during one Compstat meeting Jack Maple repeat- Compstat programs examined in edly flashed up a slide of Pinocchio while a mem- our national survey. ber of the command staff struggled to explain crime in his precinct (Maple later apologized). In

14 Compstat and Organizational Change in the Lowell Police Department

contrast, Davis recognized that commanding of- tor captains, in particular, accountable suggests ficers might be apprehensive or fearful about be- that there is some general recognition that ac- ing held accountable. Referring to this infamous countability is experienced most acutely by “beating-up” scenario in New York, he argued that middle-level managers. it was “counterproductive” to humiliate individu- The conventional notion that grilling precinct als during Compstat and adopted a more human- commanders on crime-reduction efforts “rein- istic approach: “Police officers are competitive by forces the patrol officer’s desire to combat crime” nature and all you really need to do is give them seems to overstate the case in Lowell (Silverman the facts and ask them a question. They will go 1999, 194–5). Our research suggested a greater from there.” Just asking them a question in front likelihood that the sense of accountability became of their peers, Davis opined, would make sector diluted as one moved down the command struc- captains feel accountable. That said, there have ture. As one patrol officer put it when asked about been occasions when Davis has been so displeased Compstat, “If you don’t go, you don’t know.” by the lack of initiative of his command staff in Nearly the entire command staff attended Comp- response to his queries that he has taken a harder stat, but only two or three patrol officers were line and “balled them out.” present at any given Compstat meeting. They Compstat’s potential for confrontation ac- might answer a question or two, and they might counts for why there is some concern that it can participate in a brief presentation, but they gener- result in command staff members being repri- ally played a peripheral role, leading one high- manded. Outside of the usual command staff par- ranking officer to remark, “Patrol officers can hide ticipants and audience, Compstat has a more pro- in the meeting and get away without saying any- nounced reputation for being brutal. Despite thing.” providing a forum for making “people do the job Just because patrol officers were not regular they do,” many are clearly discouraged by the con- attendees at Compstat did not mean that they did frontational atmosphere that can characterize not learn about or experience accountability; it Compstat meetings. One officer described Comp- suggested, however, that they experienced it less stat as a forum where officers had their “balls intensely. Several comments made to us suggested ripped off” and surmised that this only served to that when someone was chastised in Compstat, make individuals “reluctant to speak up . . . reluc- news of their plight spread quickly throughout the tant to do their job.” department. The roll-call survey results, however, Compstat’s reputation as a pressure-cooker indicated that what happened in Compstat was environment that holds sector captains account- more likely to be communicated informally than able was widespread at Lowell. Even patrol offic- through a systematic process. Stories, anecdotes, ers, who rarely attended Compstat, recognized this and jokes expressing sympathy for, or humor feature. Based on the survey we distributed at roll about, an individual who was unfortunate enough call, 61 percent of officers reported that holding to be chastened by Davis seemed to feed Comp- sector captains accountable for crimes in their stat’s powerful reputation for holding officers ac- beats was “very important” (26 percent) or “some- countable. However, the absence of a formal what important” (35 percent). Only 34 percent mechanism for transmitting this message fre- reported that holding sector captains accountable quently and directly to patrol officers could miti- for crime in their beats was “not at all important.” gate its impact on those at the bottom of the po- Five percent responded that they “don’t know” to lice hierarchy. When asked, “How often does your the question. In contrast, a slightly smaller pro- supervisor discuss what has happened at Compstat portion of officers, 56 percent, responded that, meetings?,” 61 percent of the patrol officers we “Holding officers accountable for crimes in their surveyed responded “never” (43 percent) or “every beats” was “very important” (18 percent) or few months” (18 percent). “somewhat important” (38 percent). The fact that Since accountability relies heavily upon a pub- a greater proportion of patrol officers believed that lic setting with high-ranking officials in atten- Compstat was “very important” for holding sec- dance, it follows that accountability is experienced

15 Compstat and Organizational Change in the Lowell Police Department

less acutely outside of this forum. In brief, a sec- regarding camaraderie and respect curtail the free- tor captain who has been rebuked in Compstat dom to share ideas by prohibiting officers from for an inadequate strategy may return to his sec- sharing potentially useful information that others tor and admonish his line officers, but the force could interpret as criticism. of the message is considerably weakened for three This type of information might appear benign reasons: (1) Compstat ultimately holds middle to outside observers who are unfamiliar with the managers, not line officers, accountable; (2) The department’s “back-stage” environment, but insid- message does not come from the highest ranking ers know the least vague suggestion that a supe- official in the police department, and (3) It, there- rior is uninformed, misinformed, or incompetent fore, does not result in public censure on the same carries significant risk. The adamant refusal of one scale. lieutenant to “embarrass captains in Compstat” Lowell’s Compstat is an independent program, reflects a feature of paramilitary organizations but its impact on accountability relies heavily upon embodied in the popular police maxim, “shit rolls the individual personality and leadership style of downhill.” Since information sharing at Compstat the superintendent, as shown by differences in the may put a member of the command staff at fault, tenor and operation of Compstat meetings. When an officer may take the safest course of action and Davis was present, he tended to dominate Comp- choose not to participate in the meeting at all. He stat meetings. In his absence, Compstat meetings might also take the path of low risk by making were more subdued. This difference in tone only short and simple comments reiterating what stemmed from a number of complex social dy- has already been said. Of course, he could also namics, as Davis pointed out. He was aware that dramatically increase the likelihood of negative Compstat meetings ran differently in his absence repercussions by directly criticizing a superior, and proposed the general nature of police depart- even though we never witnessed this behavior ments as “paramilitary organizations” as a partial during our research at Lowell. explanation for the difference. He observed that Davis openly acknowledged that displays of these kinds of organizations, with their presumed deference limited the value of information ex- emphasis on the command structure, didn’t en- change by remarking that it was “extremely dif- courage “free-thinkers,” but he was optimistic that ficult” to facilitate questions at Compstat. He noted a “system” like Compstat could encourage indi- that an officer who is questioned by a peer or sub- viduals to think differently. Davis’ comments re- ordinate during Compstat might later confront the veal that even though the individual responsible inquisitor and complain, “You really screwed me for running Compstat influences how accountabil- in there.” To avoid this scenario, some members ity is experienced within the department, the or- of the command staff made a practice of giving ganization is simultaneously constrained by larger advance warning to colleagues whom they in- structural limitations. These present significant tended to call on at Compstat. One officer men- challenges to Compstat’s purported goals to en- tioned that he preferred to discuss Compstat- courage information sharing and hold officers ac- related matters in private with Davis, rather than countable for crime in their beats. bring them up at meetings; and Davis similarly The department’s organizational hierarchy, appeared to choose discrete rather than visible embodied in the rank structure, hampers the free settings when directing his most vociferous criti- exchange of crime-related information and prob- cism at command staff. Members of the CAU also lem-solving strategies at Compstat. Personal his- shared this tendency to bring discretion to the tories between officers intertwine with rank to Compstat accountability process and described form a complex set of social norms and relation- doing their best to inform sector captains of any ships. Mutual exchange is clearly restricted by the changes in the Compstat format because Compstat expectation that, on account of their social posi- was about “providing information, not catching tion, higher-ranking officers are primarily respon- people out.” sible for communicating crime information dur- We have already seen how Compstat’s stress ing Compstat. In addition, department norms on accountability conflicts with its emphasis on

16 Compstat and Organizational Change in the Lowell Police Department

brainstorming, limiting effective collaboration but Davis indicated that he would move a sector between department members. The size of the captain who was not doing his or her job at organization and its administrative rules and regu- Compstat to another position. This is not to sug- lations, both factors independent of Compstat, gest that there was no turnover of sector captains. limit, in turn, the level of accountability to which Respondents recalled that some individuals were a chief can formally hold his sector captains. These uncomfortable with the Compstat process (the factors also determine whether Compstat can live public speaking and the intense scrutiny) and up to its claim for predictably improving perform- asked to be reassigned. However, responding to a ance by helping departments achieve greater ac- request for reassignment obviously differs from countability than they have in the past. Salient removing sector captains for failing to fulfill the issues in this regard include how much a chief obligations of their position. Not only was Davis’ can punish those who fail to meet the goals of the capacity to remove a sector captain limited by the organization, whether he can remove them from small pool of qualified candidates, the magnitude their position, and how much reward he can offer of the responsibility commensurate with the po- them. sition, not to mention the administrative burdens, Our research in the relatively small department also made it unlikely that all those who qualified of Lowell suggests that incentives are usually lim- would choose to serve as one of the three sector ited to public acknowledgement of good or bad captains at Lowell. One interviewee pointed out performance. This can be attributed to two orga- nizational features that exist independently from Compstat and are generalizable to other police Our research in . . . Lowell suggests departments: the size of the agency and the rules and regulations that govern its administration. In that incentives are usually limited small departments, the chief’s power is limited by to public acknowledgement of good the small pool of qualified candidates for the po- or bad performance. sition of district commander, lowering the risk of an individual being fired for marginal perform- ance. There are, furthermore, usually few positions above the rank of district commander, as was the that given the significant responsibilities of the case in Lowell, thus undercutting the incentive of position “maybe four out of eight captains [total] promotion. Finally, the strict regulations that gov- are willing to do this job.” Finally, some respon- ern personnel decisions in police departments dents mentioned forcefully that the superinten- constrain a chief’s ability to take action, though, dent could not just remove sector captains with in Lowell, Davis indicated that he would transfer impunity. They noted that his authority in mak- a sector captain who was not performing up to ing personnel decisions, similar to chiefs across Compstat’s standards to another position in the the country, was limited by civil service and union department. regulations: “A chief cannot hire, fire, or promote Commissioner Bratton was able to replace people at will” (Walker 1999, 368). Additionally, more than half of all the NYPD’s middle-level the provisions regarding discipline, such as nu- managers during the first year of Compstat. There merous avenues for appeal, make it extremely dif- was some turnover of older officers when Davis ficult for chiefs to dismiss bad officers, or even to took over, but it appeared that this was attribut- discipline officers for poor performance (Mas- able to the department’s commitment to commu- trofski 2002, 158). nity policing rather than a direct result of its imple- The sense of accountability varied through- mentation of Compstat. An officer recalled that out the Lowell Police Department. Sector captains, the incentive program to buy out officers was a who clearly recognized their responsibility for “sign of the times,” since the department was their beats, apparently felt more accountable than “old.” No one remembered Davis removing any other members of the department. The failure of of the original sector bosses from their positions, many department members to pose probing ques-

17 Compstat and Organizational Change in the Lowell Police Department tions at Compstat meetings in Davis’ absence also Geographic Organization of suggested the sense of accountability was less per- Operational Command vasive in Lowell than it was in Compstat sites like New Orleans whose department held district-level Compstat holds police managers to a high level of Compstat meetings to foster accountability accountability, but it does not do so without pro- through the ranks (Gurwitt 1998). viding these commanders with the authority to Davis plays an integral role as facilitator of carry out the Compstat mission. Mid-level man- questions and judge of sector captains’ account- agers are empowered in this model through the ability, and his centrality raises concerns regard- concept of geographic organization of operational ing Compstat’s potential longevity in the Lowell command. Operational command is focused on Police Department. It remains unclear whether the policing of specific territories, so primary de- Compstat is institutionalized at Lowell and cision-making responsibility is delegated to com- whether the department would be committed to manders with territorial responsibility (e.g., pre- continuing Compstat if Davis moved on. One re- cincts). That is, the organization gives a higher spondent addressed this concern directly by sug- priority to commanders who specialize in terri- gesting that Compstat relied so heavily upon Davis’ tory rather than function. Functionally special- distinct leadership style that it would disappear if ized units—such as patrol, detectives, school re- he left. Compstat also inspired some ambivalence source officers, and traffic—are placed under the among patrol officers, whose sentiments were command of the precinct commander, or arrange- equally divided along response categories, accord- ments are made to facilitate their responsiveness ing to our survey. to the commander’s needs. Considerable ambivalence also exists among In this section, we describe how operational top brass, middle management, and patrol officers command was organized in the LPD, and we as- about the use of Compstat as a forum for holding sess how much authority sector captains had over officers accountable for crime. There is some evi- department sections and units that had tradition- dence, however, that Compstat can be a powerful ally operated independently of them. Furthermore, means for achieving this goal at Lowell. Sector we highlight some of the specific challenges faced captains understood that being interrogated in by a department that was only partially decentral- front of one’s peers is a guaranteed method for ized, with some detectives assigned to the sectors ensuring some level of accountability. Sector cap- and others to a central division, and that was struc- tains expected Davis to ask numerous probing tured both geographically and temporally, with questions about crime and openly acknowledged individual sector captains working eight-hour their responsibility to come to Compstat fully pre- shifts and bearing responsibility for their specific pared with an array of questions. This sense of territory over a twenty-four-hour period. accountability did not filter down to the rank and Results from our national survey suggested file with the same potency, but patrol officers un- strong support for Compstat’s emphasis on geo- derstood that their sector captains carried a graphic organization of command. When we asked significant burden of responsibility. One lieuten- whether departments authorized middle manag- ant felt that Compstat held everybody account- ers to select problem-solving strategies for low- able since no officer was anxious for command level problems, 90 percent of Compstat-like de- staff to see a “crime spike” on his route or for fel- partments claimed to have given such authority low officers to leave a Compstat meeting and pass to district commanders, line supervisors, or spe- around the news that “So and so got his ass handed cialized unit commanders. This response clearly to him.” Humiliating as this is, hearing indirectly demonstrated that empowering middle managers about someone else’s misfortune is a less effective with decision-making autonomy is an important method of holding officers accountable than ex- feature of Compstat departments (Weisburd et al. posing them to the possibility of public censure 2001). on the same scale as their sector captains at Since Davis assumed control of the Lowell Compstat. Police Department, decision making has become

18 Compstat and Organizational Change in the Lowell Police Department

more decentralized, a change that we can quan- attended roll calls at department headquarters tify by considering which levels in the organiza- under the supervision of a single sergeant or lieu- tion bear geographical responsibility for spaces tenant indicated that on one level the LPD re- such as beats, sectors, or cities. Before the advent mained fairly centralized. of community policing and Compstat, the depart- ment was organized temporally. A captain oper- Decision making ated as a shift or watch commander and was re- sponsible for supervising officers on a single shift Despite this traditional structure, decision making for the entire city; once the shift was over, his or at Lowell is now significantly more participatory her responsibilities ceased. Under the old system, than it ever was under the traditional policing therefore, the lowest-ranking officer with twenty- model that preceded the widespread introduction four-hour responsibility for a geographic location of community policing in the 1980s. Police ex- held the command-level position of deputy chief ecutives in Lowell, as elsewhere, were responsible in charge of operations. Compstat, however, was for making decisions on department policy and grafted onto the department’s community-polic- procedure without much input from middle man- ing model and accordingly devolved this power agers. This began to change, however, when the down one level by giving sector captains twenty- LPD adopted community policing and moved to- four-hour responsibility for their sectors. So, for ward decentralization in 1996. Since then the the 60 percent of officers assigned to the sectors, department’s command staff of lieutenant through geographic organization of operational command deputy superintendent, as well as civilians, have had shifted from the very top of the organization played a more integral role in the decision-making (unitary command) down one level to the three process through informal private conversations, sector captains, effectively cutting the scope of weekly Operations Staff meetings, and annual stra- operational command to one third the previous tegic planning meetings, which also involve sepa- scope (if measured by number of sworn officers). rate conferences for patrol officers and adminis- trative personnel. Prior to this development, meetings between middle managers and the su- Organizational structure perintendent were infrequent, formal, and pro- In some respects the Lowell Police Department foundly hierarchical in character (Thacher 1998, under Compstat remained quite traditional. The 17). physical structure of the LPD continued to be cen- The decision-making autonomy of Lowell’s tralized with department headquarters dominat- sector captains in managing their own beats has ing operations. While some of its precinct substa- increased significantly under Compstat. Accord- tions were equipped with computer terminals ing to the department’s organizational chart, sec- (connected to the department’s mainframe) and tor captains reported to the deputy superinten- could function as command bases, others were dent of operations, but they really reported directly little more than storefront operations that were to the superintendent, leading one former sector closed at night. Detectives could not use the pre- captain to state that he felt very much “like the cincts to conduct investigations because they did captain of his own ship.” Another former sector not have interview rooms. The fact that all police captain, now in charge of a large enterprise, ex- officers, no matter what their sector assignment, pressed a similar sentiment by commenting that Davis allowed him to run his project how he wanted. Davis, who preferred his sector captains to rely upon their own judgment when making In some respects the decisions about their sectors, echoed these state- Lowell Police Department ments. For example, he explained that he did not under Compstat remained interfere when sector captains decided to allocate personnel equally within their individual sectors. quite traditional. Were he to make an executive decision in this in-

19 Compstat and Organizational Change in the Lowell Police Department

stance, he continued, it would appear to sector coordinating these plans with other beat officers. captains as an edict and foster resentment within Finally, a single sergeant, who had twenty-four- the department. hour responsibility for the team, was authorized Despite the increased autonomy of sector cap- to work any shift he chose so that he had the flex- tains, some evidence suggests that the authority ibility to meet with all the members of his team. to make decisions without supervisory approval The system in Saint Petersburg became unwieldy does not extend further down the organization. and collapsed after only three years, but it still In contrast to the autonomy of sector captains, serves as an example of how the level of decen- one of Davis’ responses to a sergeant’s question tralized decision making can vary dramatically during a Compstat meeting suggested that line between police agencies. At Lowell, sector cap- supervisors feel far less certain about their capac- tains definitely felt that they had the authority to ity to make decisions without consulting a supe- make decisions without getting approval from rior. When a sector sergeant mentioned to Davis above, but this level of clarity did not exist at lower that he should probably discuss his decision to levels in the organization. reassign back-up cars with his immediate super- Despite the extensive autonomy enjoyed by visor (a lieutenant), Davis used this as an oppor- sector captains, the exact nature of this autonomy tunity to reinforce the message that sergeants under Compstat is more complex than it might at generally had a great deal of decision-making au- first appear. Our national survey revealed that 90 tonomy within the department. Davis explained, percent of departments with Compstat-like pro- “Some people might not be happy with sergeants grams gave district commanders, line supervisors, taking initiative, and there could be some union or specialized unit commanders the authority to grumbles,” but he was willing to deal with this select problem-solving strategies for low-level and was disappointed that sergeants did not real- problems, but this percentage dropped to 70 per- ize “they could move people around” within their cent when it came to authorizing middle manag- sectors. He reiterated this point at another Comp- ers to choose how to deal with higher visibility stat meeting by stating that the official detail was problems (Weisburd et al. 2001). These results only a guide, and that sergeants had the authority suggested that middle managers under Compstat to put men “where they were needed.” Although did possess a significant amount of decision-mak- he noted that sergeants should still consult with ing autonomy, but Davis and his deputies might their sector captains, he was trying to encourage curtail their independence at times. When we ex- his street supervisors to recognize that they should amined the extent to which departments are will- exercise more initiative. ing to give middle managers greater responsibil- Davis has decentralized decision making down ity for determining beat boundaries or staffing to sector captains, but in contrast to Lowell some levels we found even less support for the concept police departments have made more vigorous at- of geographic organization of command. Only four tempts to take geographic organization all the way in ten departments that claimed to have imple- down to line officers (Parks et al. 1998). In Saint mented a Compstat-like model gave district com- Petersburg, Florida, for example, the police de- manders, line supervisors, or specialized unit partment assigned a small team of about five pa- commanders the authority to determine beat trol officers to one of the city’s forty-eight beats boundaries. Even fewer, only 18 percent, gave for their shifts. In addition, each beat was assigned these commanders the authority to determine rou- a community-policing officer who could work any tine staffing levels (Weisburd et al. 2001). time during the day as long as he worked forty In Lowell, the sector captains were largely re- hours per week. Freed from calls for service, he sponsible for redrawing their beat boundaries was primarily responsible for working on devel- during the department’s bid process.8 During our oping problem-solving strategies for the beat and research period, the sector captains used their

8. As part of their union contract, every eighteen months, patrol officers were allowed to request reassignment to any beat in the city; priority was then given to those officers with the greatest seniority, or most years on the force.

