Become a Patron


Brandon Garrett

excerpted from: Brandon Garrett, Remedying Racial Profiling, 33 Columbia Human Rights Law Review 41-148, 107-140 (Fall 2001)(343 Footnotes omitted)

One police chief noted: "Until they define [racial profiling], we can't really discuss it . . . . It means too many things to too many people." Indeed, no social or legal consensus exists on what degree of disparity in police treatment is tolerable and under what conditions. Defining racial profiling has befuddled courts and legislators dealing with the question of what to do if disparity is uncovered. The Justice Department consent decrees suggest a way of sidestepping this problem: they focus on eliminating disparity by making the policing process more responsive and informed. However, the decrees only use information supplied by officers during traffic stops and only provide for a few years of monitoring by an independent monitor or state attorney general's office. Further, as noted above, neither Justice Department consent decrees, state legislation, nor remedies pursuant to private litigation require that racial profiling data be made public. Additionally, all three of the mechanisms discussed above fail in significant ways to include outside actors in the decision-making process. If police departments can be made more responsive, the question is: Responsive to whom?

Police may have embraced collecting detailed data about stops because they have had complete control over the circumstances of the collection and interpretation of the results. Maybe advocacy groups have not questioned data collection as a remedy because they believe that lawsuits and consent decrees allow them to retain some control over interpreting data and to ensure that police efforts are not a sham. In order for lasting change to result from these lawsuits and consent decrees, permanent institutional monitors need to be in place when they end.

Racial profiling remedies have left key constituencies without a role in the definition of policing norms. This is true, above all, of the community groups that acted as a critical catalyst in calling for an end to racial profiling. If these groups are shut out of the process, they will demand a role in defining remedies--by litigation, as in Los Angeles, or by establishing, via referendum, a civilian board to supplement the consent decree, as in Pittsburgh. Any approach attempting to redress discriminatory policing should employ outside expertise and permit key outside actors to play a role in defining the remedy.

In explaining a more comprehensive approach to racial profiling, this Part discusses two examples of participatory reform, showing how police departments can begin to implement a more effective remedy. It will then analyze what makes participatory police reform work, examining three key elements of successful remedies: 1) information gathering; 2) institutional problem solving based on reflection upon, and assimilation of, collected data; and 3) partnership with outside participants.

A. Participatory Reform: A Look at Two Police Departments

After making information public, police must take more steps to include outside groups in the long-term project to solve police problems. Several police departments have moved towards a model of interpreting data and problem solving that stresses collaboration with experts, researchers, service- providers, and community groups. These projects show how police and the community can sustain a long-term commitment to work together and how remedies can be institutionalized. They also indicate a trend in the evolution of racial profiling remedies.

One recent project in Philadelphia demonstrates how community groups can rapidly become involved in interpreting data, designing how it is collected, and contributing to police department efforts to respond to problems and identify new ones. In 1999, the Philadelphia Inquirer uncovered a twenty-year practice by which the city police were "dump[ing]" reported sex crimes. Police either failed to investigate crimes at all, investigated them poorly, or downgraded their seriousness by reporting them as lesser or non-criminal offenses. Reports indicated a pervasive attitude of blaming victims and an unwillingness to find their stories credible and worth pursuing. City Council hearings were held, and the new police commissioner, John Timoney, assigned forty-five detectives to look at over two thousand closed sexual assault cases, including hundreds of rapes, that dated back to 1995. In the midst of this unfolding scandal, the Commissioner did something dramatic--he invited women's and domestic violence groups in Philadelphia to form a special task force with police to review all relevant police files and help to decide if cases should be reopened. The police courageously admitted that they could not solve a problem without outside expertise and the credibility which those community groups would bring. This was a first-- permitting outside groups to review police records without litigation. The groups involved were the Women's Law Project, Women Organized Against Rape, the University of Pennsylvania Women's Center, and the Pennsylvania National Organization for Women (NOW) Chapter. As a condition to their participation, these groups agreed that they would review only closed cases, take no notes, and keep all information confidential. They did so, as it allowed them to continue assisting people who needed advice or advocacy on open cases while preserving police investigations. The review process most often ended in agreement, with a few cases provoking dispute over police handling of the case. The project quickly led to joint problem-solving and reinterpretation of how data is collected, as well as the incorporation of new community groups--the sorts of remedies needed to address racial profiling. Using data to ask questions about the source of original problems, advocates and the police soon found that more information and expertise were needed to address broader underlying problems. For example, they found that part of the reason that sexual assaults were characterized as lesser offenses was that Philadelphia's police code manual had either vague definitions of sexual assault or definitions inconsistent with Pennsylvania law. They began to study how codes are written and began a project of rewriting the police code manual. They also found that files were often vague or incomplete, and suggested the need for training on interviewing techniques. Furthermore, they saw that supervision in the sex crimes unit was lacking. In response, they sought data from the police department on how it responded to sex crimes. Interestingly, they asked the police department to have the data broken down in ways useful to the outside groups--by sex, race, and zip code. They hoped to see if race or class affected disposition of complaints, although given the scope of the crisis, the problems may have been systemic. Finally, they discovered that many of these mishandled cases involved children, and lacking expertise in the area, they brought in two children's advocacy groups: the Support Center for Child Advocates and the Philadelphia Children's Alliance. What helped this collaboration come together so quickly? The police were willing to make the initial overture to grant access to outsiders. The existence of a crisis may have helped, though police departments can just as easily close ranks when there is a scandal or crisis. The Philadelphia police have also been willing, so far at least, to enter into a partnership extending beyond the immediate problem of reviewing case files, to allow an exchange of advice and information, and actually to cooperate on problem solving. Philadelphia also has a particularly cohesive group of public interest organizations that were ready to address the issue and made a show of good faith by being willing to take the risk of foregoing litigation in favor of collaborating. The existence of cohesive and able institutions with strong ties to the community must have helped make the police commissioner's decision much easier. Other institutions also played an important role: the city council insisted on change and held hearings, and the Philadelphia Inquirer broke the initial story and then maintained continuing press coverage.