20 Compstat and Organizational Change in the Lowell Police Department

extensive familiarity with their sectors in redraw- actions, and was not averse to exerting his author- ing cruiser and walking routes. This recon- ity over them. On many matters, Davis invited the figuration involved the deputy superintendent of participation of sector captains in identifying prob- operations and the sector captains, with Davis lems and driving solutions. They engaged him on rarely interfering with his sector captains’ deci- these matters, and he had the final say. This pro- sions. This did not mean that he was unwilling to cess is more accurately characterized as participa- overrule any decision that he felt was inappropri- tory management than as a pure delegation of ate. Under pressure from both the community and Davis’ authority to the sector commanders. In con- the city to increase the total number of walking trast to mid-level managers, our observations sug- routes, Davis overturned a sector captain’s deci- gest, although we cannot state this definitively, that sion to cut one of his officer’s walking time in half street supervisors seemed to feel that their deci- (to four hours). Davis’ commitment to commu- sion-making autonomy was narrowly confined. nity policing and the problem of scarce resources In addition to giving district commanders the in the face of an increasing demand for beat officers authority to make decisions that once belonged contributed to the complexity of this issue, and to top management, Compstat also requires that good arguments were presented on both sides. functional differentiation, a conventional feature This scenario illustrated that sector captains of police organizations, be replaced with “geo- might have significant autonomy, but they always graphically based management” (Silverman 1999, operated within certain limits defined by Davis. 149), in which largely autonomous specialized One observer summed up Davis’ style as, “You can units that focus on specific crimes or vices are do anything you want, as long as I agree with you.” placed under the direct command of district com- Decisions over routine staffing levels also revealed manders or made responsive to their needs. While the persistence of the traditional hierarchical com- the accounts of respondents at Lowell suggested mand structure. A sector captain who required that specialized units would try to accommodate additional personnel was supposed to approach a sector captain’s request for assistance, it was clear Davis or his deputies and they would accommo- that the supervision of these units had not been date him/her if possible. transferred to sector captains because of Compstat. In sum, the decentralization of decision mak- Before and after Compstat, captains remained re- ing to mid-level managers in Lowell is complex. sponsible primarily for patrol officers, and the re- First, Compstat appeared to have consolidated assignment of detectives from headquarters to the decision-making power by creating the job of sec- sectors was part of the department’s community- tor captains, putting more of the department’s policing program rather than its Compstat initia- personnel under their command, and using that tive. Compstat, in fact, did not lead Lowell to pro- job as the focal point for twenty-four-hour respon- vide sector captains with additional specialists. sibility. Previously, the organization’s decision Lowell differed, in this respect, from the NYPD, making was done at levels both above and below which used Compstat to reconstitute some pre- that rank (by the command staff at headquarters cincts so that “all detectives, drug investigators, and by shift commanders). Second, sector captains and housing police who had previously reported possessed a great deal of autonomy. Davis encour- to their separate borough and headquarters supe- aged them to solve problems in their sectors, to riors” came under the direct control of the dis- decide how to assign sector personnel, and to co- trict commander (Silverman 1999, 149). At Lowell, operate with one another. This suggested that de- on the other hand, the detectives responsible for cision making was considerably more decentral- drug and serious crimes did not come under the ized at Lowell than at more traditional police direct command of the sector captains following agencies. However, this did not mean that sector the implementation of Compstat, and the Lowell captains possessed unrestricted authority. Our ob- Housing Authority continued operating as a sepa- servations suggested that Davis liked to be kept rate precinct. informed of his sector captains’ decisions, required Besides significantly increasing the input of that they provide supportable reasons for their sector captains in decision making and slightly

21 Compstat and Organizational Change in the Lowell Police Department

modifying the department’s division of labor, the In relation to the sector captains, he defined his geographic organization of operational command role as offering assistance and seeing “problems under Compstat contributed to a number of orga- develop . . . to be aware of what is going on.” nizational challenges faced by the department. The continuous flow of information among These included the difficulty of coordinating tasks sector captains, their detectives, and the Criminal between central and sector detectives, harmoniz- Investigation Section was the primary means of ing duties across sectors and shifts, and dealing coordinating investigations within the department, with the disproportionate burden of responsibil- a loosely defined coordination structure that has ity that falls on sector captains in an organization both benefits and drawbacks. Its most important that is structured both geographically and tem- advantages are to facilitate the flow of informa- porally. tion and potentially enhance productivity by re- ducing red tape and turf battles. In other words, it can help mitigate bureaucratic dysfunction, Coordination issues though it does have the drawback of depending Lowell differed considerably from more traditional heavily upon individuals taking the initiative for police departments in the decentralization of its contacting and coordinating with each other. Criminal Investigation Section. The process be- Under these circumstances, there is obviously gan in 1998, when the section was split into a some potential for organizational dissonance. handful of detectives, who were assigned to each Some members of Lowell’s command staff were sector, and a number of central detectives, who indeed concerned that coordination between sec- retained responsibility for the most serious crimes, tors was not clearly defined and might breed con- such as homicides and armed robberies. The task fusion over the responsibilities of individual sec- of coordinating investigations between sector and tor captains, particularly in the event of an ongoing central detectives that began at this time was on- multi-agency investigation. One sector captain, for going at Lowell, according to our observations. example, asked other members of the command The officer in charge of the section made a point staff to define his responsibilities in a homicide of using sector detectives on homicide divisions investigation, given that homicide investigations in order to provide them with much needed ex- in Massachusetts are the primary responsibility of perience. He also acknowledged the authority of the local district attorney and state police who a sector captain over his or her own detectives coordinate their activities with local detectives. and did not interfere in a sector investigation un- Officers responded thoughtfully by proposing that less the sector captain requested assistance. He his responsibilities were to ensure a controlled took this hands-off attitude toward sector captains response, such as “canvassing the neighborhood,” because he believed his role was “to be aware of or to “provide support by securing the crime scene what is going on” and to offer assistance when he and offering assistance.” Davis also responded by saw “problems develop.” This approach was evi- commenting that the sector captain was doing dently productive according to a number of at- exactly what was expected by empowering the tendees at Operations Staff meetings who re- community through crime prevention strategies. marked that sector detectives were much more The outcome of the investigation in question seasoned than they were two years ago and no was a success, but this scenario illustrates how longer had to rely upon the central detectives for the lack of a clear process fostered a need for clari- direction. The officer in charge of the Criminal fication and assurance. Speaking more generally, Investigation Section made a point of using sec- one high-ranking official suggested, “The depart- tor detectives on homicide investigations. This was ment needs to put a plan together as to what su- a way of providing them with much needed expe- pervisors should do.” The absence of clear opera- rience. Since the sector captain supervised his or tional guidelines makes it more likely that the her own detectives, the officer in charge of the autonomous actions of lower-level managers may section would not interfere in a sector investiga- circumvent the chain of command and create dis- tion unless the sector captain requested assistance. cord at higher levels of the organization.

22 Compstat and Organizational Change in the Lowell Police Department

Communication breakdowns are inevitable in any organization, and it is unclear to what extent It is the sector captains who are stricter rules and procedures would improve co- ordination. There was still some ambivalence in entirely responsible for crime in the the department over whether the Criminal Inves- city and must answer to the tigation Section should be entirely decentralized superintendent during Compstat. and whether the current hybrid system was work- ing effectively or, as one officer jeered, “It is work- ing in spite of itself.” Still, it is much easier for us to point to cases where coordination broke down that it was difficult, and somewhat unfair, to hold than where it succeeded, since failures are much someone continuously accountable while off duty. easier to detect. The issue of coordination between shifts was a topic of much discussion at Lowell and came up often at Operations Staff meetings. One observer Geographic versus temporal organization suggested that the issue was confusing enough that Conflict between the department’s geographical some officers, in particular the department’s ser- and temporal organization was the factor that most geants, might not actually know who their direct complicated coordination among department supervisor was. “If they [the sergeants] are con- members. Sector captains were directly account- fused,” he commented, “what do you think the able for whatever happened in their sectors over a patrol officers think?” He also described the twenty-four-hour period, even though they were department’s hierarchy metaphorically, in terms only on duty for eight hours during the day shift. of parents, or sector captains, and baby-sitters, or When the sector captains were off duty, the early shift commanders, who are responsible for enforc- or late night shift commander or lieutenant was ing the parents’ rules. This description might have responsible for carrying out each sector captain’s been a slight exaggeration, but it suggested the orders while simultaneously supervising all three existence of a feeling among some officers that sectors. This obviously raised difficult organiza- they had two bosses, a sector captain and a shift tional issues, such as the need for effective com- commander. One sector captain disagreed, stat- munication between the sector captains and shift ing that his personnel clearly knew that he was commanders and the problem of clearly establish- their boss. After all, as he said, “I approve their ing command responsibilities when the sector vacation time, their schedule, and their time off.” captain was off duty. He did, however, acknowledge that a sector cap- Once a problem strategy was recommended tain’s relationship with a shift commander was at Compstat, it was the sector captain’s responsi- complicated because the shift commander was bility to inform the shift commander (most often working “in a sense” for the sector captains but through a phone call or e-mail) about any tasks simultaneously exercised a great deal of decision- that needed to be implemented and supervised on making autonomy. the early and late night shifts. The sector captain The core of the coordination problem takes was also responsible for informing his or her lieu- us back to the challenges associated with account- tenant and sergeants of his actions so that the de- ability. It is the sector captains who are entirely cision was conveyed down the organizational hi- responsible for crime in the city and must answer erarchy. It was up to the shift commander to ensure to the superintendent during Compstat. This es- that these instructions were followed, but it ulti- sentially moves a major burden for managing the mately fell to the sector captain to “complete the police organization onto the shoulders of only loop” by making sure that the decision was actu- three people. The shift commanders, by contrast, ally implemented at the street level. Technically a “get a break” by being divorced from the whole sector captain had twenty-four-hour responsibil- Compstat process and “have little on their plates,” ity for everything that happened on his or her sec- when compared to sector captains, as one respon- tor, but members of the command staff recognized dent said. In a thoughtful paper entitled “An Or-

23 Compstat and Organizational Change in the Lowell Police Department

ganizational Plan for Patrol Operations,” one sec- were primarily responsible for controlling crime tor captain argued that the current system let ev- in the city. To support them in this endeavor they eryone “else off the hook” and placed a dispro- possessed significant decision-making autonomy portionate workload on the sector captains. In that was only, for all intents and purposes, directly support of this perspective, another respondent curtailed by the superintendent. In comparison, noted that some shift commanders did not worry Compstat had not resulted in a significant change about conveying command decisions to the lower to the overall structure of the department. Spe- ranks because “the heat falls on the sector com- cialist units were, in theory, made available to sec- manders.” The sector captains did have some com- tor captains, but they did not fall under their di- mand support, since each sector was assigned a rect supervision. Sector captains continued only lieutenant who served as executive officer. The to be responsible for patrol officers. Decentraliz- primary responsibility of the executive officer, ing decision making clearly encouraged middle however, was to convey the sector captain’s or- managers to take initiative and responsibility for ders down the chain of command rather than act policing specific territories, but it also created as a key decision maker. some difficult problems. An organization that is There are potential dangers that sector cap- managed both geographically and temporally will tains, who bear significant responsibilities, may have to establish clear and effective communica- burn out or become frustrated. The potential for tion channels that bridge sectors and specialized burnout is high since sector captains carry a heavy units and also extend across shifts and sectors. administrative burden, including weekly Opera- This helps ensure cooperation between top and tions Staff meetings with Davis, regular depart- middle-level managers and helps minimize break- ment meetings, weekly meetings with community downs in coordination and in the chain of com- groups, and the preparation of information for mand. There is also a danger that a few middle Compstat meetings. In addition to all this, they managers are burdened with too much responsi- have significant command responsibilities, includ- bility in comparison to their peers and subordi- ing managing and staffing large events, such as nates. Based on the impressions of our informants, Lowell’s Winter Fest, and dealing with personnel Compstat may encourage many to slide, while issues, such as abuse of sick leave by officers. In increasing the likelihood that sector commanders fulfilling these responsibilities, sector captains had may burn out or become frustrated with their con- to satisfy Davis that they were handling his pri- siderable commitments and the high level of ac- orities; they had to coordinate with each other; countability to which they are held. However, and they had to exercise initiative in identifying without any specific examples of burnout within and solving important problems. They also had the department, these findings remain speculative. to ensure that their subordinates fed useful infor- mation up the chain of command, carried out di- Organizational Flexibility rectives that came down the chain of command, and exercised the kind of problem-solving initia- Having the authority to respond to problems is a tive that Davis hoped to foster in his department. necessary but insufficient condition for middle As a result of these myriad responsibilities, sector managers to implement their decisions effectively. captains spent so much time going to meetings They must also have adequate resources. Compstat that there was hardly any time for them to imple- promises to respond rapidly and effectively to ment anything or to sit at their desks and talk to problems as they emerge, but it also requires flex- their staff, as one respondent observed. A sector ibility in an organization’s use of resources since captain described this difficult situation simply problems often arise in patterns that are difficult and poignantly by stating that, “Time is precious.” to predict. Unfortunately, police organizations are To sum up, the geographic operation of com- ordinarily designed in ways that restrict their ca- mand at Lowell under the Compstat model did pacity for flexibility. First, they are highly bureau- provide middle managers with considerable au- cratic and operate according to detailed rules and thority. To a large extent, just three individuals regulations that are often imposed or embraced

24 Compstat and Organizational Change in the Lowell Police Department

by police administrators themselves. This makes In this even more constrained environment, them much more successful at performing rou- how can police departments become more flex- tine tasks in response to highly predictable work ible? They can attempt to negotiate with collec- demands than at developing the capacity for cop- tive bargaining units for greater flexibility in mak- ing with unexpected problems. Second, adminis- ing job and shift assignments. They can budget trators’ ability to make their organizations flex- for overtime, which allows managers to work ible is constrained by other organizational forces, around the constraints of set work shifts. They mainly those concerned with employees’ rights can create special units that are not bound to work and management’s duty to negotiate changes in in a particular geographic area or work shift, serv- workplace routines. For example, the inflexible ing as “taxi squads” that are portable in when and regulations that labor union contracts impose tend where they work. They can also require or reward to discourage middle managers from changing cooperation and teamwork within and across or- work shifts and job assignments (Mastrofski ganizational units. Ultimately, the flexibility of the 2002). Third, the department is subject to pres- police organization is displayed when managers sures from a host of interest groups, including are able and willing to alter the allocation struc- politicians, business, the press, and neighborhood tures and routines in response to non-routine work associations, that want to determine how and demands. where officers spend their time. Sometimes these In our national survey, we measured organi- groups reinforce what management believes to be zational flexibility by the department’s response the department’s priorities, but they often con- to two questions: (1) To what extent had the de- strain what it may do to cope with unexpected partment reallocated resources such as reassign- problems and work demands. ment and overtime to the primary crime or disor- One of the ways that police organizations tra- der problem of the last twelve months; and (2) To ditionally coped with unpredictable fluctuations what extent did middle managers have general in work demand was to overstaff their patrol units authority to approve requests for flexible hours relative to the demand they expected on a given or to mobilize SWAT units to specific operations? work shift (Reiss 1992). That is, they built in slack Although these two items reflected the depart- time, during which officers engaged in preventive ment’s commitment to geographic organization of patrol, but their principal function was to be avail- command, they also focused directly on whether able for work that might arise, especially that there was flexibility in the allocation of depart- which might require an emergency response. Po- mental resources. Departments in the national lice departments have tried to build in enough survey that claimed to have implemented a Comp- slack to deal with significant, and unexpected, stat-like program did appear to have a great deal catastrophes since the 1960s, when former Chi- of organizational flexibility. Eighty-four percent of cago police chief and policing theorist O.W. Wil- these departments had reassigned patrol officers son advocated the systematic incorporation of this to new units, areas, or work shifts to address their flexible capacity into the calculation of staffing primary crime or disorder problem of the past year, needs (Wilson and McLaren 1972). and 80 percent of them had responded by using For many years this approach was widely prac- overtime to provide additional personnel (Weis- ticed by police departments and accepted by those burd et al. 2001). who approved their budgets, but the ethos of more What did Lowell do to develop flexibility? efficient, results-oriented policing has changed all Officers were sometimes reassigned to other units, that. Community policing, problem-oriented po- geographic areas, or work shifts, but by far the licing, and Compstat all call for officers to engage most common practice of reallocating resources in directed activities focused on preventing and outside of normal patterns was to do so on an ad solving problems, spending time with the com- hoc, informal basis that actually minimized dis- munity, and engaging in crime-reduction activi- ruptions to department routine. The most likely ties—all of which diminish the slack time avail- occurrence was for sector captains to direct a po- able through “preventive patrol.” lice officer or detective away from some of their

25 Compstat and Organizational Change in the Lowell Police Department

more routine activities during their shift and ask signed some specialists (some of the detectives), them to pay particular attention to a specific prob- a measure that increased within-sector flexibility lem area. In response to the ongoing problem of in responding to problems, but in general the de- motor vehicle crimes in one sector of the city, Davis partment did not rely on “taxi squads” to enhance requested that a central detective be assigned to flexibility. The department’s top leadership spoke work more closely with the CAU. In addition, if of a desire for greater teamwork but had to con- necessary, the department would pay officers over- tend with a number of disincentives and limita- time as part of a crime-reduction strategy. For ex- tions on the organization’s capacity to reward it. ample, in response to the brutal mugging of a Thus, while the LPD did undertake to increase woman in the highly visible downtown area, Davis flexibility in resource allocations, it was signifi- authorized overtime funds for additional patrol cantly limited in its capacity to do so and ham- over a two-week period. pered by traditional internal and external chal- A clear example of the reassignment of a spe- lenges that included lack of manpower, city cialist to a crime problem was the decision made politics, and rivalries within the organization. The at an Operations Staff Meeting to temporarily following sections discuss these in greater detail. move an assistant crime analyst from the CAU to the criminal division. In order to provide central Manpower detectives with greater access to gang information, the crime analyst was relocated and assigned spe- The paradox of resource flexibility is that the fewer cific gang-related cases. In relation to this move, resources the organization has to devote to flex- ible assignments, the more it needs that flexibil- ity. Despite the considerable increase in the num- The paradox of resource flexibility ber of sworn personnel since Davis took over, several respondents still felt that there was a real is that the fewer resources the need for more patrol officers in the police depart- organization has to devote to ment: “More troops on the street,” as one sector flexible assignments, the more it captain explained. This sentiment appeared to be widely shared, especially among the sector cap- needs that flexibility. tains, and was particularly strong in the West sec- tor where a disproportionate amount of gang and motor vehicle crime was concentrated. Silverman the crime analyst stated that there was a need for writes, “To its credit, Compstat is not tolerant someone in the detective division to “go over pat- when precinct commanders automatically respond terns” in regard to specific crimes. According to ‘we need more people’ to questions about below- the national survey, almost three quarters of de- par crime fighting” (Silverman 1999, 203). That partments responded that they allowed district said, several members of the LPD did seem to feel commanders, line supervisors, or specialized unit that such a response was justified when discuss- commanders to decide on flexible hour requests, ing how to tackle crime increases. One sergeant and 65 percent allowed them to mobilize SWAT believed that the department could use more pa- units. In line with this trend, Lowell’s sector cap- trol officers. He questioned the logic of keeping tains were responsible for making these kinds of patrol officers in administrative or specialty posi- decisions with little or no review by superior of- tions, such as computing or training, when they ficers. could be used more effectively on the street. An- Other methods to increase flexibility were less other line supervisor believed that his officers felt common. We did not observe the department seek- that they were understaffed but added, “The ad- ing concessions from labor. It did use overtime, ministration doesn’t want to hear this.” but the funds available were not always sufficient Given his budget constraints, Davis must re- to meet the demands, at least in the view of some spond to these general criticisms by exhorting his below top management. Sector captains were as- officers to work with what they have. At one

26 Compstat and Organizational Change in the Lowell Police Department

Compstat meeting, Davis expressed concern over deployment, as it did in early 2001 after a city a recent increase in motor vehicle crime in one councilor expressed concern about a spate of drug- area of the city, inspiring a sergeant to state that related crimes in one area of the city. Contending due to injuries, “We don’t have a full contingent,” that police should be more visible in the neigh- and “this really hurt us.” Davis pointedly reminded borhood, he told Davis and other city councilors the sergeant, “There is a propensity to want more that he wanted to see “more cops walking.” Davis people, but there are one hundred more cops out responded with some reservations, through he is there than five years ago.” He continued by tell- a staunch advocate of community policing and ing his audience that one way he was trying to foot patrol. Placing a higher priority on walking increase the number of officers available was to routes, he said, would reduce the number of avail- put a system in place that closely monitored of- able cruisers and provide some areas of the city ficer injuries and sick days. It was his hope that with less patrol. The only way to avoid this prob- this would increase available personnel by decreas- lem, he continued, would be for the department ing the number of inexcusable absences. Davis and to hire more officers or increase overtime. Since his deputies also tried to increase the number of the city council was unable or unwilling to sup- patrol officers by asking those in specialist posi- port either of these measures, the department’s tions to consider moving to patrol during the only alternative was to come up with a plan for department’s bid process. Davis, however, was not “Advanced Community Policing” that assigned keen on “pulling” specialists from their key posi- patrol officers to all the city’s neighborhoods be- tions because he felt, according to one respondent, fore filling cruiser route assignments. that they fulfilled a necessary purpose where they In prioritizing the deployment of patrol offic- were. ers, the police department was influenced by the Compstat may not be particularly sympathetic politically powerful editor of the local newspaper, to claims for more officers on the street, but by local business owners, and the Lowell Housing constantly bringing attention to crime problems Authority, as well as city councilors. They too were and the need for rapid and effective solutions, it keenly in favor of the department maintaining a also brings a sharper focus to the old bugbear of high level of police visibility in the downtown area the police organization—the lack of manpower. where their businesses were located. Davis re- Since middle managers are held most accountable sponded to this political pressure by requiring his for crime problems in their sectors, they are the sector captains to maintain foot patrols in the main proponents of an increase in personnel as downtown area on the day and early night shifts. one of the most effective strategies for reducing These patrols were in addition to six officers and crime. a lieutenant who were assigned to the Lowell Housing Authority precinct, according to an agree- ment between Davis and the city. In taking these City politics actions, Davis emphasized that he was ultimately The department’s budget must obviously be ap- responsible for his department and would not do proved by the city council, which is unlikely to anything that that he considered “inappropriate,” fund significant increases in police personnel dur- no matter who was putting pressure on him. He ing times of fiscal constraint. This reality of city ultimately decided to assign additional patrols to politics has affected Davis who began his tenure the downtown area and housing precinct because in office with “carte blanche” to run things as he he was just as committed to community policing wished, according to one officer. Five years later, as those outside the department. however, municipal funds were more limited, lead- This prioritization of patrol in certain areas ing city hall to propose that the police department placed considerable restrictions on the ability of had enough resources to effectively maintain or- sector captains to allocate resources and select tac- der, control crime, and provide services to Lowell. tics. On one occasion, a sector captain had much City politics can also reduce the department’s difficulty arranging for the regular deployment of organizational flexibility by influencing officer a decoy car to catch perpetrators of motor vehicle

27 Compstat and Organizational Change in the Lowell Police Department

crimes because of a lack of manpower. A respon- resources to another sector for assistance? In other dent commented that using a decoy car “ was not words, is it more important to demonstrate per- a practical application of manpower.” Even though formance in your own sector or to help out your Compstat encourages the flexible deployment of neighbor? resources, the department’s organizational capac- Clearly, this is an issue for top management to ity is limited by city politics. resolve, which it can do in two ways. One is for This daily reality became apparent during a the chief to get involved on a continuing and ad reassignment meeting, when a sector captain noted hoc basis, refereeing requests for additional re- that the department “was trying to do with doing sources between sector commanders. Of course, without” and suggested that it use Compstat more doing this undercuts the integrity of the principle scientifically to allocate manpower to high-crime of delegating operational control to sector com- areas. Another sector captain responded by point- manders. How can a general run his army when ing out that “there are politics” which make it the commander in chief may at any time take a impossible to base manpower entirely on Comp- significant portion of his troops from him to help stat. Given the limited supply of officers and over- out another general? The other alternative is for time pay, the goal of maintaining a highly visible the chief to create one or more special units that police presence competes with that of flexibly are not assigned to a specific sector, but that he deploying officers to problem areas. Sector cap- uses to fulfill special requests from sector com- tains and those they command are consequently manders, like a “special forces” unit in the mili- under frequent pressure to find alternative ways tary. This provides for greater stability among sec- of controlling crime that do not increase the ex- tor units, but in most departments the creation of isting strain on manpower. One sergeant, whose such a unit would require reductions in the allo- territory covered the downtown area, noted, for cation of resources to sector commands, thereby example, how difficult it was at times for him to reducing the organization’s commitment to geo- fill his route. Just that weekend, he complained, graphic organization of command. Davis occasion- he had been unable to fill his cruisers because four ally opted for the first strategy and not at all for officers had called in sick and he had to keep “two the second, but, for the most part, resource allo- walkers” in areas designated as high priority by cation between sectors remained stable, suggest- Davis and the city. ing that organizational flexibility was a constant challenge. In this sort of environment, flexibility depends Rivalry between sectors a great deal on informal cooperation between sec- One way of improving deployment flexibility tors, which depends, in turn, on the nature and would be to increase the organization’s ability to extent of rivalries among sectors. There is an ap- move officers from one sector to another on a tem- parent contradiction in Silverman’s praise for the porary basis. Such flexibility requires that sector NYPD’s use of Compstat to compare statistics commanders have a strong sense of shared respon- among precincts, such as coming up with top ten sibility for the overall good of the organization lists for those with the most arrests or the greatest and city, but an ethos like this clashes directly with drop in crime. Silverman argues that “rankings another Compstat principle—accountability. We spurred analysis of precinct activities, crime noted earlier that one of the paradoxes of account- trends, and results, and drove commanders to ability was that it limited Compstat’s ability to fa- perform better” (Silverman 1999, 101), but this cilitate innovative problem solving through brain- competition also had potential to hinder collabo- storming. Accountability similarly conflicted with ration. In Lowell, the department did not encour- organizational flexibility. How does one balance age competition directly, but Compstat did help the need of sector commanders to simultaneously foster rivalries since it compared crime statistics demonstrate that they are effectively addressing across sectors. crime in their assigned areas and promoting organ- Our observations suggested that competition izational flexibility by readily “donating” their between the sector captains over crime numbers

28 Compstat and Organizational Change in the Lowell Police Department

was relatively subdued. Most often they expressed turf. Any measure, therefore, that threatened to solidarity and sympathy when a crime spike oc- reduce their ability to reduce crime in their own curred in one of their sectors. A sensitivity to com- sector would encounter resistance. One sector petition was not, however, entirely absent. On one captain commented that if he was “down a ser- occasion a respondent said that he was slightly geant” on one of his shifts, he could ask a ser- uncomfortable that detective follow-ups and ar- geant from another sector “to keep an eye on rests had been displayed at a Compstat meeting— things.” He recognized, however, that, “It would this was not the norm. The respondent felt that be an exceptional sergeant who would be willing this might reflect negatively, and unjustly, on the to take on the added responsibility of stuff going hard work that was being carried out in those sec- on outside of his/her sector.” tors that were understaffed—in this case, due to Officers further down the ranks were also injuries. aware that Compstat did not encourage sectors to Most supervisors did not react like this respon- share manpower. One officer described this reluc- dent, though many line officers did have an acute tance to share “route men” between sectors as “a perception of Compstat as a tool for fostering com- turf thing” and explained that sector captains petition and a “big contest between captains to “don’t go into each others’ turf” because “they want have impressive stats,” as one patrol officer their own officers to ride around in their own sec- phrased it. Responses to tor.” A sergeant painted our line-officer survey a similar scenario in re- lent some support to this . . . many line officers . . . have sponse to a question description since 75 per- an acute perception of about the problem of cent of those surveyed understaffing and the “agreed” or “strongly Compstat as a tool for likelihood of borrowing agreed” that Compstat fostering competition and a officers from another sec- had made supervisors “big contest between captains tor to deal with a press- place too much empha- ing crime problem. “It sis on statistics. Sixty-five to have impressive stats” . . . could happen,” he con- percent, moreover, “dis- jectured, “but everyone agreed” or “strongly dis- in their own sector has agreed” with the statement, “Compstat has in- their own problems they are dealing with.” He creased teamwork between my unit and specialist went on to note that it had been much easier to units in the department.” This latter response, get the people you needed when the department though it did not specifically address teamwork was centralized than it was under the existing sec- among sectors, did suggest that Compstat was not tor system because sector captains think of the particularly successful at fostering the sharing of sectors “as their own little kingdoms and don’t resources within the organization. want to let people go.” On one occasion, for ex- One of the areas where this competition mani- ample, detectives carrying a heavy workload re- fested itself most clearly was in regard to man- ceived overtime pay when several injuries left the power. Since sector commanders were held ac- department understaffed. The department did not countable for their own individual sectors, they decide to provide additional detectives from other focused primarily on what happened in their par- sectors, however, suggesting that it is easier to ticular area of the city. Several current or former provide overtime than to shift resources across sector captains commented that they paid some sectors. attention to what was going on elsewhere, but the There appeared to be greater organizational reality was that during Compstat they were only flexibility in the LPD than in more traditional asked questions on their own sector. In addition, police organizations because Davis encouraged sector captains exercised ownership over their teamwork and pared down bureaucracy. This flex- sectors and were proud to take individual respon- ibility was significantly limited, nonetheless, by sibility for any problems that occurred on their both internal and external forces including: the

29 Compstat and Organizational Change in the Lowell Police Department

lack of available manpower, city politics, intra- agement or praise to his sector captains it tends organizational rivalries, and the demands on an to be for the achievements of individual sectors organization that was trying to pursue a number rather than for teamwork that occurred across of conflicting objectives simultaneously. On the sectors. one hand, the organization was trying to be fair to In sum, the department’s limited resources its employees, by giving them stable shift assign- were frequently requested by, and provided to, ments for example. On the other hand, it wanted sector captains for use within their own sectors to have enough flexibility in shift assignments to and were only rarely shared among sectors for the successfully address unforeseen community needs overall good of the department. For instance, the as they arose. It wanted to be fair to its citizens increasing number of motor vehicle crimes in the and provide them with equal protection against West sector led the department to assign two cen- crime, but it also wanted to be responsive to po- tral detectives to help solve the problem, but it litically powerful individuals and groups who want did not result in a request for, or assignment of, more resources for their priorities, regardless of additional personnel from the two other sectors. how it affects safety throughout the community. The absence of inter-sector cooperation that this Contemporary police organizations are not at lib- instance indicates arose from several factors. The erty to pursue a single objective for any length of first was that budgetary constraints did not leave time, so they resolve the conflicting demands on sectors with resources to share. The second was them by making compromises that will not pro- Compstat’s apparent failure to systematize team- voke a crisis or inspire negative publicity. work as part of accountability by making sector In Lowell, many respondents seemed to feel captains only accountable for their individual that “manpower” was an “issue,” despite the turfs. Last is the irony that Compstat’s emphasis significant increase in patrol officers over the past on organizational flexibility conflicts with the de- six years. Unlike some of the other Compstat de- velopment of geographic organization of opera- partments we visited, organizational flexibility was tional command. When a department strives to not clearly facilitated within the structure of the increase geographic organization of command by LPD. For example, there were no “taxi squads”— giving more resources to sector captains, it auto- officers that are part of a sector’s operational force matically increases the capacity of each sector cap- that do not have a specific geographic assignment tain to operate within his or her sector. At the same but are available to a sector captain to use where time, however, the department also decreases its he or she pleases. (One sector captain’s comment capacity to shift resources between sectors, a pro- that the amount of collaboration between sector cess that is easier when more personnel are moved detectives was due to “personalities” nicely en- from sector commands to special units that can capsulates the ad hoc quality of organizational be allocated from one sector to another as needed. flexibility in the LPD.) Even though a sector cap- tain would never interfere in the management of Data-Driven Problem another sector—“Everyone needs to take care of Identification and Assessment their own house”—there was some cooperation across sectors. One sector captain acknowledged Compstat relies heavily on crime statistics to re- resources were tight, but this did not dissuade structure information for managerial decision detectives from different sectors from collaborat- making by shifting the focus from highly selec- ing with one another. However, he observed that tive anecdotes of individual cases to the larger he only had a few detectives and would make it patterns and trends that might exist. The result, clear to any of them who wanted to help work as our national survey suggests, is for Compstat another sector’s case that their primary responsi- departments to have the capacity to manage and bility was to their own sector’s cases. This focus analyze data in more sophisticated ways. Over 90 on intra-sector responsibilities rather than inter- percent of these departments claimed that they sector collaborations is also evident during Op- conducted “crime trend identification and analy- erations Staff meetings. When Davis offers encour- sis,” and nearly as many claimed that they used

30 Compstat and Organizational Change in the Lowell Police Department

“database or statistical analysis software for crime staff; (3) the importance of ensuring timely and analysis” (Weisburd et al. 2001). Lowell was typi- accurate data; and (4) the heavy workload carried cal of these departments in having a Compstat pro- by civilian crime analysts in a small department. gram designed to provide information that would help managers decide on their priorities in handling crime or safety issues. As one Lowell officer put A history of Compstat—overcoming technical and learning obstacles it, “Information is always the name of the game.” Lowell designed its Compstat around account- It is impossible to understand how Compstat op- ability mechanisms so that systematically collected erated at Lowell without considering how the de- and analyzed data pulled from officer reports partment constructed its crime database and pro- would guide decision making. The doubling in duced crime statistics. At the time of our visit, the size of the department’s CAU and the rapid ability of the department to gather and process changes in the way crime data was collected and crime data was seriously hampered by the absence analyzed attest to the importance of crime data of a Records Management System (RMS), or elec- for instructing decision making. There were, how- tronic filing system, that would have allowed the ever, a number of challenges to data collection in CAU to download crime information directly from Lowell. We hope to serve two purposes by briefly a central database rather than enter it manually examining how Lowell attempted to overcome from individual police incident and arrest reports. these challenges and routinely gather, analyze, and The department contracted Lucent Technologies use crime data to understand and tackle problems. to remove the old Larimore RMS and install a new The first is to show police organizations how one one by December 1999, but a host of technical department overcame serious, technical obstacles. and financial difficulties prevented Lucent from The second is to provide a clearer understanding doing more than shutting down the old system of the sophisticated, “data-driven” processes that and installing a Computer-Aided-Dispatch System fuel Compstat. or CAD. Members of the department were frus- Our discussion also highlights a specific di- trated by Lucent’s failure to completely honor its lemma associated with using data to act strategi- contract, but they managed to implement a series cally: What data should be selected for Compstat of stopgap solutions that allowed Compstat to and what should be overlooked? In Lowell, when operate. members of the department became aware that The fact that any crime data could be system- their high-priority problems were being dropped atically gathered and analyzed in a timely fashion from Compstat, they exerted pressure by bending was testimony to the innovative thinking and un- Davis’ ear. Alternatively, they raised concerns dur- common work ethic of members of the CAU. From ing Compstat or made suggestions to the CAU to Compstat’s inception until the summer of 1998, a include or exclude a particular crime category. clerk in Records was responsible for photocopy- There was no forum for any kind of systematic ing every page of every individual police report, a discussion on what to include in Compstat, and process that could take up to eight hours for re- both these ad hoc approaches contributed to ports filed over the weekend, according to the Compstat’s constant state of flux. officer in charge of the CAU. Copies of these re- Finally, in addition to some of the technical ports were distributed to the three sectors, as well obstacles Lowell faced in implementing its Comp- as the detectives, juvenile section, and CAU. At stat program and the decisions surrounding what this time, the principal crime analyst inputted data to include, we discuss some other challenges crime data from the Larimore RMS into MS Excel associated with gathering, analyzing, and present- and used this to run Compstat. The crime ana- ing data. These include: (1) the burden put on lysts also exported CAD call data from the Lari- district commanders to learn crime analysis; (2) more system, “eyeballed” the resulting printout the problem of encouraging detectives and mem- to identify locations that appeared to generate a bers of the rank and file to participate in a pro- high volume of calls, and distributed hard copies gram that is heavily oriented toward command of this data to the sector captains and shift com-

31 Compstat and Organizational Change in the Lowell Police Department

manders who attended Compstat meetings. With- Records Division could use this to enter data from out any kind of mapping software, however, crime police officer reports. From January to March analysts had to rely on their own experience and 2000, the CAU extracted the data it needed to run observations for identifying hot spots. One crime Compstat from this MS Excel database and con- analyst conjectured that “We probably missed a tinued to supplement it with information from lot,” though he thought that most of the CAD call officer reports on PaperPort. Compstat, however, data concerned simple quality-of-life problems, depends upon up-to-date information, and the two such as complaints about barking dogs. In Janu- input clerks—out of the four clerks in Records— ary 1998, Davis recommended that the department were generally one-and-a-half to two months be- begin using crime mapping, according to one hind schedule with their data entry. The cause of crime analyst, and rely on officer reports rather this delay was that they were still entering old than CAD call data, which tended to mislead the police reports into the temporary database for gen- department by failing to tell the specific time or eral departmental access. In order to run Comp- location of an incident. The members of the CAU stat, the CAU needed to find out how far Records responded to Davis’ recommendations by input- had proceeded with its data entry, locate the re- ting information from officer incident and arrest ports that had not been entered, and then enter reports into MS Excel and then importing these data from these reports into its own CAU data- data into MapInfo in order to generate maps. CAU base. In addition to this cumbersome process and members also had to learn to use MapInfo on their the tardiness of the data, the MS Excel data were own since they were unfamiliar with mapping not in a Compstat-friendly format. One of the software. department’s crime analysts explained that there Another innovation took place around sum- were no standardized codes for crimes. For ex- mer 1998 when Records began using PaperPort ample, simple assault might be entered as “s.a.” to scan police officer reports into the department’s or “simp. ass.,” and street addresses were fre- server. PaperPort expedited the CAU’s access to quently misspelled. police reports, which had previously depended The tensions that these problems created be- upon the speed with which the Records clerks tween Records and the CAU came to a flashpoint could input data into the old RMS and provide in March 2000, when a sector captain complained the CAU with copies of officer reports. Now that the CAU was mapping data incorrectly. Mem- Records clerks could generally scan all officer re- bers of the CAU thought they were being blamed ports from the previous day onto the department’s unfairly for inputting errors, while those in server by 10 a.m. the following morning. The CAU Records felt they were being penalized for the limi- could then print them off, read them, and use the tations of their database. The officer in charge of information on crime they contained to supple- the CAU and members of the Management Infor- ment data that input clerks had entered into the mation Systems Division responded to the incipi- Larimore RMS. This process provided timely in- ent crisis by replacing the temporary MS Excel formation to Compstat participants, who had for- database with an MS Access database that provided merly used data that was often more than two users with drop-down menus containing standard- weeks old. ized data fields. This helped eliminate inputting Real administrative and technical problems in errors but was only intended as a short-term so- running Compstat began to materialize with the lution until Lucent installed the new RMS. shutting down of the Larimore RMS at the end of MS Access provided a more user-friendly data- 1999, leaving the LPD without any kind of cen- base, but the CAU still depended upon Records tralized database. In January 2000, the officer in for inputting crime data. By September 19, 2000, charge of the CAU and members of Lowell’s Man- the CAU decided that the whole process of coor- agement Information Systems Division tried to dinating data entry with Records was too difficult, overcome this problem by setting up a temporary particularly because the clerks in Records were MS Excel database as an indexing system on the entering only “the bare bones” of police reports, department’s server. The input clerks in the and the CAU demanded more detailed informa-

32 Compstat and Organizational Change in the Lowell Police Department

tion for Compstat. Since then, the CAU alone has ments that arose during the meetings, where Davis been responsible for inputting all the crime data often made suggestions. They would then use this that Compstat draws from police reports into its information to make additional changes and im- own MS Excel database. provements, such as choosing to display motor vehicle crimes at Compstat, on account of their prevalence. What kinds of data are used? At the time of our site visit, Lowell police officers Visibility of crimes at Compstat generated approximately sixty incident and arrest reports on any given day, and the department drew Since 1998, the crimes selected for Compstat have heavily on these reports for its crime data.9 The generally remained the same. The disappearance process of data collection and analysis began when and reappearance of domestic incidents—assaults, an officer filled out an incident or arrest report at disputes, arrests, and violations of restraining or- the end of his or her shift. The officer then gave ders—demonstrates the dynamism of the Comp- the report to the sergeant or commanding officer stat process, however, and shows that the high responsible for approving it before it was scanned visibility of “Compstat crimes,” as distinct from onto the department’s server and made accessible the low visibility of non-Compstat crimes, can to all personnel. A subsequent stage of the pro- influence how the department perceives and ad- cess took place every weekday morning when dresses crime. Domestic assaults were mapped members of the CAU accessed the PaperPort data- separately at Lowell’s Compstat until January 1999 base on the department’s server and printed off when people began to recognize that domestic the previous day’s reports for eight categories, in- violence incidents did not lend themselves to a cluding traffic accidents and crimes dealing with resolvable situation or “strategic technical re- assaults, burglaries, drug activity, breaking and sponse.”10 Although they were a priority within entering motor vehicles, motor vehicle damage, the department, the nature and prevalence of this motor vehicle theft, and other crimes. Informa- type of crime made it difficult to identify patterns. tion from these reports would then be presented Officers were bothered by seeing so many “dots” at Compstat meetings. on display month after month and found the Compstat at Lowell was in a constant state of department’s way of mapping domestic violence flux as feedback from participants in the program to be “a bit overwhelming and not instructive.” generated future change. This was shown by the The department responded by grouping domestic choice of crimes depicted during Compstat meet- assaults into the general category of “assaults,” ings. When Compstat first started, Davis, along which only distinguished simple from aggravated with members of his command staff, selected those assaults. By January 2001, Davis publicly ex- crimes they felt deserved the greatest focus. Part pressed concern about the change in Compstat’s of this selection process involved dispatching a mapping procedures, a sentiment that several com- community liaison into Lowell’s neighborhoods mand staff members had also voiced over the past to identify their concerns. Over time, sector bosses few months. The opening of Lowell’s new Domes- would visit the CAU and make requests that might tic Violence Center reinforced this concern a few result in changes. In addition, members of the months later, and the department reintroduced the CAU would review each Compstat and any com- mapping of domestics at Compstat meetings.

9. The CAU did use CAD call data for Compstat, though this was mainly for reporting drug activity and to a far lesser extent than police/ incident reports. In addition to its limited utility, the CAD system was a “locked” or self-contained system at Lowell. If crime analysts wanted to use CAD call data, they would have to print it out and then manually type it into MS Excel, putting additional time constraints on a system that was already seriously hampered by the absence of an RMS. The disadvantage of not using CAD call data, one captain pointed out, was that one address could be responsible for twenty-five calls a month, but it would not register at Compstat. 10. In contrast, the Compstat process is well suited to identifying and solving traffic problems. Between 1997 and 2000, the American Automobile Association (AAA) gave Lowell several awards acknowledging the department’s considerable success in reducing traffic accidents and problems.