Successful problem solving may require the right kinds of institutions, not just in the police department, but in the community as well. One fear is that partnerships cannot occur in places where, unlike in Philadelphia, community groups do not exist, are at odds with each other, or are antagonistic towards the police or other parts of the community. Such divided communities may find it especially hard to accept the legitimacy of lawsuits or decisions over which they have had no input. However, just bringing different groups together to work on a problem may also foster institutions that can cooperate with police and represent the community. Conversely, institutions can be created to encourage community-police cooperation. For example, civilian complaint review boards have been created as oversight bodies with varying degrees of success. While most are organized as internal boards, or as external monitors, some have been more independent and have collaborated with police. For example, in Phoenix, citizens have been invited to sit on use-of-force review boards as well as hiring and promotion boards, to provide input, and to inform the public. Another concern is that public interest and legal advocacy groups may not adequately represent the community; to some degree, they are elite and removed. However, these groups have stronger ties to the community than the impact groups that now control racial profiling litigation. To some extent, the process of litigation itself creates a hierarchical relationship that distances the Justice Department and groups like the ACLU from the community. In control of litigation, disclosing negotiations only to the court at the time of settlement or decision, these groups have only periodic informational meetings with the community. In contrast, collaboration permits a greater role for the community by giving people the opportunity to participate in an ongoing process of problem-solving--one in which they receive constant information about the progress of reform efforts.

The strongest partnerships have resulted in two-way communication between the police and community groups where police share data with community groups and invited their input. While community and advocacy groups have not traditionally shared responsibility for data collection and interpretation, this process can be collaboratively designed and implemented. Police departments can bring together representatives from the department, the community, and expert and advocacy groups to discuss ways to design data collection as well as the means to implement and improve these measures. For example, Rhode Island took this approach when it convened an interdisciplinary group of community leaders and experts to decide how to interpret statewide data on race. Along these lines, a Chicago project in the late 1980s involved both the police and the community in a collaborative data gathering and problem-solving effort to map crime in neighborhoods. Like collecting data on complaints and race, crime mapping has normally been exclusively in the hands of police. However, residents often have very different ideas of where trouble spots exist in their neighborhoods. With better computer technology, mapping was for the first time providing a powerful tool for community members to visualize crime patterns and to evaluate police response. Community groups in Chicago had to overcome a history of strong police resistance to providing community groups with access to police information on neighborhood crime. In a pilot district, police agreed to supply community groups with daily crime reports that were already kept in a database. As in Philadelphia, there were organized community groups that worked well together--the Chicago Alliance for Neighborhood Safety (CANS) and a consortium of local neighborhood watch and crime-prevention groups. The consortium had already cooperated with police, receiving some crime reports, and had worked with six local neighborhood groups that ran block meetings and crime watch programs. CANS analyzed the crime data, prepared the basic neighborhood maps, performed daily data entry, linked the information to the maps, and evaluated and distributed the maps to the six other neighborhood groups. They also aspired to gather data on "incivilities," or other disturbing non-criminal incidents reported by community members, which would then be mapped together with crime data and discussed in a joint meeting. Unfortunately, this did not occur. One of the great successes of the project was learning from the two-way exchange of information. Community input provided crime maps that showed trends and compared data in ways the police had never thought of before, pointing to a more comprehensive approach:

In one instance, the police thought that residents' fear of crime had overcome their objectivity and that their alleging that one busy corner was especially dangerous (particularly on weekends) was a clear error. But long-term data for the area, which community analysts were able to provide, and calls-for-service data, which were added to the mapped analysis, verified the accuracy of the residents' perceptions; relying on crime statistics alone gave a misleading picture of the corner's problems. This experience resulted in a vigorous response by the district commander, and each party's respect for the other was solidified. Identification of hot spots, in meetings that presented crime maps and gathered data to make a case to the police, resulted in successful crime prevention. Both the police and the community found they had a great deal to learn from each other. The researchers emphasized that any information system must be designed in collaboration with its users. Institutionalizing information sharing can occur with varying degrees of cooperation. The Chicago pilot project allowed community groups to prepare maps, to make information accessible, and to use their own knowledge and maps to solicit police response to their needs. Today, with its community policing approach, the Chicago police still create maps of crime using the techniques learned in the pilot project, and they bring them to community beat meetings as a basis for discussion. Making information public seems much easier when the information is in such a useful form.

However, the community has lost its stake in the process, its ownership in producing the maps, and its ability to decide how the maps would be organized, when they will be displayed to police, and in what cases maps are inadequate. Making the maps public and explaining them is helpful, but without giving the community an equal stake in interpreting the maps, there is no longer the same kind of problem solving. Similarly, the project in Philadelphia does not yet involve neighborhood groups, nor is collaboration on interpreting data formalized yet; hopefully the police will allow the external review to continue.

B. Creating Problem Oriented Partnerships

The examples provided by Chicago and Philadelphia suggest how police can institutionalize a problem-oriented approach that incorporates information gathering and analysis. Neither deals specifically with racial profiling, but the same kind of approach can apply in any problem solving effort. The prior section argues that this approach is particularly suited to dealing with problems of race in policing. Three features stand out: 1) police departments took the initial step of setting up a system of information gathering and making public the results; 2) they began an ongoing process of reflection, questioning, and problem solving based on that information; 3) they opened participation to outside groups that assisted in problem solving, lent support to the effort, helped rearticulate the process of information gathering, and even raised new problems. 1. Information Gathering

"Information is the lifeblood of the police." Police departments increasingly rely on information systems to guide their policing. New York led the way early in the 1990s with its Compstat system, used to monitor crime and pinpoint areas where more resources should be deployed. Police departments can use the same problem- solving and information systems to improve their services and to monitor themselves. Racial profiling, itself defined through statistics, has led to a new expansion of police information gathering. Laws passed and remedies in consent decrees have encouraged police to take advantage of technology in order to modernize their approach to routine traffic stops. One example of the revolutionary possibilities information technology can provide to communities is a website maintained by the state of North Carolina. The website allows one to select a police department and view statistics each month, broken down by race, sex, age and ethnicity, initial reason for stop, basis for search, and what enforcement action was taken. The most critical aspect of the data collection is that its results be disclosed. This should be the goal of any initial remedy. Disclosing data from early warning systems would create a dramatic change in community access to policing information. It could give communities access for the first time to in-depth information about how police work in their neighborhoods. Thus, residents would find it easier to show that a department is deliberately ignoring the abuses of 'problem' officers.

More importantly, making information public allows police to announce their policies and goals, and build the trust of the public by allowing them to assess the success of a policy. Thus, public information helps police accomplish key law enforcement goals. "[T]ransparency requires not only visibility of policy choices but a publicly declared rationale for these decisions. Requiring government to announce both positive action and normative justification acknowledges the human desire to assess the bona fides of public officials." Transparency of government action also provides stronger democratic legitimacy. "[T]ransparency is a prerequisite of legitimate, democratic government. Nowhere is this mandate more important--in terms of rights, interests, and costs--than in the criminal justice system." Despite the benefits, police have strongly resisted sunlight, so much so that they have even resisted sharing information with other law enforcement organizations. Police have historically resisted any attempt at civilian review and public access. This distrust of the public leads to insularity. Further, a police culture of silence can then take on an institutional life of its own. Thus, although police departments increasingly applaud data collection requirements, they adopt tepid forms of data collection. If they eventually acquiesce to the consent decrees and state laws, it is only because the information is kept out of the public eye.