33 Compstat and Organizational Change in the Lowell Police Department

Crimes displayed at Compstat maintained high categories. The types of data that the CAU visibility among the command staff, ensuring that punched in for each crime included the field re- they were subject to both ample attention and fre- port number, the date, day of the week, platoon, quent discussion. Some sector captains may have and sector, as well as specific information that may continued to focus on domestic violence incidents or may not have appeared on a map. CAU mem- following their removal from Compstat but found bers generally shared data inputting, though the it difficult to give them the same level of scrutiny. most senior crime analyst took personal respon- This is because the department may ignore crimes sibility for any errors by saying that any mistakes that are absent from Compstat. After all, Compstat were “on her.” Besides potential errors, the lack resembles police radar, and “If something is not of an RMS made it difficult for the CAU to com- on the radar, it is invisible,” as one officer put it. pare current Compstats with those conducted be- It was therefore unlikely that the department fore September 19, 2000, when the unit estab- would hold a sector captain to the same level of lished its own database. This difficulty was only accountability for a crime that was not on the underlined by requests for longer-term compari- Compstat screen. A general rule of thumb as stated sons from department members who did not seem by one respondent was, “If something is not shown aware of this restriction. In addition, the CAU’s at Compstat, no one cares about it . . . it means capacity for data analysis was seriously curtailed that you are not paying attention to it . . . you are because half a day was spent on data entry. Theo- not accountable for it.” It may be an overstate- retically, it would have been possible for Lowell’s ment to argue that non-Compstat crimes are “in- crime analysts to make comparisons of crime over visible” to the department, since those that are periods longer than twelve months, but this would reported are usually documented and often receive have been incredibly time consuming, requiring some police response. However, these observations analysts to print hard copies of reports compiled do raise an important issue: constantly holding before September 19, 2000 and to input the nec- district commanders accountable for the same essary data. Given that they were entering data crimes at Compstat may lead to potentially useful every day in order to remain current, this was a crime-related information being overlooked. tall order indeed. The CAU, however, was able to The decision to include a particular crime in make shorter-term comparisons, such as compar- Compstat reflected the pressures exerted by com- ing consecutive Compstat periods. mand staff members through conversations with During this same period, the CAU made the each other, Davis, or the CAU, and, at times, following preparations for Compstat meetings. On through discussions that arose during Compstat. the Monday or Tuesday before Thursday’s meet- For the most part, therefore, individual initiative ing, the CAU provided each sector captain with rather than a systematic decision-making process copies of all the police reports for the previous determined the types of crime data that were ulti- two weeks. It also used MS Excel to manufacture mately displayed at Compstat. Indeed, some non- spreadsheets that organized individual crimes by crime-related issues that Compstat excluded could city sector and gave details, such as the suspects still be very important to the department, accord- involved in an assault and battery case and the ing to one captain, because Compstat directed dis- case’s disposition, as well as recording variables, cussion to specific crimes rather than quality-of- such as day, time, and total number of crimes. By life issues in Lowell’s neighborhoods. late Tuesday night or early Wednesday morning, it gave sector captains copies of their sector crimes located on Compstat maps and MS Excel sum- How are the data collected, mary sheets that reported aggregates, including processed, and analyzed? details on the number of assaults and the platoon The absence of an RMS created problems, includ- or shift where they occurred (see Appendix IV). ing the lengthy three to four hours per day that it This procedure remained in effect until 2001, took the CAU to input data from officer reports when the department instituted a number of into its MS Excel database for the eight Compstat changes. As of February 2001, when each sector

34 Compstat and Organizational Change in the Lowell Police Department

captain began presenting every six, rather than Maps and patterns ______every two weeks, the CAU began giving sector The originators of New York’s Compstat program captains copies of their sector’s reports and the were very aware of the importance of crime map- additional spreadsheets a week ahead of time, since ping. Jack Maple observed, “It’s easy to lose sight there were now six rather than two weeks of re- of the power of mapping. Maps are superior to ports to review. As of September 2001, the CAU numbers or narratives as a means of communi- has been using SPSS, which is a more powerful cating to individuals at every level of an organiza- statistical package than MS Excel. It has also tion the immediate challenges in front of them. employed MapInfo, rather than PowerPoint, to Maps tell a story in a way numbers and narratives project slides at Compstat meetings because simply can’t” (Maple 2000, 105). The visual pre- MapInfo, unlike PowerPoint, allows presenters to sentation of crime maps was a central component zoom in on specific geographical areas. The Lowell of Lowell’s Compstat meetings. Crime maps were Police Department generates all the data displayed displayed for the presenting sector and for the at Compstat, but it encourages other agencies to entire city. This ensured that the department did share information. The department, for example, not focus solely on one sector’s crime figures— collaborated with Lowell’s chief probation officer, the sector that was being presented—since any of who exchanged information with an assistant the sector captains could be asked to account for crime analyst during daily phone calls and regu- crime in their beats. larly attended Compstat meetings where he par- Sector captains did not receive maps daily, only ticipated in the department’s discussion of crime during the week before Compstat. They used data. them, in combination with other available crime information, to identify crime patterns and prep How are the data used? themselves on incidents that were unfamiliar. Throughout the preparation process, they collabo- Sector captains ______rated closely with their executive officers and com- Everyone in the department had access to officer municated with the officers under their command reports on PaperPort, and each day the sector cap- to gather relevant information. Sometimes they tains and their executive officers printed off the also called or e-mailed the CAU in order to clarify previous day’s sector reports. Some sector captains any questions they had about data for their sec- read all the reports to familiarize themselves with tors. All sector captains noted that preparing for the crimes in their sector, while others relied more Compstat was time consuming, usually taking heavily upon their executive officers for this pur- several hours over a period of two days. Particu- pose. The executive officer read all the reports and, larly daunting were the maps. Since they displayed in consultation with the sector captains, assigned crime incidents over the long period of six weeks, detective follow-ups. The executive officer was one sector captain was inspired to playfully com- also primarily responsible for notifying his ser- ment, “Oh crap! I did not think that there are as geants of these joint decisions. The sergeants then many of these [crime incidents],” when he recol- relayed priorities to patrol officers. The sector cap- lected his feelings upon receiving them. tain also found out what was going on in the sec- Sector captains and their executive officers tor by communicating directly with line officers. were primarily responsible for identifying patterns The week before Compstat, the presenting sector captain would examine all of the sectors’ police reports in order to reacquaint himself/herself with what had been happening in the sector. The sec- All sector captains noted that tor captain might also skim reports from other preparing for Compstat was time sectors but paid far more attention to the crime consuming, usually taking several reports from his or her own sector. The primary hours over a period of two days. purpose of examining the reports, maps, and sum- mary data was to identify crime patterns.

35 Compstat and Organizational Change in the Lowell Police Department

and figuring out what was driving them. This was the “dead air” with details of specific incidents. because Lowell, unlike a larger department such Though some presenters related incidents that as the Richmond Police Department, did not have were serious, unusual, or entertaining, others used enough resources to provide sector commanders the occasion to engage in tedious monologues, with their own civilian crime analysts. District leading Davis to interrupt and ask them to move commanders in smaller departments may have to on. develop crime analysis skills on their own, a chal- lenge faced by Lowell’s sector captains, none of The Daily Bulletin ______whom had received any specific training on Comp- Since March 2001, the CAU has tried to make stat. It was clear from our discussions with them Compstat more inclusive by putting together sum- that some were more comfortable with this role mary handouts for distribution to all attendees at than others, even though none of them directly Compstat meetings. Those who did not attend addressed the challenges of becoming a crime ana- learned about some of the main issues raised at lyst. From time to time, sector captains did con- Thursday’s Compstat from the Daily Bulletin, the tact the CAU at headquarters, but this tended to department’s quotidian publication. At the time, be for additional information or clarification on a the Daily Bulletin was a two-page summary of particular report rather than an in-depth discus- crime-related material that was e-mailed to all sion of crime patterns. This may have been be- members of the department and distributed in cause members of the CAU were so busy that sec- hard copy at roll call. The importance of the Daily tor captains did not want to bother them with Bulletin should not be underestimated. The officer additional requests or because sector captains were in charge of the CAU praised it as “the most vital so busy that they did not have time to consult with information tool that we have in the department,” the CAU. It also may have been because sector an opinion supported by a patrol officer who de- captains felt they were better equipped to identify scribed it as a “very useful . . . little newspaper patterns, since they were able to combine Comp- kind of thing.” According to a crime analyst, the stat data with their detailed street knowledge. Al- CAU took most of the material that appeared in ternatively, it may even have been because sector the Daily Bulletin from police reports and regu- captains felt no need for assistance since crime larly included announcements made at roll call. analysis consisted of identifying clusters on maps Occasionally, the CAU added a synopsis of a crime rather than engaging in sophisticated statistical pattern or issued an officer safety warning, and analysis. Friday’s edition often included a short summary The importance of discovering patterns was of one or two crime incidents from each sector most apparent during Compstat meetings, where that had merited attention at the previous day’s sector captains felt compelled to acknowledge Compstat meeting. Patrol officers never made sug- even the absence of a pattern during their presen- gestions for the bulletin, the crime analyst noted, tations.11 The failure to make an explicit statement, although command staff and detectives occasion- such as “there are no patterns here,” could lead ally requested that the CAU include certain infor- another audience member to identify spurious mation. relationships between incident clusters. When this occurred, sector captains might respond that a Line officers ______cluster of dots on a map did not necessarily mean The dissemination of information through the that a pattern existed because the dots might in- Daily Bulletin raises questions about the availabil- dicate a set of incidents that occurred at different ity of crime data and Compstat to line officers. times and on different days. This tended to make Granted, Davis demanded that a couple of offic- meetings drag on because sector captains, who did ers from the presenting sector attend Compstat, not have crime patterns to discuss, tended to fill but members of the department’s command and

11. Sector captains understand the word “pattern” to mean an underlying factor that explains or ties together the occurrence of a number of crime events.

36 Compstat and Organizational Change in the Lowell Police Department

administrative staff dominated the meetings. Pa- at Compstat. Comments from other respondents trol officers and detectives played only a marginal suggested that information is usually exchanged role in both Compstat meetings and the overall when line officers have informal conversations Compstat process, and line officers, as a group, with captains, lieutenants, or sergeants. Follow- received no training in Compstat. When we dis- ing the chain of command, the sector captain or tributed surveys at roll call, feedback from patrol lieutenant may inform a sergeant to tell his or her officers suggested that they were, in fact, surprised patrol officers to “keep an eye on” a problem or to be asked their opinions on Compstat. They did “increase visibility” in a problem area identified not feel that their limited attendance at, and par- by Compstat. ticipation in, Compstat qualified them to answer Members of the command staff and CAU were any questions concerning its implementation and concerned that information generated at Compstat impact on the department. Compstat belonged to was not reaching the rank and file. One captain the “brass,” not to them. Fifty-six percent of pa- stated, “Quite honestly, I am not sure all the stuff trol officers had not attended a Compstat meet- at Compstat filters down to them.” To help rem- ing, and of the 41 percent who had attended, edy this situation, the CAU placed a large pin approximately half responded that they had “ob- map—measuring about six feet by four inches— served meeting only,” rather than participating or that marks crimes for a given Compstat period in answering questions. Before March 2001, when the roll call room. This has been in place since the CAU began formatting Compstat data in SPSS February 2001, and the hope is “it will bring and MapInfo, all personnel could easily access the policy-level decisions down to the patrol officer information collected for, and presented at, Comp- level.” Still it would appear to have limited utility. stat by using a series of folders and drop-down Before roll call, a handful of officers wandered over menus on the department’s server. Several respon- to the map and examined it for a couple of min- dents suggested, however, that line officers very utes, but, for the most part, it was ignored. This rarely accessed Compstat in this way due to a range did not necessarily mean that officers were unfa- of factors that included lack of knowledge that miliar with the major crime problems in their sec- Compstat could be accessed via a computer ter- tor, only that they were not involved in the Comp- minal, lack of interest, inadequate computer skills, stat process of collecting, analyzing, and discussing and the inability of computer terminals in the crime data. Our visits to other Compstat depart- department’s cruisers to access crime data on the ments suggested that they have made more of an department’s server. Compstat data on the depart- effort to engage detectives and lower-ranking per- ment’s server became even less accessible after the sonnel at the sector level in Compstat. In New switch to SPSS and MapInfo because most of the Orleans, for example, the department held sec- department’s computer terminals did not contain tor-level Compstats the day before the precinct either of these programs. As a result, even officers commander had to appear at the department who wished to access Compstat information after Compstat. These could last as long as two-and-a- March 2001 were unable to do so. half hours and involved detectives, patrol super- The majority of officers learn about Compstat visors, and some patrol officers. and crime-related issues through information pro- vided at roll call and in the Daily Bulletin. They Special presentations ______can also access crime maps by “asking someone.” At the request of Davis or a sector captain, the Rank-and-file officers learn what happened at CAU may assemble a short presentation for Compstat meetings from their fellow officers and Compstat or for organizations outside the depart- supervisors, though the latter source tends to be ment, such as the city council. The CAU, for ex- fairly haphazard. Forty percent reported that their ample, gave two ten-to-twenty minute presenta- supervisors discussed what happened at Compstat tions on motor vehicle crime in the city during either “every week” or “about once a month,” our research in Lowell. These presentations, which while almost the same proportion reported that often took up to eight hours of preparation, gen- their supervisors “never” discussed what happened erated more interest when Davis was present.

37 Compstat and Organizational Change in the Lowell Police Department

Detectives ______accuracy of data entry and mapping by the CAU. Apart from the officers in charge of the Criminal Moreover, sector maps sometimes varied slightly Investigation Section and the Special Investigation in the way they displayed addresses or dates, per- Section, only one or two other central or sector haps because different crime analysts prepared detectives were present at any given Compstat, a data for different sectors. The variations were mi- low rate of attendance that disappointed one re- nor and seldom perceptible to Compstat partici- spondent. Fortunately, the officers in charge of the pants. The officer in charge of the CAU worried, central detectives could tell them what happened however, that the lack of standardization across at Compstat. In addition, the Daily Bulletin pro- maps could do harm by undermining the cred- vided detectives with yet another means of gain- ibility of all the data. On occasion, an incident ing information about Compstat. might be mapped incorrectly, but this was rare and The CAU did receive requests from detectives, usually caught prior to the Compstat meeting but they were infrequent and generally originated Two major factors that contributed to the ac- from concern over a specific “hot” issue, such as a curacy of data were that the most senior crime series of robberies, motor vehicle crimes, or tag- analyst felt personally responsible for inputting ging incidents in a high-profile downtown area. errors and that close collaboration among mem- The requests generally required the crime analysts bers of the CAU encouraged the discovery of data to collect all the reports on a specific crime and errors. These safeguards notwithstanding, the identify characteristics, such as day of week and Lowell Police Department did not systematically location, which they had in common. On one oc- verify its data by drawing a sample of records and casion, for example, a detective asked an assistant double-checking them to see what percentage were crime analyst to access his gang database in order entered incorrectly. Since information drives de- to help identify suspects involved in a rash of cision making in Compstat departments, unreli- graffiti incidents. These types of requests, as one able data can be a major concern. For example, member of the CAU related, were “retrieval re- the accuracy of data on the frequency of juvenile quests,” in which detectives asked, “can you get crime in Lowell inspired heated debate at a com- me all the reports on . . .,” as opposed to calls for mand staff meeting between Davis and his cap- in-depth analyses of crime incidents. They re- tains after a lengthy report by the department’s quired the CAU to simply collect the information juvenile crime analyst showed that very few crimes and then present a short summary report to the against persons occurred in the local high school. detectives. Some captains proposed the analyst’s figures re- flected attempts by school officials to bolster the school’s safety record by underreporting crimes, Quality of the data—timeliness and accuracy and they worried about making policy decisions Effective decision making under Compstat de- on the basis of inaccurate information. Davis, on mands data that are timely and accurate. Police the other hand, was clearly apprehensive that com- reports are generally available the next day on ments like these from his command staff would PaperPort, and updated crime statistics and maps undermine the department’s general commitment for Compstat are available from the CAU every to using data to drive policy. Such an approach, two weeks. Before the implementation of Comp- he feared, might lead officers to reject valid infor- stat, Davis related, they only looked at the Uni- mation that could help them tackle crime. form Crime Reports, which provided statistics that It is crucial for officers to record information could be as much as a year old. “I believe in the correctly because almost all the information gen- utility of timely data,” he added, “so Compstat erated at Compstat comes from their reports. At seems like a smart approach.” Lowell, the shift supervisor was responsible for The accuracy of the Compstat data was pri- ensuring that officers wrote accurate reports, un- marily affected by officer reporting errors and in- like New York where precinct commanders an- putting errors by the CAU, compounded by the swered for any errors. Davis frequently urged com- absence of a formal procedure for checking the manding officers at Compstat meetings to make

38 Compstat and Organizational Change in the Lowell Police Department

sure that officers’ reports were accurate and suffi- Rather than being careless, patrol officers may ciently detailed. Despite his efforts, one respon- simply not have known how important it was to dent still felt that officers’ reports had not im- write complete reports, a problem exacerbated by proved very much over the last three years. He the department’s ostensible failure to give patrol noted that he had never heard of a report being officers enough positive reinforcement for accu- sent back for revision to either a shift commander rate reporting. Patrol officers did express consid- or a patrol supervisor because it was so poor. He erable ambivalence about the amount of praise added, “Somewhere the message is not getting they received for writing accurate reports, when down [to the rank and file] crystal clear.” Though responding to a question in our survey that asked there were some problems with officers’ reports, them to rank ten activities for which they might one of the crime analysts estimated that only a receive praise from their superiors. Responses to small fraction of about 2 percent were illegible or the question were divided between 54 percent of incomplete enough to be considered terrible. Oc- officers who ranked writing accurate reports one casionally the CAU would put up a caption on a to five and 46 percent who ranked it from six to Compstat slide that reminded the audience about ten, with one being the activity that would receive the problem of substandard officer reports, for the least praise from superiors and ten the most.12 example “only 9 of the [stolen motor vehicle re- These closely divided responses showed clearly ports] listed damages” (italics in the original). In that the importance of accurate report writing, as terms of the bigger picture, a respondent noted indicated by perceptions of praise from superiors, that poor reporting hindered attempts to identify seemed unclear to patrol officers. It seems likely crime patterns. For example, an incident could that the lack of contact between the CAU and line be gang related, but this was left unrecorded. By officers also hindered the impact of this message. relating this to the lack of direct communication It might facilitate accurate reporting if CAU mem- between the CAU and patrol officers, the respon- bers were able to express their needs directly to dent was drawing attention to how much less line line officers and explain to them that high quality officers participated in Compstat than sector cap- reports are indispensable to the Compstat process. tains did. Officers need to understand that all decisions that In sum, CAU members had to depend upon emerge from Compstat rely on the quality of in- their experience and familiarity with the data, their formation they provide to the CAU. If the data are collaboration with one another, and their interac- unreliable, decision making will be misguided. tions with supervisors and managers in order to ensure reliability. In the absence of an RMS, CAU members had to input a great deal of data and had The Crime Analysis Unit (CAU)—autonomy, time constraints, and frustrations no time to systematically check for the kinds of errors associated with manually entering data or The CAU’s authority to prepare for Compstat and transferring data from one format to another. The run meetings with a high degree of autonomy CAU also joined Davis in expressing concern mirrors the standard division in police depart- about the quality of officer reports. As far as we ments between civilians who provide technical could tell, Davis conveyed this priority to sector support and sworn personnel who focus on field captains and shift commanders by urging them to operations (Crank 1989). Sometimes the crime exhort their patrol officers to focus more intently analysts received requests from the brass, but they on writing accurate reports. were left alone for the most part. Formal meet-

12. If we sort the activities according to the percentage of officers who ranked them one through three in terms of how much praise they might receive we get the following ranking: (1) recovering a weapon (62 percent); (2) making a large drug bust (53 percent); (3) reducing overall crime in their beat (43 percent); (4) writing accurate reports (33 percent); (5) identifying a crime pattern (32 percent); (6) working with community members to solve local crime or disorder problems (26 percent); (7) assisting crime victims in obtaining services and cooperating in prosecutions (22 percent); (8) rapid response to 911 calls (21 percent); (9) making misdemeanor arrests upon sight of persons known to the police (17 percent); providing legitimate activities for potential offenders (13 percent).

39 Compstat and Organizational Change in the Lowell Police Department

ings between the CAU and the department’s com- mand staff were also very rare, occurring only once Given the importance of data- in the eight months we spent at Lowell. Some driven policing in Compstat, it members of the department were disappointed by may seem strange that crime this lack of executive oversight in a process so analysts were so undervalued in central to the organization’s operation and felt that command should invest more interest and energy the Lowell Police Department. in constantly evaluating and improving Compstat. This hands-off attitude also troubled the CAU when command asked it to run an analysis on a of the CAU acknowledged that members of the particular problem, such as motor vehicle crimes CAU were “undervalued and underpaid,” while in city garages, but failed to explain the reasons another recommended that they and other civil- for its request or provide any substantive feedback ian staff be integrated more fully into the depart- on the CAU’s subsequent report. ment by becoming “a part of police business” and In addition to operating with little guidance “a formal part of the organization.” from management, the CAU was under consider- Given the importance of data-driven policing able time constraints. Preparing data for Compstat in Compstat, it may seem strange that crime ana- was labor intensive and placed a heavy burden on lysts were so undervalued in the Lowell Police a couple of crime analysts and the unit’s full-time Department. This might have been because Comp- interns. Assisting detectives, preparing special stat was a relatively new program that operated presentations, and producing the Daily Bulletin on a shoestring budget from soft money grants. added to the CAU’s duties. In addition, some mem- One of the indicators of Compstat’s long-term bers of the CAU were responsible for time con- importance in Lowell is, in fact, what happens to suming and onerous tasks that were unrelated to crime analysts’ positions and the resolution of is- Compstat. One crime analyst, for example, had sues surrounding the soft money that funds them. to take a week out of every month to count police Davis and members of his administrative staff ex- reports by hand for the Uniform Crime Reports. pressed concern about the sustainability of the Another analyst had a monthly commitment to CAU, given that none of its members beside the tracking and recording information on recently officer in charge are on the payroll. Davis com- released inmates as part of a national Prison Re- mented that the department was experiencing a entry Initiative program. bit of a “backlash” to the generous funding it had Besides this heavy workload, members of the received in the previous five years. He also felt CAU also had to contend with other disadvan- that it was clearly necessary to make Compstat a tages. CAU members were unable to fully utilize permanent part of the budget, even though he their research and technical skills in the absence knew there was an “issue surrounding the inter- of an RMS and were frustrated at having to spend nal capacity of Compstat.” many tedious hours in cramped quarters simply William Walsh has recently argued that small inputting rather than analyzing data. They also departments may be unable to implement Comp- had to perform menial tasks, such as locating miss- stat as a result of the costs of purchasing appro- ing reports or retrieving reports from Records, at priate software and more sophisticated computer the request of officers who were unfamiliar with technology for collecting and analyzing accurate the integral role of crime analysts in the Compstat and timely data. (Walsh 2001, 357). This may be process. Additionally, crime analysts were funded true in some cases, but the implementation of by soft money from federal or state grants so they Compstat at Lowell demonstrates how even a earned a relatively low wage, did not have com- small department with limited funds can use ex- prehensive health benefits, had little job security, isting resources in personnel and software to get a and generally lacked employment privileges, such Compstat-like program off the ground. A depart- as regular pay raises, that were enjoyed by mem- ment can also be inventive and include modest bers of the police union. One respondent outside requests for Compstat funds while applying for

40 Compstat and Organizational Change in the Lowell Police Department

external grants earmarked for more popular pro- embarrassed by a significant data error. Finally, in grams, as Lowell did when requesting funds for spite of general recognition that crime analysts community policing. The department can then use play an invaluable role in data-driven policing, these funds to develop Compstat by hiring trained smaller organizations might be unable to secure crime analysts, installing powerful computer serv- the permanent and adequate funding necessary ers and terminals, and purchasing sophisticated to attract and keep highly skilled civilian crime mapping software. analysts. For all these reasons, it seems impera- Despite Lowell’s considerable success in imple- tive that police departments should at least hold menting Compstat, obstacles remain. The contin- periodic meetings with members of the CAU and ued gathering, processing, and analysis of data, representatives from all police ranks to systemati- that provides the foundation for the Compstat cally monitor and evaluate the entire Compstat process, presents additional challenges, especially process for weaknesses in order to make improve- to small police departments that are under- ments. resourced. Compstat requires district command- ers to become skilled data analysts, but the com- Innovative Problem-Solving Tactics plexity of advanced statistics makes it unlikely that they can achieve this without additional training. Police have long collected data and compiled sta- Our observations at Lowell, as well as the other tistics, but they rarely used these data to make departments we visited, suggested that crime important decisions about how to solve problems analysis there focuses on descriptive statistics— (Mastrofski and Wadman 1991). Despite its cen- such as percentages, frequencies, and means—not tral purpose of reducing crime, Compstat’s rela- the kind of rigorous, analytic, and “evidence- tionship to innovative problem-solving tactics is based” research ideal advocated by Lawrence probably the least developed element in the exist- Sherman (1998). Without the resources to assign ing Compstat literature. Middle managers are ex- civilian crime analysts to individual precincts or pected to select responses to crime problems that sectors, it is unlikely that crime analysis within offer the best prospects of success, not because smaller police departments will become more so- they are “what we have always done,” but because phisticated. a careful consideration of a number of alternatives It is more likely that CAUs in smaller depart- showed them the most likely to be effective. In- ments will be overwhelmed with commitments novation and experimentation are encouraged: use and feel underappreciated, given their significant of “best knowledge” about practices is expected. workloads and lack of stable funding. The unavail- In this context, police are expected to look be- ability of adequate resources might also increase yond their own experiences by drawing upon the likelihood that the reliability of the data will knowledge gained in other departments and in- be compromised. We found similar problems to novations in theory and research about crime con- these in one form or another in several of our six- trol and prevention. Silverman argues, “This in- teen short site visits. Without more time or man- novative process [Compstat] radiates throughout power, it is difficult for CAUs to conduct system- the NYPD as the energizer of strikingly creative atic checks for reporting and inputting errors. The decision making at headquarters and in the field” peripheral role played by line officers in the (Silverman 1999, 123–4).13 Compstat process may further undermine the in- In the national survey, this did not seem to be tegrity of the data. Since rank- and-file officers are the case very often in Compstat programs across unaware of the importance of carefully written and the country. When departments were asked how unbiased police reports, they may contribute to they decided upon a problem-solving strategy to the inaccuracy of the data—though we did not address “the one crime/disorder problem that used observe an instance at Lowell where anyone was more of the department’s efforts than any other

13. Although in New York, units were also pushed to implement the same strategies that had guaranteed success in the past (Silverman 1999, 113).

41 Compstat and Organizational Change in the Lowell Police Department

problem in the last twelve months,” they were sponses. The department tended, however, to re- most likely to tell us that they relied on the depart- sort to traditional law enforcement strategies. ment’s previous success with an approach. Sixty- Thus, the problem solving stimulated by Lowell’s seven percent of departments stated that this was Compstat generally replicated the classic street “very important” followed by 39 percent who cop’s ethos that it is more important to act deci- stated, “research evidence suggested that this was sively—even with an incorrect diagnosis and hast- the best approach” (Weisburd et al. 2001). In other ily considered solution—than it is to ponder the words, departments with Compstat still tended to evidence and carefully weigh options before taking rely on their own experience with traditional law action. enforcement approaches, such as saturating an area with police and making arrests. In many respects, Lowell mirrored this na- Problem solving and brainstorming during Compstat meetings tional trend. Compstat facilitated the gathering and effective dissemination of information to a Attendees at Compstat felt that there was a great much greater extent than it contributed to the deal of “reporting and brainstorming,” as one sec- implementation of innovative problem-solving tor captain put it. Our observations suggested, tactics, such as mobilizing other public/private however, that this was an overstatement. The pre- agencies. On one occasion, a member of the com- sentation of crime data from each sector did fa- mand staff recommended the innovative strat- cilitate some sharing of information, and a little egy—in comparison to merely increasing patrol— prompting from Davis encouraged participants to of contacting insurance companies and asking offer some insights on how sector captains might them to notify owners of high-risk vehicles for tackle important problems. In other words, audi- theft. During Compstat, he suggested that some ence members did exchange factual knowledge of the insurance companies might have been will- and provide suggestions drawn from personal ex- ing to provide anti-theft devices, such as “the club” periences, but they were far less likely to engage to their customers. Unfortunately, the idea fo- in in-depth discussion of the advantages and dis- mented little discussion and never came to frui- advantages of a broad range of problem-solving tion. We noted earlier that one of the significant alternatives. challenges to brainstorming, an important element The pressure that internal accountability of the innovative problem-solving process, is placed upon sector captains limited the utility of Compstat’s heavy emphasis on holding district the suggestions made at Compstat. In response to commanders accountable for identifying and re- this pressure, sector captains felt compelled to sponding to problems quickly. This pressure hin- have a crime strategy already in place before Comp- ders careful sifting of crime patterns, deliberation stat meetings. In addition, the onus on individual about them, and judicious review and discussion sector captains to come up with quick and effec- of the benefits and drawbacks of various ap- tive solutions to crime problems—solutions that proaches—the very kind of analysis that Compstat were likely to involve the sector captain investing claims to encourage. A sector captain’s failure to precious resources in their successful implemen- have a problem-solving strategy already in place tation—increased the likelihood that potentially by the time of the Compstat meeting is an addi- helpful criticism made during Compstat would be tional limitation to the innovative problem-solv- regarded as threatening or unhelpful. This likeli- ing process. The absence of a strategy can be seen hood increased when the source was perceived as as a sign of incompetence, and questions and state- being ill informed or inexperienced. At this point ments made in Compstat can be perceived as at- the sector captain had already invested his or her tempts to undermine a sector captain’s authority. reputation, as well as the sector’s resources in the These stumbling blocks did not prevent Davis chosen solution. The Compstat meeting might and his sector captains from encouraging prob- have revealed a problem-solving approach not lem solving, and there were some instances in previously considered in the sector, but at this which Lowell police implemented innovative re- point the sector and its commander had already

42 Compstat and Organizational Change in the Lowell Police Department

invested in the chosen solution. Radical changes comments or suggestions from audience members would be unnecessary, since the strategy had al- at Compstat as a challenge to their authority and ready solved the crime problem, or unwelcome, expertise, superfluous, ill informed, or merely as given the investment of resources and reputation. an attempt to try and impress Davis. Audience Sector captains at Lowell informed us that they questions very noticeably interrupted the rhythm could start to identify patterns and problems well of Compstat presentations whose familiar routine in advance, since they were familiar with daily seemed almost scripted at times. One respondent officer reports. Their knowledge of the geography observed that if Compstat was really intended to of their sectors also enabled them to map crimes solve problems then “it would not be in its present in their heads while they were occurring, infor- form.” Eliciting problem-solving responses from mation that they supplemented through discus- administrators or civilians, who lacked experience sions with their staff and drives around their in- in the complexities of police work, was unhelp- dividual sectors. Given the importance that ful, since their comments were sometimes per- up-to-date crime maps play in Compstat, it seemed ceived as “slings and arrows” designed to expose curious that sector captains felt no need for daily the sector captain to criticism. One respondent or weekly maps in order to identify crime prob- remarked trenchantly that Compstat, despite all lems in their beats. An explanation for this dis- the hype, still promoted a traditional “reactive” crepancy may lie in one sector captain’s approval response to crime. He supported his statement by of the department’s decision to have each sector noting, “People have never comprehended tech- present every six weeks, rather than every two, nology to be proactive.” Since the information and to come to each meeting with more data. If displayed at Compstat was as much as six weeks crime patterns become more detectable over old, Compstat participants were clearly reacting longer periods of time, as this sector captain be- to incidents rather than implementing preventive lieved, daily maps would probably not be very measures. He asked, “Why are we still seeing so helpful in identifying long-term patterns. many dots?” meeting after meeting if Compstat The ability of sector captains to discern pat- was so effective. terns without access to daily maps and have a crime strategy in place before Compstat would Follow-up seem to undermine the value of maps as a crime- fighting tool. A cynical interpretation of this According to Silverman one of the key features of development would be that the maps were prima- Compstat is the “particular emphasis on follow- rily for show or entertainment, but a more bal- up and constant monitoring” (Silverman 1999, anced assessment would be that the maps con- 193). This is essential for ensuring that tactical tributed usefully in a number of ways. They responses are being implemented and desired re- encouraged and reminded sector captains to think sults are actually being achieved (Walsh 2001, geographically and “see” patterns. They informed 355). At Lowell, an administrator took notes dur- department members who were less familiar with ing Compstat, but these were cursory—maybe a a sector’s geography about its crime topography; line or two on each Compstat slide—and were for and their display at Compstat identified patterns Davis’ personal use rather than for command staff that sector captains may have missed before meet- review. Immediately following Compstat, there ings. In short, our observations at Lowell suggest was no meeting to discuss “reinforcement and fol- that Compstat maps may be helpful but not indis- low-up.” One respondent expressed disappoint- pensable to the problem-solving process—a ment that there was not a more effective “feed- finding that challenges the highly touted utility back” loop in the Compstat program. He noted of crime mapping claimed by Compstat advocates. that when a problem came up at one meeting, there The Compstat format raised other problem- was no guarantee that it would be addressed at a atic issues. Since sector captains were held ac- subsequent meeting. He commented that prob- countable for having effective crime control strat- lems were “lost” because of this. His suggestion egies already in place, they regarded some was to encourage an “intensive and formal exami-

43 Compstat and Organizational Change in the Lowell Police Department

nation of the data.” An important crime problem could be addressed six months after its identifica- Follow-up at Lowell may not be tion. If the problem had not abated, according to “relentless,” but sector captains must crime analysis, the comparison could be used to provide some kind of answer to hold the sector captain accountable: “You have had six months to work on this problem, why is there Davis’ inquiries at Compstat . . . no change?” Another regular Compstat attendee noted that the hot topic at one Compstat was usu- ally forgotten two weeks later. Pointing vigorously enough prepared to have something intelligent to to a handout showing a 132.14 percent increase report—is enough to succeed. In contrast, getting in motor vehicle breaking and enterings for the a good grade for “doing well” would imply a dif- last Compstat period, he asked rhetorically why ferent measure of success—one that ranked per- it had not been addressed at the following Comp- formances according to much more selective stat: “Where did it go?” He added cynically: “Can criteria, in this case, the careful selection and you tell me that anything is honestly getting done evaluation of an effective problem-solving strat- because of Compstat?” Our observations sup- egy; and (2) They suggest that Compstat is far ported the assertion—one made by other respon- removed from the “real” world of police work. By dents—that there was a discontinuous relation- being a student taking a test or turning in a paper, ship between Compstat meetings. Invariably, little they seem to imply a disjuncture between what reference was made to discussions that took place goes on in Compstat, or the classroom, and what at a prior Compstat, as if each meeting was her- is happening out there on the streets. metically sealed in time and space. An unfortunate side effect of this lack of fol- Follow-up at Lowell may not be “relentless,” low-up is that innovative ideas can go unrewarded, but sector captains must provide some kind of since there is no collective memory or positive answer to Davis’ inquiries at Compstat—for ex- reinforcement encouraging their implementation. ample, “We have increased patrol in that area.” The vulnerability of new ideas becomes clearer We observed that Compstat was similar to a gun- by returning to the insurance example cited ear- fight in the Old West. What appeared to matter at lier. At one Compstat, as we have seen, partici- Compstat was that sector commanders were pants bounced numerous ideas around the table “quick-to-the-draw” and able to return fire with a concerning how to reduce the problem of stolen suitable response. The speed and decisiveness of motor vehicles in the city, including encouraging the response was more important than a clear di- insurance companies to contact owners of certain agnosis of why a particular course of action was high-risk vehicles. Since there was no concerted chosen and why it was considered the most effec- follow-up, this potentially effective approach was tive approach. One respondent used another anal- soon forgotten. Fortunately, Davis and other de- ogy to illustrate that it was more important to ap- partment members recognized the problem, which pear knowledgeable about a problem than to increased the likelihood it would be remedied. In demonstrate the development of strategies to ame- general, however, as Davis admitted, “We do a lot liorate it. He said that presenting at Compstat was of analysis here, but we don’t track a lot of plans similar “to dealing with a term paper: You do it, for a response.” turn it in, and then get ready for the next one.” Another respondent compared it to taking a test: Traditional police responses “Compstat is in some ways like being a student. As a student you study just so that you can pass a Most of the problem-solving approaches identi- test; just like with Compstat you just prepare in fied at Compstat relied on traditional police strat- order to pass it.” These analogies point to a couple egies that had been used before—in particular, of interesting implications: (1) They draw atten- asking patrol officers to identify suspects and tion to a pass/fail standard where turning some- “keep an eye on things,” area saturation, stepping thing in—or, in relation to Compstat, being well up traffic enforcement, “knock-and-talks,” and

44 Compstat and Organizational Change in the Lowell Police Department

increasing arrests. For example, in response to two mean that it will be more effective than a tradi- unrelated incidents—an increase in prostitution tional law enforcement response. Nevertheless, it in a particular location and the constant use of a does appear that Compstat’s claim to encourage specific pay phone by suspected drug dealers— responses that rely upon cutting-edge research and the sector captain put “extra cruisers” in the area. experimentation could easily be a serious over- When one Compstat showed that the street-side statement. windows of several parked cars had been smashed, During our research at Lowell, the department Davis asked his deputy: “What kinds of things continued its attempts to reduce the long-term have we done in the past?” The deputy suggested problem of motor vehicle crime—theft, damage, that they clamp down on motor vehicle violations: and breaking and enterings. This was a citywide “You know, chief, sometimes you just get lucky. concern, but it was a major problem in the West You catch a kid and they just talk. We need to get sector. The high visibility of these types of crimes people in to talk to them.” in the department and city put pressure on sector captains to remain “on top of the problem.” Over the eight-month period we visited Lowell, the sec- Traditional and innovative police responses tor captain from the hardest hit sector responded When a relatively minor problem, such as that of by blanketing particularly vulnerable neighbor- the smashed side windows, appeared isolated, ran- hoods with crime- prevention information, pre- dom and offered few leads—“We don’t know the senting this information at weekly community ‘who, what, when, and how?’” of the crime—the meetings, identifying targeted motor vehicles and department was more likely to turn to traditional publicizing this information in the local newspa- responses. When the problem was more serious pers and on local radio stations, increasing motor and intractable, there was greater likelihood that vehicle enforcement and patrol, and making more the sector captains would employ a mix of tradi- arrests. He also took less traditional approaches, tional and more innovative police responses. There such as requesting the CAU to run an analysis of are probably a couple of reasons for this: (1) Sec- motor vehicle crime for Compstat, soliciting a tor captains have limited resources of time, money, problem-oriented-policing approach from one of and manpower, and they prefer to allocate them his sergeants, using the assistant crime analyst’s to more serious problems; and (2) Sector captains gang database, contacting twenty other police de- are under more pressure to solve serious than partments about suitable responses—Davis sug- minor problems in their sector. Consequently, it gested that he contact Newark—and using this is likely that they will try a broader array of strat- information to deploy a decoy car. egies to demonstrate to Davis that they really are In most cases, the sector captain had overall doing something and to increase the likelihood responsibility for coming up with problem-solv- that a combination of these strategies, as opposed ing tactics and implementing a crime-reduction to a single approach, will succeed in mitigating strategy in the sector. While assuming the lead in the crime problem. We might also surmise that a the decision-making process, the sector captain combination of traditional and innovative strate- may have taken suggestions from Davis and other gies reduced the overall risk of censure to the sec- participants at Compstat and may have also elic- tor captain. Should a sector captain implement an ited advice and feedback from lieutenants and ser- experimental and creative problem-solving ap- geants; it was less likely that he would approach proach in lieu of a more traditional response, the line officers for suggestions. In other words, there failure of the more innovative strategy may open may have been some “bubble-up” from the lower him to possible ridicule and reprimand. Put sim- ranks, but ideas tended to originate from the top ply, the implementation of a traditional response and trickle down the chain of command. poses the lowest risk of opprobrium to district Since Compstat rewards “something being commanders in departments that do not particu- done”—otherwise, as one sector captain put it in larly encourage or reward innovation. Of course, reference to the persistent motor vehicle problem, just because an approach is innovative does not “You might get your head broken”—problem-

45 Compstat and Organizational Change in the Lowell Police Department

solving efforts tended to center on the implemen- Innovative police responses tation of a series of alternative approaches rather than the meticulously researched, carefully coor- Many incidents that are identified at Compstat are dinated, in-depth problem-solving response ad- temporary crime blips that will disappear by the vocated by Herman Goldstein (1990). If approach time the next Compstat period rolls around. For A did not work at first, the sector boss might try instance, following a period of heavy January A again, move on to approach B, or try a combi- snowfalls, several motor vehicles had their tires nation of A and B, and so on. In all fairness, it is slashed on one of Lowell’s streets during a single important to acknowledge that motor vehicle night. Compstat encourages a focus on patterns crime is a difficult problem to resolve. A sector over time and the identification of underlying captain explained that motor vehicle crimes were causes, but this does not mean that there are any. very often scattered throughout the city rather In this case, the crime series only occurred once than concentrated in specific areas or hot spots, and may have been motivated by frustration over took place at different times, were perpetrated by the lack of available parking, which was already many different types of suspects, and were com- at a premium in the congested area and exacer- mitted for a variety of reasons, including vandal- bated by snow clearance. Once crime analysis did ism, gang-related retaliation or initiation, drug- not reveal any patterns, a routine response was related thefts, joyriding, selling the parts to chop appropriate—take a report and assign follow-ups. shops, and defrauding insurance companies. There There are some incidents, in contrast, that are was no magic bullet to the problem, Davis sur- clearly related, and their resolution may demand mised. Another respondent wondered how useful significant time, resources, and creative efforts. Goldstein’s approach was for solving these types Such was the case when the department dealt with of crime. Given the many victims, perpetrators, a dilapidated boarding house whose residents en- and locations, it seemed unlikely that there was gaged in a great deal of criminal activity. The sec- one underlying cause for the problem. In decid- tor captain’s response was at once one of the most ing upon a suitable response, a sector captain must innovative and effective problem-solving strate- also balance how much attention can be focused gies we witnessed at Lowell, an illustration of how on these crimes in comparison to other sector the department can use city agencies to tackle dif- concerns. Putting out a decoy car might look good ficult crime problems and a demonstration of how on paper, but “in terms of resources—time, money, problem-solving strategies can “bubble up” from manpower,” as a respondent observed—“it was not line officers. This was not a commonplace occur- a good method . . . it pulled cops away from other rence, but it cautions us not to overgeneralize things, and it was not cost effective.” about a department’s limited capacity to imple- Despite these attempts to reduce motor vehicle ment innovative strategies and use the decision- crime, some still felt that more needed to be done: making abilities of its rank and file. “We need to put more emphasis on this.” In Janu- When we arrived at Lowell, the department ary 2001, one respondent suggested that it might was having serious and persistent problems with be time for the department to “look at this differ- the residents of an old, derelict rooming house, ently” and perhaps put together a coordinated re- known as the Surf Building, in downtown Lowell. sponse between sectors and outside agencies, in- Located near several bars and occupied by va- cluding other police departments. Another felt that grants, criminals, and down-and-outers, the Surf there was a pressing need for a task force, and he Building was recognized by the police as a “stag- suggested collaborating with the state police and ing area” for crime. The department responded insurance companies. At the beginning of 2002, almost every day to numerous incidents involv- the department eventually formed an auto theft ing alcohol abuse, drug activity, vandalism, and task force with the city’s fire department inspec- serious crimes involving robberies and stabbings. tors and city inspectional services. The task force These problems continued unabated despite in- focused on closing down auto shops operating creased police activity and enforcement and ulti- without a license. mately inspired the sector captain, who inherited