Police may fear liability from data, though such concerns seem unwarranted. After all, if plaintiffs decide to sue, they can obtain data through discovery even if none has yet been collected. Racial profiling lawsuits have only required data collection as a remedy. Any individual damages are premised on the harm caused by use of force and unreasonable searches; no damages have been awarded for the purely stigmatic harm caused by being part of a statistical pattern of disparate stops. Concerns about liability or access to privileged information could be easily avoided by having a monitor decide which data could safely be made public, perhaps aggregating data by neighborhood without names of officers. The New Jersey Decree and the Rhode Island state law both require release of aggregate information. Police also justifiably fear releasing data that might be misunderstood or seized upon by opponents. They worry that if information is made public, advocates will exaggerate this data to make it appear that they are 'racists.' Given that such data is often released into such highly charged situations, it stands to reason that its interpretation will be in dispute. On the one hand, police departments are very quick to declare that the data proves that they do not engage in racial profiling, despite deep-seated community feelings to the contrary. On the other hand, advocates often latch on to incomplete data given their limited access.

In order to defuse this conflict, departments should release enough information to be meaningful to communities, broken down by neighborhood, and compared with relevant demographic information. Police can release the information in different ways depending on the audience. Aggregate information may be of interest to groups addressing systemic problems, while more specific information will prove critical to individuals, victims, and their advocates. Public information will make police information gathering less suspect, but only if the information released is comprehensive.

Collecting highly personal information during stops also presents a civil liberties threat that justifies public scrutiny of the information. While police departments may resent requirements that they collect data, these new databases give police access to an unprecedented amount of information about the people with whom they come into contact. Two of the consent decrees permit officers to ask for personal information even though these requests may serve to intimidate people and may be of dubious value in reforming policing. The New Jersey Decree asks that police collect license plate numbers and dates of birth of those stopped, and that they videotape or tape record the stops. The Steubenville Decree mandates collection of the names and contact information for those stopped, as well as for witnesses. Personal information itself should not be made public. For example, the names of officers or of people stopped should be withheld, with detailed figures from databases being made public. The hope of the decrees is that increased police accountability will minimize opportunities for abuse. However, the threat of misuse of so much personal information justifies some accountability to the public, especially where it originally sought this information to prevent, and not further compound, police abuse.

One way to disclose very specific information, without compromising the police or public, is to ensure that officers provide badge numbers when they make stops, so that citizens can follow up on complaints. The New Jersey Decree provides that, "The State Police shall require all state troopers to provide their name and identification number to any civilian who requests it." Making badge numbers clearly visible also reassures people during a stop that they have some recourse. Even better, the new Colorado law requires police to give those stopped business cards to explain how to file a complaint. Recording or videotaping stops also provides a more comprehensive public record. Ultimately, the effectiveness of this information depends on whether a real review follows a civilian complaint, and whether police also publicize how to make a complaint. Not only will data transform the relationship between police and the public, it will transform internal relationships within the department. The examples above show how information systems can help organizations abandon habits and traditional assumptions in order to become more effective. Information makes general police policies and the specific actions of officers more transparent to the community and to problem-solvers within the department. Moreover, information can spearhead a new, more open approach to organizational change; Malcolm Sparrow describes a "new philosophy of information management" where agencies broadly share data, embrace public access, and encourage cross- analysis and integration of different kinds of data.In designing information gathering systems, departments should begin with a comprehensive approach. Consent decrees and lawsuits have sought to broaden the data set, including information about the use of force, neighborhood, and time of day, in order to capture the quality of the stop and not just its rationale or racial focus. Unlike those mandated by state legislation, questionnaires must avoid becoming simplistic litmus tests. On the other hand, if they are too complicated, police, feeling burdened, will be less likely to fill them out. As discussed earlier, data must also be made public so that the process is transparent. The data should not just be released in the aggregate, but also by neighborhood, so that communities know how policing affects them specifically. Further, information must be organized in an accessible database, so that data can be assembled or analyzed easily. Finally, and critical to the legitimacy of the project, the data must be compared against control groups, such as driving populations or demographic samplings of neighborhood street traffic. Internal comparisons, against data from other patrol units, should also be done. The way that police gather information will also be critical. First, data collection design efforts will necessarily transform how stops, searches, and arrests are conducted. Second, data collection itself sends a strong message of concern and accountability to the community, indicating that racial disparity is being watched and is not tolerated. For example, data collection can send a negative message if police gather information in an intrusive way, leading to perceptions of racial profiling or unfair singling-out. Citizens also may be suspicious of police asking too many personal questions, such as where they live. Such concerns may themselves be diffused by providing an explanation for the stop at the time it occurs, or by apologizing for its inconvenience. Further, if police fill out a stop questionnaire in front of the citizen, it may diminish any outrage over the encounter. The citizen knows that the stop is documented, that its rationale is being recorded, and that she can check the identity of the officer. However, if the form is filled out after the fact without any questions being asked, then the citizen may be less likely to trust in the reporting.