46 Compstat and Organizational Change in the Lowell Police Department

them upon assuming command, to make the Surf dred calls to the department between January and Building his sector’s major concern. To address the November 2000. The department then illustrated problem he implemented some traditional strate- the magnitude of the problem by presenting this gies and also pursued two innovative problem- data at a city council meeting. Its efforts to thus solving approaches to great effect. The first was end the city’s legal foot-dragging succeeded in for patrol officers to conduct bed checks after 10 December 2000, when the city evicted and relo- p.m. and to arrest all those who were not legal cated the Surf’s residents and finally shut down residents of the building. The second was for the the building. department to encourage city administrators to The Surf example shows how innovative prob- secure a court order requiring the recalcitrant land- lem-solving strategies that come from both upper lord to close down the building. These police ac- and lower levels of police organizations can work tivities, as the sector captain noted, required ac- together to solve intractable, long-term crime tive involvement from both administration and problems. Compstat also played a role in this pro- patrol and represented a clear example of a coor- cess by enabling the department to identify recur- dinated response to a difficult problem. ring incidents and maintain its focus on a specific Sector sergeants and patrol officers contrib- problem area. The department then used the data uted to the department’s response by coming up to convince external parties, including the mu- with the idea of asking officers who walked the nicipal government, local newspaper, and city resi- Surf beat to conduct bed checks. Thrilled by this dents, that only prompt and vigorous action would display of initiative, the sector captain proposed provide an effective and lasting solution that the police use this method to take names, Since the appearance of Goldstein’s (1990) conduct arrests, and generally make it clear that seminal work on problem-oriented policing, in- they would no longer tolerate crime and disorder novative problem-solving responses have been in this area. In addition, the information gathered closely associated, at times uncritically, with pro- by officers promised to help to identify suspects, gressive or “cutting-edge” police departments. provide evidence for the wide variety of criminal Compstat claims to encourage innovative prob- incidents associated with the Surf Building, and lem solving, but crimes that police regard as par- facilitate the city’s efforts to resolve the problems ticularly prevalent, random, difficult to prosecute, associated with this location. and obviously symptomatic of much larger struc- One obstacle to this goal was that Lowell resi- tural problems—poverty, unemployment, drug dents generally recognized the Surf as a historic addiction, and disenfranchised youth—may be building, so city officials worried that closing it less likely to be the focus of a Compstat-related, and displacing its residents would be legally diffi- problem-oriented-policing strategy. Certainly the cult and politically unpopular. This official reluc- likelihood of a successful outcome to these types tance notwithstanding, the sector used crime data, of crime is limited. There may be some patterns with the aid of the CAU, to demonstrate that the to domestic violence and motor vehicle crimes, Surf Building had accounted for some three hun- but their complex and persistent nature generally interacts with other important considerations, such as limited police resources and a need to fo- cus on more serious crimes, to mitigate the likeli- . . . crimes that police regard as hood that they will be the focus of a particularly particularly prevalent, random, difficult innovative crime-control strategy. Interestingly, the to prosecute, and obviously sector captain who helped close the Surf Building symptomatic of much larger structural believed that having a physical location on which to focus significantly enhanced her problem-solv- problems . . . may be less likely to be ing success. She contrasted this to the “multi-di- the focus of a Compstat-related, mensional location” of motor vehicle crimes, as problem-oriented-policing strategy. well as the multiple reasons for why youths were vandalizing, breaking into, and stealing automo-

47 Compstat and Organizational Change in the Lowell Police Department

biles. These factors made finding approaches that vorable stories and “to help push the police force would dramatically reduce motor vehicle crimes into an aggressive law enforcement mind-set” significantly more elusive (Sherman and Weisburd (Silverman 1999, 91). The department did not 1995; Weisburd and Braga 2003). focus on disseminating Compstat maps and crime In sum, based on our findings at Lowell, the data more generally to facilitate input on crime claim that Compstat fosters problem-solving in- issues and its performance, but Compstat could novation and experimentation, based upon brain- provide the police with the opportunity to garner storming in Compstat meetings, collaborations feedback from a mass audience. between departments, and a close examination of In communicating information about the research evidence, is a serious overstatement. department’s mission, Compstat can be used to Compstat’s data orientation did seem to affect what create knowledge about what the department is problems were identified and, thus, when and doing and accomplishing and to acquire informa- where responses would be mobilized, but it had tion from others. Members of the organized com- generally done little to stimulate analysis of how munity play an important role in this process. By to actually respond on the basis of the data. This inviting members of the press, representatives from is not entirely surprising, inasmuch as the sort of other municipal organizations—including public sophisticated data analysis that might go beyond works, inspectional services, and probation—local targeting problems and hot spots was not yet part dignitaries, and visitors from other police depart- of the department’s routine. Unable to break down ments, a police agency can use Compstat to se- general crime patterns, such as those involving cure community support and enhance its legiti- motor vehicles, into the multiple patterns that macy. With its sophisticated maps, officers in full contributed to them, department personnel were regalia, and tension-filled atmosphere, Compstat “flying blind” as to how to innovate effectively. is impressive theater. Consequently, it is a power- That they did find ways to innovate on some seri- ful means of helping the organization earn credit ous problems must be duly acknowledged. Inno- for its accomplishments and enhancing its pro- vation in the police response was not absent from fessional reputation among peer agencies and the Compstat process in Lowell, but it was hardly other police leaders. In addition to public recog- the hallmark that Maple said was an essential nition, Compstat meetings allow municipal actors element. to share useful information and offer their services. These collaborations might help district com- External Information Exchange manders implement more effective crime strate- gies and enable the department to attain its goals A final element in the Compstat process is exter- At Lowell, members of the press, command nal information exchange, which is composed of staff from neighboring police departments, local two critical components. Compstat provides a dignitaries, such as a visitor from Boston’s Attor- police department with: (1) the means of com- ney General’s office, and academics from nearby municating information to the public, other mu- universities all attended Compstat at some time nicipal actors and organizations, the press, and during our research. These visitors often appeared peers about what the department is doing and how dazzled by the show, and Compstat made a simi- well it is accomplishing its mission; and (2) a po- lar impact on the local press. When it was first tential mechanism for outside constituents to pro- implemented, it received significant and positive vide departments with input in the form of evalu- publicity in The Lowell Sun, but interest seemed ations and expressions of needs and priorities from to have waned by the time we arrived at Lowell. the community and even to create consequences Even though reporters continued to be granted for performance. In New York Bratton, gave the free access to Compstat meetings—although they “elite” community—the press, researchers, local were not allowed to release crime details—they government officials, and visitors from other po- attended far less frequently than other visitors. An lice department—greater access to the depart- archival search of The Lowell Sun only revealed ment’s inner sanctum in order to encourage fa- one reference to “Compstat” in the past year—

48 Compstat and Organizational Change in the Lowell Police Department

subsumed within an article on Davis’ recent lead- But Compstat certainly could operate as a ership award from the Police Executive Research mechanism for acquiring input from ordinary citi- Forum (PERF). zens on crime problems and making the depart- The chief probation officer of Lowell’s district ment accountable for its performance—the sec- court was the only representative of a municipal ond principal component of external information organization who attended Compstat regularly. He exchange. Perhaps Compstat does not resonate as often participated in discussions with the sector powerfully with local citizens as community po- captains, providing information on the names and licing because the responsibilities of district com- addresses of suspects, and he was especially manders and the utility of computer maps do not knowledgeable about the city’s gangs—their mem- have the same allure as the promise of extra officers bers and turfs. We did not observe officials from walking their beats. When Davis mused that he other local municipal organizations attending was unsure whether Compstat had “won the pub- Compstat on a regular basis, but this did not sug- lic relations battle like community policing” and gest that they were absent from the Compstat pro- added, “It is just management people that look cess. Sector captains communicated frequently with interest on Compstat,” he might have been with city agencies in order to help them imple- acknowledging these particular characteristics. ment their crime strategies. For example, in an Nevertheless, a department could use Compstat attempt to reduce thefts from automobiles in the to disseminate crime data and information on its city’s parking garages, a sector captain was in con- efforts to reduce crime through media outlets, such stant contact with Lowell’s parking authority. as the press and department Web site, and through Through phone calls and memos, the sector cap- existing community policing programs, such as tain urged officials in city hall to post signs en- neighborhood meetings. Thus, Compstat has the couraging citizens to remove valuables from their potential to encourage a mass audience to play a parked cars. more integral role in policing. Davis was keen on Although the department did not actively dis- the using the local newspaper and the department courage community members from attending Web site to disseminate Compstat information, but Compstat, we did not observe any of them at the he was constrained by outside forces. The local meetings. That information was shared primarily paper’s editor was worried about portraying the at the macro-level with local officials and the press city in a negative light, Davis told us, and the city is consistent with our more general observation council was unable to provide the department with that Compstat functions as a top-down approach additional funds to improve its information tech- to policing. Much of the information sharing and nology. brainstorming occurs among the high-ranking Despite these constraints, our observations at members of the department and some of the “in- Lowell suggested that the department had moved siders” who attend Compstat. In addition, these Compstat in this direction. The department pro- meetings help publicize the department’s efforts vided timely crime data to the local newspaper, to reduce crime. Certain visitors are afforded privi- mentioned Compstat on its Web site, http://www. leged access because they can help generate posi- lowellpolice.com/index.htm, and discussed tive attention for the department’s accomplish- Compstat maps at community meetings. Never- ments and offer information that may contribute theless, its use as a mechanism for eliciting com- directly to the department’s success in meeting its munity input was infrequent and extremely lim- goals. In comparison, ordinary citizens and mem- ited, and Compstat hardly functioned at all to bers of the rank and file do not generally attend promote external accountability. We should note or participate in the Compstat process. In this re- that time constraints restricted our ability to ex- spect Compstat differs from community policing, plore this element of Compstat as fully as we where departments are encouraged to elicit input would have liked. We only attended one formal from community members at a grass-roots level meeting between a sector captain and a local rather than just contact city leaders and munici- neighborhood group, and we did not have the pal organizations. opportunity to interview local residents. As a re-

49 Compstat and Organizational Change in the Lowell Police Department

sult, our analysis of this component of external complishing its goal of crime reduction. It is pos- information exchange draws more on our general sible to envision Compstat as a more comprehen- impressions and on comments made during in- sive mechanism for encouraging community terviews than on systematic observation of the members to provide the department with helpful interactions between police and community. information. Should departments post Compstat One way that Compstat can be used to foster maps on their sites, residents could learn the types community feedback is if its products—maps and and general locations of crime in their neighbor- statistical analysis—are made available to the lo- hoods and could offer specific insights into pos- cal media. At Lowell the department did use sible causes and solutions. Even though the LPD monthly crime statistics in its statements to the did post descriptive information on each of the press. For example, one newspaper story identi- city’s sectors, it did not make any Compstat- fied how an increase in motor vehicle crimes in generated data available. one area of the city was contributing to the city’s Sometimes Compstat maps and crime statis- annual crime increase. This Compstat-generated tics were made more visible at regular monthly information was in the form of percentages and meetings between sector captains and local neigh- did not contain any maps, crime analysis, or ref- borhood groups. Since there were four or five erence to Compstat. Our neighborhood groups for observations and a brief It is possible to envision each sector, on average a review of The Lowell Sun sector captain would at- from 2002 gave us the Compstat as a more tend a community meet- impression that the LPD comprehensive mechanism ing every week, though used Compstat data to for encouraging community sometimes another of- help explain annual or ficer would appear on the monthly crime increases members to provide the sector captain’s behalf. At or decreases but did not department with these meetings, sector provide the press with helpful information. captains were respon- the sophisticated and up- sible for soliciting crime- to-date crime statistics related information, ad- and maps that were produced regularly by the dressing community concerns, and suggesting CAU. This was a small change from the conven- crime-prevention techniques. In order to help ac- tion of releasing statistics from the Uniform Crime complish these goals, a sector captain might men- Reports once or twice a year; a more significant tion Compstat and provide a few Compstat maps. change might have involved reporting crime pat- Several respondents observed that the frequency terns or providing an assessment of problem- with which Compstat-generated maps were used solving strategies to the local newspaper. Thus, depended on the style of the individual sector cap- Compstat seemed to serve more as a means for tain, but the maps were present at the neighbor- the department to establish legitimacy by show- hood group meeting we attended. Our observa- ing that it was knowledgeable and responsive to tions suggested that these products mainly served crime trends than as an important accountability to reassure residents that local crime problems function through the press. were being quickly identified and tackled, and, Many police departments have constructed, or on a much more modest level, to get some feed- are in the process of building, their own Web sites. back from the community. Currently these are used to give some general At the one community meeting we attended, background on a department, including its key the sector captain provided a one-page crime map personnel, community programs, organizational depicting burglaries, motor vehicle breaking and structure, and mission statement. Some of these enterings, traffic accidents, and stolen cars over a Web sites, like that of Newark, New Jersey, are two-week period. The rest of the handout con- also used to disseminate crime statistics that serve tained contact information and crime-prevention to illustrate how well the agency is doing in ac- tips, such as “How to keep you and your property

50 Compstat and Organizational Change in the Lowell Police Department

safe.” The sector captain stated that he selected (Manning 2001). Other external constituents, the crimes because, “From my experience, these however, were not entirely removed from Comp- seem to be what people care about.” In addition, stat. The community meeting we attended dem- he noted that any information from community onstrated that it could be used as a tool to en- members on these crimes would really help the courage community members to support their police. Finally, the sector captain explained that local police department. Giving community mem- the Compstat map “not only gives a geographical bers less restricted access—no matter how small— illustration of neighborhood crimes to my com- to “internal” information and police crime con- munity members, but it also provides me with a trol strategies than under more traditional policing specific focus” and ensured that the community models may help increase local residents’ overall meeting “did not wander all over the place.” The investment in the success of the department’s ef- maps could also be used to show residents that forts to reduce crime. the department was successfully tackling quality- of-life and crime problems. A precipitous decrease in traffic accidents provided the sector captain with VI. Discussion the opportunity to convince his audience of Compstat’s utility and a notable department suc- Results from our national survey suggested that cess: Later in the meeting he revisited this theme Compstat was a policing phenomenon sweeping of closely monitoring crime problems by telling across the nation. We have argued elsewhere that his audience that the police “visit everything [not this is probably due to the publicity it received as just the crimes depicted on these maps] daily.” a creation of the NYPD which is the nation’s most He also introduced another key Compstat element visible police department, the fact that some de- when he stated, “We are truly held accountable partments had already implemented Compstat for what is going on.” elements before the term Compstat was coined, In raising the issue of accountability, the sec- and Compstat’s easy adaptability to the traditional tor captain was clearly referring to the pressure hierarchical organization of U.S. police depart- that Compstat places on middle managers to re- ments (Weisburd et al. 2003). However, the popu- duce crime. But Compstat may also serve as a larity and praise afforded Compstat by many po- mechanism for holding an entire department ac- lice practitioners and scholars is fueled more countable for meeting its goals. For example, in- heavily by rhetoric and anecdotes than a body of formation on a department’s problem-solving strat- systematic research. Rather than focus on the egies could be disseminated to the public to enable NYPD, we decided to contribute to a more com- community members to assess the department’s prehensive understanding of Compstat by exam- efforts to reduce crime. During our time at Lowell ining how it was implemented and operated in a we did not observe Compstat operating in this way, police agency of much smaller size and different but this is not surprising. Such a finding would organization. What insights could Lowell teach be consistent with research that suggests that even us about Compstat? with the advent of community policing and its We focus our main findings around three ba- promise of closer police-community partnerships, sic questions: (1) What was the dosage or amount police departments remain reluctant to cede any of each Compstat element within the depart- of their decision-making autonomy to community ment—very high, high, moderate or low?; (2) members (Lyons 1999). What were the opportunities and challenges as- In sum, Compstat’s symbolic capacity, or “dra- sociated with each element?; and (3) How well maturgy,” to impress audiences about departments’ did each element integrate with the others to form efforts to reduce crime and its focus on the utility a coherent program? Our responses to these ques- of information exchange with other municipal tions suggest that each core element can be quickly officials help explain why members of the elite adopted by police agencies, even smaller ones with community were granted much greater access to more limited resources, and can have some posi- Lowell’s Compstat program than local residents tive effects on how they operate. However, the

51 Compstat and Organizational Change in the Lowell Police Department

dosage level of those specific elements that de- year encapsulated this approach in a bold mission mand significant change to a department’s daily statement that left little uncertainty about his vi- work and management is likely to be lower than sion inside and outside of the department. Aware those that reinforce traditional structures and ac- of the department’s focus on formal bureaucracy, tivities. Thus our qualitative research supports the one that had exacerbated divisions between man- finding from our national survey that Compstat is agement and the rank and file, Bratton’s rallying easily adapted to the traditional police organiza- cry was positively received within the NYPD. Upon tion but does not move the entire organization in his arrival, patrol officers had ranked fighting a new direction. Resistance to change in police crime a lowly seventh on a list of activities that organizations is well documented (Guyot 1979; they believed the department wanted to see from Sparrow, Moore, and Kennedy 1990), but our them. They afforded administrative and manage- findings suggest that Compstat contains some hid- ment concerns the highest priority; writing sum- den dilemmas and paradoxes that further limit its monses and holding down overtime ranked first ability to work effectively as a fully functioning, and second respectively (Heskett 1996). By ap- crime-fighting program. The most important of pealing to the self-image of the police as crime these in Lowell is the premium that Compstat fighters, Bratton reconnected line officers with the places on making district commanders responsible organization’s primary purpose. In addition, he for having an immediate strategic response to made its symbolism more authentic and endur- crime in their specific beats. We argue that the ing by repeatedly rewarding officers who came up powerful effects of internal accountability com- with effective crime-reduction strategies. He also promise the integrity of other elements of the pro- packaged his mission for consumption outside the gram: They undermine the innovative problem- department through frequent and highly visible solving process, and they discourage the flexible press releases and news conferences. Realizing that reallocation of resources to crime problems across the public was disenchanted with department sectors. scandals and rising crime rates, Bratton’s promise of a more aggressive approach toward corruption Mission Clarification and public safety appealed directly to city residents (Heskett 1996, 5–7; Silverman 1999, 88–91). In In contrast to the confusion surrounding the spe- essence, the power of Bratton’s mission as a belief cific goals of community policing (Bayley 1988), system rested upon it being “concise, value-laden, Compstat’s objective is elegantly simple: to fight and inspirational” (Simons 1995, 82). crime. Compstat revitalizes the attitude that the Davis, like Bratton, provided a bold and vis- police are primarily crime fighters, one that has ible mission statement with crime reduction as its fallen into disfavor over the last thirty years among sine qua non. The mission to make Lowell “the progressive reformers who underscore the impor- safest city of its size in the United States” appealed tance of community problem solving (Goldstein to city residents and has been widely accepted in 1979; Wilson and Kelling 1982). More ambi- the department. In our survey, nine out of ten of tiously, Compstat’s proponents argue that the po- the rank and file recognized that reducing violent lice actually have the capacity to influence crime crime and improving the quality of life in city rates (McDonald 1998; Bratton 1998; Kelling and neighborhoods were “very” or “somewhat impor- Sousa 2001). The articulation of a clear mission tant” to the department’s Compstat strategy. Since statement centered on results, not effort, is de- the value of fighting crime was widely shared, we signed to make a major contribution to the thrust assess the overall dosage of this core element as of this message. “high.” We also note that the Compstat mission When Bratton became commissioner of the does not need to be antithetical to community NYPD, he rejected the “kinder-gentler” commu- policing. Bratton may have dramatically shifted nity-policing style of his predecessor, Lee Brown, the NYPD’s focus away from community policing, in favor of a stance that was tough on crime. His but Superintendent Davis continued to balance his promise to reduce crime by 10 percent in the first department’s crime-fighting efforts with a focus

52 Compstat and Organizational Change in the Lowell Police Department

tackling and reducing crime, as Compstat’s prin- Lowell’s Compstat meetings put cipal element. This component did have some sector captains under positive effects: The top echelon and district com- considerable pressure, as widely manders certainly seemed more aware of their crime environment than they were before. This recognized by the rank and file. was a major change from earlier operations, and sector bosses were motivated to do something about crime problems that arose—getting on top on working with citizens, city agencies, and com- of problems—much more than before. munity groups to identify and solve problems. Lowell’s Compstat meetings put sector cap- In contrast to Bratton, Davis established a tains under considerable pressure, as widely rec- benchmark for success that might have been over- ognized by the rank and file. Even though line ambitious, as he himself recognized. The superla- officers did not feel this pressure directly, what tive “safest” leaves no room for failure, and the transpired at Compstat took on mythic propor- goal implies a continued ability to meet the same tions among the lower ranks. This suggested that exceptional standard for success—it is not enough internal accountability was a very important com- to become the safest city, the city has to stay in ponent of daily operations, and we assessed the pole position. Ironically, the same audacity that dosage of this Compstat element as “very high.” helped motivate residents and mobilize police of- Compstat appears to be a mechanism that ficers also fomented criticism when crime, after a makes district commanders highly responsive to long decline, began to rise. Mission statements the chief, but it does not provide a similar struc- present chiefs with a dilemma: How do they rec- ture to facilitate the responsiveness of patrol of- oncile the benefits of a bold objective with the lack ficers to district commanders. This depends on of confidence and public disapproval that might the will and skill of the individual district com- accompany an organization’s failure to attain its mander. This finding differs from Kelling and goal? As leaders of their organization, chiefs are Sousa’s assertion that accountability, “Drive[s] the particularly vulnerable to criticism from external development of crime reduction tactics at the pre- constituents and their own officers. Despite these cinct level” (Kelling and Sousa 2001, 11). To over- risks, Compstat requires that chiefs raise expecta- come the challenge of disproportionately burden- tions of performance both inside and outside of ing district commanders with reducing crime, the department. Perhaps that leaves three imper- holding smaller meetings at the precinct level, and fect “solutions:” (1) Only propose a bold crime- having sector captains, lieutenants, line super- reduction goal in cities where the existing crime visors, and patrol officers attend them, could help problem is so bad that department efforts are likely spread the sense of accountability among all to meet with a high level of success; (2) When officers. success is less likely, sacrifice the motivational In fact, we discovered that the onus of account- value of the mission for a more modest crime-re- ability placed upon sector commanders might ac- duction goal; or (3) Leave the organization after tually sabotage Compstat’s capacity to reduce an initial success before crime rates have the crime. Paradoxically, two other Compstat ele- chance to rise again. Given the pressure on chiefs ments—the practice of innovative problem solv- to succeed, it is not surprising that John Diaz, an ing and the organization’s capacity to allocate re- assistant police chief in Seattle stated that, “The sources to the most pressing problems—were politics of being a police chief have become so undermined by the burden of accountability insane no one wants the job” (Bratton 2001). placed on district commanders. Compstat is in- tended to facilitate the creation of an information- Internal Accountability seeking, data-driven, collegial, and reflexive en- vironment that thrives on the willingness of its In Lowell, many regarded internal accountability, members to think creatively, to share information, or making a specific individual responsible for to be critical of themselves and others, and to act

53 Compstat and Organizational Change in the Lowell Police Department

on this criticism in a positive way. In many re- to cover anti-terrorist activities, sector personnel spects, this represents the ideal of how profession- may be reluctant to share resources with one an- als, such as academics, engineers, and scientists, other. In spite of its claim to encourage the shift- are supposed to operate within their different in- ing of resources to the most pressing problems, stitutional environments. Atypical problems are Compstat’s emphasis on holding middle manag- presumably best resolved in a less hierarchical ers accountable for lowering crime statistics in environment, or at least one that is buffered from their precincts hinders a more collective approach the very sort of rapid-response-time accountabil- to crime. ity that Compstat inspired in Lowell (see Mas- trofski and Ritti 2000 for an overview of the lit- Geographic Organization of erature). Davis tried to foster brainstorming and Operational Command believed that a system like Compstat could en- courage individuals to think differently. Ironically, Operational command at Lowell had been low- it is the operation of this system within a para- ered one level to the sector captains who were military, hierarchical organization—one that so- given twenty-four-hour responsibility for their cializes officers to support one another and defer specific beats—a change from the period prior to to rank—that hinders innovative thinking and Compstat’s implementation. Sector captains were collaborative learning. Given the high standard of now empowered to make a wide range of admin- accountability to which sector captains were held, istrative, managerial, and tactical decisions that many members of the command staff were reluc- included approving vacation time, solving person- tant to share their insights at Compstat out of con- nel problems, and deploying officers to problem cern that these might appear as criticisms of the areas. In addition, sector captains were granted sector captains. Lower-ranking officers were es- broad access to other departmental resources, such pecially reticent to make suggestions. Since there as specialized units, in their problem-solving is an unwillingness to appear disputatious, par- efforts. ticipants at Compstat express broad and support- Despite this devolution of decision-making ive opinions rather than engage in incisive and authority to the sector level, Compstat continued critical debate. to reinforce the ideal of top-down control. Davis Furthermore, since Compstat put pressure on encouraged his sector captains to take initiative, Lowell’s district commanders to have a crime strat- but he also exercised his decision-making author- egy in place before the biweekly meetings, the util- ity directly over them. In approving a request for ity of discussing alternative crime strategies was overtime or abrogating a beat change, Davis dem- seriously curtailed. The pressure of Lowell’s system onstrated that he was willing to intervene when of internal accountability hindered innovative he disagreed with a sector captain’s decisions. approaches to crime by demanding so-much-so- Furthermore, a high level of centralized command quickly in a work culture that avoided contradict- and control penetrated down the organizational ing those who felt this accountability most in- hierarchy. Significant decision-making autonomy tensely. It did not, as Kelling and Sousa claim, may have been granted to each sector, but it did “constantly challenge precinct commanders to not exist at the precinct level. In addition, spe- develop new responses to crime problems” (2001, cialist units were not placed directly under a sec- 12). tor captain’s control, suggesting that many enforce- A final paradox associated with the premium ment and deployment decisions continued to be Compstat places on accountability is that it re- functionally, not geographically, based. We esti- duces organizational flexibility by fostering com- mated the dosage of this element as “moderate.” petition rather than collaboration across geo- In sum, our findings suggest that operational graphical sectors or precincts. When resources are command under Compstat may subvert one of the scarce, which is especially likely to be the case in central tenets of community policing—decentral- smaller departments or departments that are now ized decision making. By reinforcing the formal under pressure to stretch their limited resources hierarchy of rank and giving sector commanders

54 Compstat and Organizational Change in the Lowell Police Department

most of the decision-making power to identify and city politicians and business owners to deploy of- solve problems, Compstat works against the no- ficers to certain areas of the city at specific times. tion of many community-policing advocates that The reallocation of officers was also hindered by line officers must be granted greater decision- Davis’ obligation to support the Lowell Housing making autonomy. Absent an attempt to devolve Authority. It would be extremely difficult for any this authority to the street level, Compstat con- department to control these pressures from out- tributes to the existing divisions between manag- side forces. ers and line officers for which professional bureau- In addition to traditional internal and exter- cratic models of policing have been highly nal constraints, Compstat itself limited organiza- criticized (Kelling and Moore 1988). tional flexibility in one of the paradoxes we previ- ously noted. By acting to ensure the responsiveness Organizational Flexibility of each individual sector captain to the superin- tendent, internal accountability helped foster com- Regularity is a necessary and desirable character- petition between the sectors and discouraged col- istic of the police organization, and it is achieved laborations. As a system, Compstat explicitly through permanent structures and routines. Many rewards district commanders for handling their of the features of the traditional bureaucratic po- own beat problems effectively, but it does not con- lice organization, such as rules and regulations tain a similar mechanism for rewarding the more governing officer behavior or the hierarchy of efficient sharing of resources—even if this could rank, deliberately restrict flexibility in favor of pre- contribute to an overall reduction in crime. A chal- dictability and order. Compstat, however, demands lenge to a police department implementing Comp- that police organizations develop flexibility to deal stat will be how to balance the requirement of with emerging or unforeseeable problems. At holding district commanders accountable for Lowell, the top leadership did encourage team- specific territories against a capacity to shift re- work and coordination, but the strategic realloca- sources across precincts—away from where they tion of resources operated on an ad hoc basis. Pa- are needed to where they are needed most. With- trol officers were commonly asked to pay attention out a structure that specifically recognizes and to a particular area during their shift. Less fre- rewards the willingness of district commanders quently, a specialist was temporarily assigned to a to share valuable resources and collaborate with crime problem. It was even less common for a sec- other precincts, it is unlikely that a small police tor captain to request additional funds, such as department can maximize its organizational overtime, and the department did not rely upon flexibility. taxi squads to enhance its flexibility. Davis told us that sector captains would willingly share their Data-Driven Problem patrol officers if a crime occurred that warranted Identification and Assessment additional resources, but we never witnessed the redeployment of officers across sectors; and sec- At Lowell, the department expended a great deal tor captains reported that it rarely occurred. We of effort to overcome several obstacles to the col- rated the treatment level of this element as “low.” lection and processing of timely data. Its success Why did the department continue to rely is testimony to how Compstat can be implemented heavily on routines in response to crime problems? despite significant technical problems. Our re- The LPD did attempt to increase its flexibility, but search also suggests that small departments can it was hampered by traditional internal and exter- start Compstat inexpensively by relying upon ex- nal challenges that are very difficult for an orga- isting computer equipment and hiring civilian nization to overcome. Budget constraints pre- analysts. Whereas crime data was formerly used vented the hiring of additional officers, and the halfheartedly once a year or so, it has now be- number of specialist assignments restricted how come an integral part of the day-to-day dialogue many officers were available for routine patrol. In between Davis and his middle managers, especially addition, the department was under pressure from sector bosses. Compstat data now establish the

55 Compstat and Organizational Change in the Lowell Police Department

structure and information used within the agency crimes with homicides in a sector to see if an in- for identifying problems, establishing priorities, crease in homicides is likely drug related—would and deciding where and when to mobilize re- be a more powerful method for detecting plau- sponses. As one Lowell officer stated bluntly, sible patterns than just relying upon maps. Re- “Concern over information is different from ten sources permitting, departments could further years ago when police officers did not give ‘two enhance the use of data to identify and assess prob- craps’ about the UCR.” lems by providing middle-level managers with Despite this increased use of data to inform their own civilian crime analyst. Given the nu- decision making, middle managers were not given merous command and administrative responsibili- any special training to help develop their data ties of district commanders under Compstat, it is analysis skills. Identification of the location of unlikely that they would have the time to con- crime problems had become significantly more duct an in-depth data analysis of crime in their rapid, but data were not analyzed to figure out beats without this additional source of support. exactly how to mobilize in order to decide exactly Finally, increased training could raise the level of what to do. Sector captains were comfortable iden- appreciation for conducting, interpreting, and tifying patterns using descriptive statistics, but for applying a more research-oriented paradigm the most part this only occurred during the two among all of a department’s supervisors and line or three days before Compstat. Instead, they re- officers. At the very least, it would help teach the lied upon serially reviewing reports, without rely- rank and file the importance of writing accurate ing upon statistics or maps, and then forming and complete reports. impressions. Furthermore, the very little time that the CAU had to analyze data and check it for ac- Innovative Problem-Solving Tactics curacy, and the modest level of appreciation among supervisors and the rank and file for the useful- Supporters of Compstat claim that it encourages ness of data analysis, suggested that this element innovation and experimentation. Police are en- was underdeveloped in the LPD. Consequently, couraged and expected to look beyond their own we assessed the dosage of this element as “mod- experiences by consulting with other departments erate.” and examining theories and research on crime It is highly unlikely that sophisticated data prevention. Our research suggests that, on occa- analysis will play a fundamental role in shaping sion, Compstat has encouraged innovation, espe- the adoption of specific crime strategies through- cially with more serious problems, and that this out police departments without additional train- might occur, albeit on an ad hoc basis, at the sec- ing. Higher levels of data analysis require the kind tor level. On the whole, however, it has made only of skills and knowledge that cannot be merely modest inroads into previous practice, so we esti- picked up while working on-the-job. For example, mated the dosage of this element as “low.” a familiarity with tabulating bivariate relation- The department’s continued reliance on tradi- ships—such as tabulating the frequency of drug tional police responses can be attributed to two other reasons: (1) Based upon their experience police officers believe these tactics work; and (2) These practices may be the only realistic option It is highly unlikely that given the limited resources available from the city. sophisticated data analysis will We heard many comments asserting the problem- play a fundamental role in solving effectiveness of increasing traffic enforce- shaping the adoption of specific ment and saturating areas with patrol. In regard to the use of a decoy car, some officers mentioned crime strategies throughout that the significant commitment of resources that police departments without this strategy required did not necessarily imply additional training. success. The time it took to organize the stake-

56 Compstat and Organizational Change in the Lowell Police Department

out, the extra personnel assigned to the case, and may be putting our future police leaders in a con- the additional disbursement of overtime still could ceptual box of their own devising. not replace the dose of luck needed to get the thieves to “hit” the decoy car. In all likelihood, a External Information Exchange chief would be pleased that his district command- ers only resorted to more creative and resource- The final element of Compstat is external infor- rich strategies when absolutely necessary. In ad- mation exchange. Given the time constraints, our dition, the department’s reliance upon traditional findings on this element are more tentative than strategies could be explained by the paradox that those on the other six elements, which were sup- we mentioned earlier. In Lowell, Compstat’s in- ported by more intensive participant-observation tensive emphasis on internal accountability de- research. Nevertheless, our research can provide mands that district commanders rush to do some- some insights into how external information ex- thing quickly about crime problems: Middle change operated at Lowell. managers are now as concerned about response Compstat can be used to generate external time as patrol officers once were. Compstat has support for a police department among peer agen- just moved reactive policing to another level. A cies and community stakeholders, and it can also more innovative police response would require be used as a mechanism for providing external that district commanders have sufficient time to constituents with information about a depart- foster, develop, and test long-term preventive ment’s goals and its progress toward achieving plans. Compstat has made this process more them. The interest with which the crime maps difficult. were received at the meeting we attended sug- Finally, the absence of any kind of systematic gested that Compstat did have a positive effect on follow-up at Lowell’s Compstat meetings hindered community members. In addition, our research innovation. Since the department did not have the suggests that Compstat facilitated the availability ability to evaluate whether different crime strate- of timely crime data to the press. Even though the gies were actually achieving their intended result, local newspaper did not print Compstat maps, it tended to resort to officers’ impressions of what department press releases sometimes used appeared to have worked in the past. The imple- monthly data to provide statements on specific mentation of a rigorous follow-up system to over- crime problems and police responses to them. Our come this challenge, however, might present an impression is that this represents a change from additional set of challenges to a police organiza- previous police practice, when the data available tion. The careful assessment of the effectiveness through the annual release of the UCRs was less of crime strategies would hold district command- detailed and less timely. ers to an even higher standard of accountability Despite these modest changes, external infor- than already exists. Not only would they be ac- mation exchange appeared to be an underdevel- countable for responding quickly to crime prob- oped element at Lowell: Compstat was not a well- lems, their responses would be scrutinized for publicized and well-understood police program possible weaknesses. The challenge, therefore, is outside of the department, and we assayed its dos- how to create a follow-up mechanism that facili- age as “low.” Certainly it enjoyed far less public tates the selection of the crime strategy that holds attention than Lowell’s community-policing pro- the best prospect for success but does not operate gram, and we are confident that those outside of to penalize individuals for their creative failures. the department could provide a far more compre- This is a tall order, given Compstat’s emphasis on hensive description of community policing than singling out district commanders and holding Compstat. In addition, Compstat was not used as them directly accountable for the department’s a tool for providing external constituents with mission of crime reduction. Long-term success detailed knowledge about the department’s probably requires innovation, and we learn more progress toward meeting its goals. Organizations from our failures than our successes. Compstat are obviously cautious about exposing their per-

57 Compstat and Organizational Change in the Lowell Police Department

formance to public scrutiny, and the provision of Lowell before Compstat was implemented, but detailed crime statistics and problem-solving ef- based on accounts from our informants, it is clear forts for external consumption could expose the that the implementation of Compstat has resulted police to the unwelcome pressure of increased in some impressive changes in organization prac- criticism. tice. Over only a brief period, crime analysis has In the absence of strong, external pressure, it come to play an integral role in police operations. is likely that police organizations will be reluc- This would have been impossible without the tant to give external constituents greater access to department’s tremendous efforts to overcome a their decision-making process and measures of number of technical data-entry and processing their performance. Doing so would expose a de- challenges. Decision makers are now more famil- partment to a higher level of accountability and a iar with the use of data and better informed about more critical assessment of its efforts to reduce what is going on in their areas. In addition, sector crime. It is much easier for a police agency to con- captains feel much more accountable for identi- vey a favorable public impression of itself if it can fying and responding to crime problems. Finally, control what types of information are made pub- members of the department have a much stron- licly available. Nevertheless, just as internal ger sense of the mission that their leader wants accountability is used to motivate district com- them to pursue. Although we determined that the manders within the police department, “public dosage of several elements’ implementation was information” about a department’s crime-fighting low or moderate, all appeared to represent a sig- efforts might “create the strongest pressure for im- nificant change from past practice. Since police provement” (Sherman 1998, 10). organizations are notoriously resistant to innova- tion and any substantial changes usually require many years to take effect, the achievements of the VII.Conclusions Lowell Police Department are especially note- worthy. When Mayor Giuliani gave his farewell address, But the actual operation of Compstat in Lowell he devoted a significant portion of his speech to has produced a pattern of practices that does not New York City’s success in reducing crime during readily fit with the idealized characterizations that his two terms in office. In the midst of the speech, have trumpeted its implementation in other ven- he described Compstat as one of the “pillars” of ues. The unevenness of the implementation has the NYPD’s policing model, and elsewhere he placed greatest stress on elements that are most praised Compstat as “a rational, reasonable, sen- consistent with goals that have long been em- sible, strategic response to crime.” His comments braced by top police managers in the United States: are not an aberration. Even though research on fighting crime and getting subordinates to carry Compstat is still in its infancy, Compstat is the out their leader’s priorities. Those elements that object of almost unconditional admiration from represent the greatest departure from past man- politicians, practitioners, and scholars who laud agement ideals—organizational flexibility, data- its capacity to operate as a sophisticated, effec- driven decision making, and innovative problem tive, and coherent crime program. However, our solving—remain the least developed. Some of this intensive site research at the Lowell Police Depart- underdevelopment is attributable to the paradoxi- ment, a department that is nationally recognized cal conflicts between Compstat’s elements that we for its adoption of some of the latest innovations have highlighted. By opting to stress internal ac- in police management techniques and strategies countability, Lowell imposed costs on the other for the delivery of police services, suggests that three. Such conflicts can be managed, of course, this fanfare is premature. Our study calls for a more but that requires compromises that must be ad- textured assessment. justed frequently, depending upon the presenting At the time of our observations, Compstat had problem. And a dynamic program of compromise been operating in one form or another in Lowell requires either: (1) continuing hands-on involve- for nearly five years. We were unable to observe ment by top management, or (2) middle manag-

58 Compstat and Organizational Change in the Lowell Police Department

ers who, with a good sense of the department’s flexibility, data-driven decision making, and in- mission, develop a shared set of “rules” about how novative problem solving. They received virtually to reach compromises that promote the desired no formal training in these areas before implemen- outcomes. In either case, a more balanced inte- tation, so they had to adapt and learn on the fly. gration of Compstat’s varied elements cannot be We should expect, then, that these three elements left to a static set of management practices that should prove most problematic for the depart- puts the Compstat process on “autopilot.” ment. If Compstat endures in Lowell, we might The underdevelopment of certain Compstat well expect these elements to be strengthened. elements may also be attributed to a second cause. Over time the captains should increase their skills Lowell seems to present a classic case of creating and inclinations to engage in these activities, and a program for which department personnel re- those who follow them may rise to those posi- ceived relatively little preparation. The prior ca- tions precisely because they have already achieved reer experiences of the sector captains did not mastery of such things as crime analysis and prob- prepare them well for developing organizational lem solving.

59

Compstat and Organizational Change in the Lowell Police Department

Appendices

Appendix I Lowell Police Department Organizational Chart

Appendix II Map of the City of Lowell and City Neighborhoods

Appendix III Patrol Officer Survey

Appendix IV Sample Compstat Prep Sheet

61

Compstat and Organizational Change in the Lowell Police Department

Appendix I

Lowell Police Department Organizational Chart

63

Compstat and Organizational Change in the Lowell Police Department

Division

Division

Division

Legal Division

Evidence Response

Special Investigation

Criminal Investigation

Dennis Cormier

Deputy Superintendent

Section

Section

Investigative & Prevention Bureau

Safety First

Officer Section

School Resource

Crime Prevention

Prevention Division

Section

Information

raining Section

T

Analysis Section

echnology Section

Intelligence Crime

T

Professional Standards

Superintendent

Edward F. Davis, III Edward F.

Administrative Division

Section

Section

Section

Detail Section

Communications

Budget & Finance

Lowell Police Department, 2000

Employee Assistance

Accreditation Section

Administrative Staff

Authority

Community

Division

Response Team

Lowell Housing

est Services Division

East Services Division

Community Response

W

Kenneth Lavallee

Operations Bureau

Deputy Superintendent

raffic Division

Detention Section

T

Headquarters Division

North Services Division

65

Compstat and Organizational Change in the Lowell Police Department

Appendix II

Map of the City of Lowell and City Neighborhoods

67

Compstat and Organizational Change in the Lowell Police Department

69

Compstat and Organizational Change in the Lowell Police Department

Appendix III

Patrol Officer Survey

71

Compstat and Organizational Change in the Lowell Police Department

Patrol Officer Survey The survey was distributed to patrol officers who attended roll call on the late night, day, and early night shifts (Platoons 1, 2, and 3 respectively). The officer in charge of the detail provided us with the department’s monthly work schedule. Patrol officers were assigned to one of three groups (Group 2, 4, or 6) and worked four days on and two days off. On any given day, one group was off-duty. Surveys were distributed on two separate days to reach officers from all three groups across all three shifts. The on-site researcher distributing the survey told officers that the study was being conducted by the Police Foundation to learn how Compstat worked at Lowell. He noted that input from patrol officers on police programs and procedures was not always solicited, but that the Police Foundation valued any feedback that they were willing to provide so that we might gather their insights into the Compstat process. Officers were advised that the survey was voluntary and responses would be anonymous, that the researchers would not attempt to identify individual respondents, and that completed surveys would be maintained in a secure space at the Police Foundation, accessible only to members of the research team. In Lowell there were 145 patrol officers. Based on our examination of the detail, we estimated that approximately twenty officers were absent from roll call on any given day. This left us with an approximate sample size of 124 officers from which we received 97 completed surveys, a 78 percent response rate.

73 Compstat and Organizational Change in the Lowell Police Department

Officer Survey Survey #L______The Police Foundation is conducting a study of Compstat. This research project is funded by the Department of Justice. As part of the study, we are conducting a survey of officers in your department. The survey is completely anonymous and you will not be identified in any report. The surveys will be kept under lock and key at the Police Foundation. The survey will take about 10 minutes to complete. We greatly appreciate your participation. Please answer as many questions as you can. You may refuse to any question or stop at any time. Again, thank you for your participation.

1. Please indicate whether each of the following is very important, somewhat important, or not at all important to the Compstat strategy in the Lowell Police Department.

Very Somewhat Not at all Don’t Important Important Important Know ❑ ❑ ❑ ❑ a. Reduce complaints against officers 1 2 3 8 ❑ ❑ ❑ ❑ b. Reduce violent crime in Lowell 1 2 3 8 ❑ ❑ ❑ ❑ c. Improve the quality of life in Lowell 1 2 3 8 d. Arrest people committing misdemeanor ❑ ❑ ❑ ❑ offenses 1 2 3 8 e. Hold sector captains accountable for ❑ ❑ ❑ ❑ crimes in their sector 1 2 3 8 ❑ ❑ ❑ ❑ f. Provide timely and accurate crime data 1 2 3 8 ❑ ❑ ❑ ❑ g. Respond quickly to calls for service 1 2 3 8 h. Identify crime patterns and choose ❑ ❑ ❑ ❑ appropriate tactics 1 2 3 8 i. Respond quickly to emerging crime ❑ ❑ ❑ ❑ problems 1 2 3 8 j. Hold officers accountable for crimes ❑ ❑ ❑ ❑ in their beats 1 2 3 8 k. Follow up to assess whether solutions ❑ ❑ ❑ ❑ were successful 1 2 3 8 l. Make officers and equipment available to ❑ ❑ ❑ ❑ different sectors as needed 1 2 3 8 m. Encourage officers to take responsibility ❑ ❑ ❑ ❑ for their beat 1 2 3 8 n. Resolve disputes among different segments ❑ ❑ ❑ ❑ of the community 1 2 3 8 o. Create and maintain open lines of ❑ ❑ ❑ ❑ communication with the community 1 2 3 8

74 Compstat and Organizational Change in the Lowell Police Department

2. In the last five years, has your department’s effectiveness in accomplishing each of the following increased, decreased, or stayed the same?

Stayed About the Don’t Increased Same Decreased Know ❑ ❑ ❑ ❑ a. Reduce complaints against officers 1 2 3 8 ❑ ❑ ❑ ❑ b. Reduce violent crime in Lowell 1 2 3 8 ❑ ❑ ❑ ❑ c. Improve the quality of life in Lowell 1 2 3 8 d. Arrest people committing misdemeanor ❑ ❑ ❑ ❑ offenses 1 2 3 8 e. Hold sector captains accountable for ❑ ❑ ❑ ❑ crimes in their sector 1 2 3 8 ❑ ❑ ❑ ❑ f. Provide timely and accurate crime data 1 2 3 8 ❑ ❑ ❑ ❑ g. Respond quickly to calls for service 1 2 3 8 h. Identify crime patterns and choose ❑ ❑ ❑ ❑ appropriate tactics 1 2 3 8 i. Respond quickly to emerging crime ❑ ❑ ❑ ❑ problems 1 2 3 8 j. Hold officers accountable for crimes ❑ ❑ ❑ ❑ in their beats 1 2 3 8 k. Follow up to assess whether solutions ❑ ❑ ❑ ❑ were successful 1 2 3 8 l. Make officers and equipment available to ❑ ❑ ❑ ❑ different sectors as needed 1 2 3 8 m. Encourage officers to take responsibility ❑ ❑ ❑ ❑ for their beat 1 2 3 8 n. Resolve disputes among different segments ❑ ❑ ❑ ❑ of the community 1 2 3 8 o. Create and maintain open lines of ❑ ❑ ❑ ❑ communication with the community 1 2 3 8

75 Compstat and Organizational Change in the Lowell Police Department

3. Listed below are 10 activities for which you might receive praise from your superiors. Please review all of these activities and then rank them from 1 to 10 in terms of how much praise you might receive. Write a 1 by the activity which would bring you the most praise, a 2 by the activity which would bring the second most praise, and so on.

Rank a. Identifying a developing crime pattern b. Rapid response to 911 calls c. Assisting crime victims in obtaining services and cooperating in prosecutions d. On-sight misdemeanor arrests of persons known to the police e. Working with community members to solve local crime or disorder problems f. Recovering a weapon g. Writing accurate crime reports h. Working to provide legitimate activities for potential offenders I. Making a large drug bust j. Overall reduction of crime in your beat

4. Have you attended a Compstat meeting? Please check one. ❑ ❑ 1 Yes 2 No [SKIP TO Q6]

5. If “Yes,” please indicate what you did at the meeting. Please check all that apply. ❑ 1 Observed meeting only—did not participate ❑ 2 Assisted someone else make a presentation but did not speak ❑ 3 Answered questions ❑ 4 Offered an opinion or information ❑ 5 Made a short presentation (less than 2 minutes) ❑ 6 Made a long presentation (2 or more minutes)

6. How often does your supervisor discuss what has happened at the Compstat meetings? Please check one. ❑ 1 Every week ❑ 2 About once a month ❑ 3 Every few months ❑ 4 Never

76 Compstat and Organizational Change in the Lowell Police Department

7. To the best of your knowledge, how much has Compstat changed your job responsibilities? By job responsibilities, we mean what you are expected to do on a daily basis. Please check one. ❑ 1 A great deal ❑ 2 Somewhat ❑ 3 A little bit ❑ 4 Not at all [SKIP TO Q9] ❑ 8 Don(t know [SKIP TO Q9]

8. Please describe any changes.

9. Please indicate your view of each of the following statements by marking the appropriate box.

Strongly Strongly No Agree Agree Disagree Disagree Opinion a. Compstat has made supervisors ❑ ❑ ❑ ❑ ❑ place too much emphasis on 1 2 3 4 5 statistics b. Compstat has made it possible for ❑ ❑ ❑ ❑ ❑ officers to get credit for doing 1 2 3 4 5 quality work c. Compstat has kept supervisors ❑ ❑ ❑ ❑ ❑ from spending enough time 1 2 3 4 5 on the street d. Compstat had made me more aware ❑ ❑ ❑ ❑ ❑ of what goes on in other parts of the 1 2 3 4 5 department e. Compstat has increased teamwork ❑ ❑ ❑ ❑ ❑ between my unit and specialist 1 2 3 4 5 units in the department f. Compstat will be an important ❑ ❑ ❑ ❑ ❑ feature of the department’s 1 2 3 4 5 organization in 5 years

10. Overall, how would you rate the impact of Compstat on the department’s performance in serving the public? Please check one. ❑ 1 Highly beneficial ❑ 2 Beneficial ❑ 3 No effect ❑ 4 Detrimental ❑ 5 Highly detrimental ❑ 8 Don’t know

77 Compstat and Organizational Change in the Lowell Police Department

11. Please briefly describe your view.

12. Overall, how would you rate the impact of Compstat on the department as a good place to work as a police officer? Please check one. ❑ 1 Highly beneficial ❑ 2 Beneficial ❑ 3 No effect ❑ 4 Detrimental ❑ 5 Highly detrimental ❑ 8 Don’t know

13. Please briefly describe your view.

14. How many years have you worked for the Lowell Police Department in a sworn position? ❑ 1 Fewer than 3 years ❑ 2 3–5 years ❑ 3 6–10 years ❑ 4 11–20 years ❑ 5 More than 20 years

15. What is your current rank? ❑ 1 Police officer ❑ 2 Rank higher than police officer

16. What hours do you work? from _____:_____ am/pm to _____:_____ am/pm

17. How old are you? ❑ ❑ 1 Under 21 5 40–49 ❑ ❑ 2 21–25 6 50–59 ❑ ❑ 3 26–29 7 60+ ❑ 4 30–39

18. What is your highest level of formal education? Please check one. ❑ ❑ 1 Some High School 5 Some Graduate School / Law School ❑ ❑ 2 High School Graduate or GED 6 Masters Degree / JD or LLB ❑ ❑ 3 Some College / AA Degree 7 PhD ❑ 4 Bachelors Degree

78 Compstat and Organizational Change in the Lowell Police Department

19. Are you male or female? ❑ 1 Male ❑ 2 Female

20. What is your racial group? Please check one. ❑ 1 White ❑ 2 African American ❑ 3 American Indian or Alaska Native ❑ 4 Asian American or Pacific Islander ❑ 5 Other

21. Are you Hispanic or Non-Hispanic? ❑ 1 Hispanic ❑ 2 Non-Hispanic

Thank you for completing the survey!

79

Compstat and Organizational Change in the Lowell Police Department

Appendix IV

Sample Compstat Prep Sheet

81

Compstat and Organizational Change in the Lowell Police Department

COMPSTAT PREP SHEETS January 11, 2001

ASSAULTS

NORTH EAST WEST CITYWIDE Total # 6141030 Arrests 08 311 Aggrav/Simple Agg-3 Agg-8 Agg-5 Agg-16 Simp-3 Simp-6 Simp-5 Simp-14 Day of Week Sat-4 Sun-7 Sun-4 Sun-11 Time/Platoon P1=3 (1-3) P1=6 P1=3 P1=12 P2=1 P2=4 P2=2 P2=7 P3=2 P3=4 (10-12) P3=5 P3=11 Liquor Related 1- Honkey Tonk 5- Brewery 2-C’est La Vie 8 Shamrock Worthen House Party Gang Related No No 1 1 Weapons Used Non-Lowell Victims= Victims= Victims= Victims= Residents Suspects= Suspects= Suspects= Suspects= Suspects Additional Info

83 Compstat and Organizational Change in the Lowell Police Department

BURGLARY

NORTH EAST WEST CITYWIDE Total # 73 919 Comm/Dwell Comm-2 Dwell-3 Dwell-9 Comm-2 Dwell-5 Dwell-17 Day of Week Sat-4 Sat-5 Time/Platoon P1=2 P1= P1=2 P1=4 P2=1 P2=1 P2=2 P2=4 P3=2 P3=1 P3=3 P3=6 Method of Entry Forced Open Forced Open Forced Open Forced Open Property Suspects Additional Info 2 Arrests 1 Arrest 3 Arrests

DRUG ACTIVITY

NORTH EAST WEST CITYWIDE Total # 44 210 Arrests 32 27 SIS Involved 112 Neighborhood Centra-2 S.Lowell-2 Acre-2 Belv-2 Down-2 Day of Week Sat-3 Sat-4 Time/Platoon P1= P1= P1= P1= P2= P2= P2= P2= P3= P3= P3= P3= Drug Type Cocaine Cocaine, Heroin Cocaine, Heroin Marijuana, Heroin Confiscated 1 Bag 25 MG Pills 1 Needle 1 Bag Cocaine 2 Needles 39 Batman Bags 9 Bags 48 Bags Heroin 3 Needles 25 MG Pills Suspects Additional Info

84 Compstat and Organizational Change in the Lowell Police Department

MV B&E

NORTH EAST WEST CITYWIDE Total # 7111230 Day of Week Wed-3 Mon-4 Mon-5 Fri-5 Time/Platoon P1=2 P1=4 (9-11)P1=5 P1=11 P2=1 P2=2 P2=1 P2=4 P3=1 P3=3 P3=4 P3=8 Neighborhood Belv-4 Down-7 High-10 High-10 Centra-2 MV Make Honda-3 Honda-3 Honda-7 Point of Entry Window-3 Window-7 Window-5 Window-15 Property Taken CD Stereo-2 Stereo-3 Stereo-7 Suspects Additional Info

MV DAMAGE

NORTH EAST WEST CITYWIDE Total # 23 11 8 42 Day of Week Sun-8 Mon-3 Sun-10 Fri-7 Time/Platoon P1=4 P1=2 P1=3 P1=9 P2=3 P2=2 P2= P2=5 P3=9 P3=1 P3=4 P3=14 Neighborhood Centra-10 Bck Cent-3 High-5 Centra-10 Belv-6 MV Make Dodge-2 Acura-3 Honda-5 Toyota-2 Dodge-5 Toyota-5 Damage Smshd Wind-9 Smshd Wind-5 Smshd-Wind-3 Smshd Wind-17 Slshd Tire-7 Suspects Additional Info

85 Compstat and Organizational Change in the Lowell Police Department

86 Compstat and Organizational Change in the Lowell Police Department

References Bratton. Flushing, NY: Looseleaf Law Pub- lications. Anderson, David C. 2001. Crime Control by the Heskett, James L. 1996. NYPD New. Harvard Busi- Numbers. Ford Foundation Report (Winter ness School Report no. N9–396–293 (April). 2001): 5. Iven, Chris. 1999. Crime in Lowell continues to Bayley, David H. 1988. Community Policing: A plummet. Lowell Sun, October 22. Report from the Devil’s Advocate. In Com- Kelling, George L., and Mark H. Moore. 1988. munity Policing: Rhetoric or Reality, ed. Jack R. From Policing to Reform to Community: The Greene and Stephen D. Mastrofski, 225–237. Evolving Strategy of Police. In Community Po- New York: Praeger. licing: Rhetoric or Reality, ed. Jack R. Greene Bluestone, Barry, and Mary Huff Stevenson. 2000. and Stephen D. Mastrofski, 3–25. New York: The Boston Renaissance: Race, Space, and Eco- Praeger. nomic Change in an American Metropolis. New Kelling, George L., and William H. Sousa. 2001. York: Russell Sage. Do Police Matter? An Analysis of the Impact Bouza, Tony. 1997. NYPD Blues—Good, Lucky, of New York City’s Police Reforms. Civic Re- or Both? Law Enforcement News. (January 31): port no. 22 (December 2001): Center for Civic 8, 10. Innovation at the Manhattan Institute, http:// Bratton, William J. 2001. A cloudy future for po- www.manhattaninstitute.org/cr_22.pdf. licing. New York Times, August 3. Lehrer, Eli. 2001. The Police Behind America’s Bratton, William J. 1998. Tu rnaround: How Amer- Biggest Crime Drop. American Enterprise ica’s Top Cop Reversed the Crime Epidemic. With (March), http://www.findarticles.com/cf_dls/ Peter Knobler. New York: Random House. m2185/2_12/71004740/p1/article.jhtml. Crank, J. P. 1989. Civilianization in Small and Lowell Sun. 2000. Where’s the good news? Decem- Medium Police Departments in Illinois, 1973– ber 11. 86. Justice Quarterly 8: 170–185. Lowell Sun Online. 2002. Lowell won’t tolerate Eck, John E., and Edward R. Maguire. 2000. Have rising crime rate. April 4, http://www.lowellsun Changes in Policing Reduced Violent Crime? .com/. An Assessment of the Evidence. In The Crime Lyons, William. 1999. The Politics of Community Drop in America, ed. Alfred Blumstein and Joel Policing: Rearranging the Power to Punish. Ann Wallman, 207–265. Cambridge: Cambridge Arbor: University of Michigan Press. University Press. Manning, Peter. 2001. Theorizing Policing: The Goldstein, Herman. 1979. Improving Policing: A Drama and Myth of Crime Control in the Problem-Oriented Approach. Crime and Delin- NYPD. Theoretical Criminology 5 (3): 315–344. quency 25: 236–238. Maple, Jack. 1999. The Crime Fighter: Putting the Goldstein, Herman. 1990. Problem-Oriented Polic- Bad Guys Out of Business. With Chris Mitchell. ing. New York: McGraw-Hill. New York: Doubleday. Gurwitt, Rob. 1998. The Comeback of the Cops. Mastrofski, Stephen D. 2002. The Romance of Governing ( January): 14–19. Police Leadership. In Theoretical Advances in Guyot, Dorothy. 1979. Bending Granite: Attempts Criminology: Crime & Social Organization, ed. to Change the Rank Structure of American David Weisburd, and Lawrence W. Sherman, Police Departments. Journal of Police Science 153–196. New Brunswick: Transaction Pub- and Administration 7: 253–84. lishers. Harcourt, Bernard. 2002. Policing Disorder: Can Mastrofski, Stephen D., and R. Richard Ritti. 2000. We Reduce Serious Crime by Punishing Petty Making Sense of Community Policing: A Offenses? The Boston Review (April/May), http:// Theory-Based Analysis. Police Practice 12: 183– bostonreview.mit.edu/BR27.2/harcourt.html. 210. Henry, Vincent E. 2002. The COMPSTAT Paradigm: Mastrofski, Stephen D., and R. Wadman. 1991. Management Accountability in Policing, Business Personnel and Agency Performance Appraisal. and the Public Sector. Foreword by William J. In Local Government Police Management, ed.

87 Compstat and Organizational Change in the Lowell Police Department

W.A. Geller, 363–397. Washington, DC: Inter- Government, Program in Criminal Justice national City Management Association. Policy and Management, http://www.ncjrs.org/ Mastrofski, Stephen D., R. Ritti, and D. Hoffmaster. nij/cops_casestudy/lowell4.html. 1987. Organizational Determinants of Police U.S. Census Bureau. 2002. United States Census Discretion: The Case of Drinking and Driving. 2000, http://www.census.gov/main/www/ Journal of Criminal Justice 15 (5): 387–402. cen2000.html. McDonald, Phyllis, Sheldon Greenberg, and Wil- U.S. Department of Justice, Federal Bureau of In- liam J. Bratton. 2001. Managing Police Opera- vestigation. 2001. Crime in the United States tions: Implementing the NYPD Crime Control 2000. Washington, DC: U.S. Government Model Using COMPSTAT. Belmont, CA: Wads- Printing Office. worth Press. U.S. Department of Justice. 1994. Violent Crime Parks, R. B., Stephen D. Mastrofski, A. J. Reiss, Jr., Control and Law Enforcement Act of 1994, R. E. Worden, W. C. Terrill, C. DeJong, M. U.S. Department of Justice Fact Sheet, http:// Stroshine, and R. Shepard. 1998. St. Petersburg www.ncjrs.org/txtfiles/billfs.txt. Project on Policing Neighborhoods: A Study of U.S. National Agricultural Library 1998. Comp- the Police and the Community. Report to the stat: A Crime Reduction Management Tool, National Institute of Justice. Bloomington, IN: http://www.nalusda.gov/pavnet/iag/cecompst Indiana University. .htm. Reiss, Albert J., Jr. 1992. Police Organization in Walsh, William. 2001. Compstat: An Analysis of the Twentieth Century. In Modern Policing, ed. an Emerging Police Paradigm. Policing: An In- Michael Tonry and Norval Morris, 11–97. Chi- ternational Journal of Police Management and cago: University of Chicago Press. Strategies 24 (3): 347–362. Scott, Christopher. 2002. Councilors’ ire rises with Weisburd, David, and Anthony Braga. 2003. Hot crime. Lowell Sun Online, April 10, http://www Spots Policing. In Crime Prevention: New Ap- .lowellsun.com/. proaches, ed. Helmut Kury and Joachim Sherman, Lawrence W., and David Weisburd. Obergfell-Fuchs, 337–355. Mainz, Germany: 1995. General Deterrent Effects of Police Pa- Weisser-Ring. trol in Crime ‘Hot Spots’: A Randomized Con- Weisburd, David, Stephen Mastrofski, Ann Marie trolled Trial. Justice Quarterly 12: 625–648. McNally, Rosann Greenspan, and James Willis. Sherman, Lawrence W. 1998. Evidence-Based Po- 2003. Reforming to Preserve: Compstat and licing. Ideas in American Policing. The Police Strategic Problem Solving in American Polic- Foundation: Washington, DC, http://www ing. Criminology and Public Policy 2 (3): 421– .policefoundation.org/pdf/Sherman.pdf. 456. Silverman, Eli B. 1999. NYPD Battles Crime: Inno- Weisburd, David, Stephen D. Mastrofski, Ann vative Strategies in Policing. Boston, MA: North- Marie McNally, and Rosann Greenspan. 2001. eastern University Press. Compstat and Organizational Change: Find- Simons, Robert. 1995. Control in an Age of Em- ings from a National Survey. Report submitted powerment. Harvard Business Review 72 (2): to the National Institute of Justice by the Po- 80–88. lice Foundation. Skruck, Jeff. 2002. Lowell chief wins award from Willis, James, Stephen D. Mastrofski, David Weis- police association. Lowell Sun, May 13. burd, and Rosann Greenspan. Forth. Compstat Sparrow, M., M. Moore, and D. Kennedy. 1990. and Organizational Change: Intensive Site Visits Beyond 911: A New Era for Policing. New York: Report. Basic Books. Wilson, James Q., and George L. Kelling. 1982. Thacher, David. 1998. National COPS Evaluation Broken Windows: The Police and Neighbor- Organizational Change Case Study: Lowell, hood Safety. Atlantic Monthly (March): 29–38. Massachusetts. Unpublished Manuscript, Har- Wilson O.W., and R. C. McLaren. 1972. Police Ad- vard University, John F. Kennedy School of ministration. 3d ed. New York: McGraw-Hill.

88 Compstat and Organizational Change in the Lowell Police Department

Authors James J. Willis is Assistant Professor of Sociology at the University of Massachusetts-Boston. His research focuses on the impact of reforms and technological innovations on police organization and practice. His other research area is punishment and penal history.

Stephen D. Mastrofski is Professor of Public and International Affairs at George Mason University and Director of the Administration of Justice Program and the Center for Justice Leadership and Management. His research focuses on police and police organizations.

David Weisburd is Professor of Criminology at the Hebrew University Law School in Jerusalem, and Professor of Criminology and Criminal Justice at the University of Maryland, College Park. He is also a Senior Fellow at the Police Foundation and Chair of its Research Advisory Committee.

Rosann Greenspan is Associate Director, Center for the Study of Law and Society at the University of California at Berkeley and Lecturer in Residence, Boalt Hall School of Law. She was research director at the Police Foundation from 1997–2000.

89