Alternatively, perhaps citizens should be asked to enter information about stops into a database after the fact, so that officers are not the only ones recording the data. If police rarely fill out forms, or citizens never witness officers performing this duty, then any statistics will be suspect. All of these decisions reflect strongly on how the citizen and officer interact, and all depend on the values and concerns of the police department and the community. Any protocol or policy for conducting stops, searches, and arrests should come to reflect the decisions that were made in the data collection design process. Police perceptions and concerns will also prove important in deciding how to collect information. Police are trained as expert observers and should be involved in the process of designing a system that takes maximum advantage of their skills. Sparrow emphasizes the importance of tapping into the 'local knowledge' of officers on the beat. Information collection systems must give officers a convenient way to submit information, but also an opportunity to raise problems and concerns. For example, consent decrees have allocated money to provide officers with personal digital assistants to reduce paperwork and make the process easier than the old system. By doing so, this policy prevents police from viewing data collection as an imposition that they are tempted to avoid. On the other hand, if a form is too easy to fill out, police may check off boxes rather than provide a nuanced description of the stop. Police should also be given regular opportunities to raise problems that cannot be expressed on a form, or to offer suggestions that would prevent repeat problems on their beats.

Standards for effective information gathering will evolve, as police learn from their mistakes. The Department of Justice and state governments should help by funding research to evaluate and compare information gathering, thereby assisting police in improving their efforts. Police departments will also have to recruit new analysts, expert in interpreting data and in asking questions about what new kinds of information would be helpful. 2. Institutionalizing Problem Solving

The great hope of the consent decrees design does not lie in setting up a computer system, but instead that by doing so, police will develop expertise in using information to question their practices and engage in ongoing problem solving. Commentators have termed this approach 'problem-oriented' policing. Under this approach, police become skeptics and form the habit of asking why problems occur, what larger social conditions contribute to the problem, and who can help assist in solving the problem. Police identify systemic issues and become willing to make systemic changes. Along these lines, the Justice Department's hope is that the information structure of the consent decrees will help to create a cadre of supervisors that will maintain capacity for reflection.

Creating a cadre of problem solvers poses unique issues in a traditionally hierarchical institution not used to tolerating, much less fostering, internal debate and questioning. For example, a traditional response to the racial profiling problem might be to merely issue a police regulation requiring or suggesting that officers follow the Fourth Amendment and not rely predominantly on race. This would not be a reflective solution, though lawsuits and consent decrees have begun by asking for such a policy as a starting point to begin problem solving. The process must proceed by asking profound questions about why officers rely on race, under what circumstances they do so, what incentives help shape their behavior, how to train them better, what law enforcement goals they are trying to accomplish by making stops the way they do, and whether the goals are themselves appropriate.

Police departments, paralyzed by accusations of racial profiling, often manifest organizational inflexibility or inability to respond to problems in order to change traditional ways of policing. They also possess insularity, an inability to reach out to others who might be able to assist them in understanding the problem. Admitting mistakes can be difficult for police; "The police culture makes it tough to be honest about operational failures. The traditional belief is that failures can only result from officers ignoring or forgetting the correct procedures." than blaming officers, the procedures themselves may need to be reconsidered. Thus, problem solving will require sustained efforts to change the organizational sub-culture and break down rigid attitudes, such as the 'blue wall of silence.' This problem solving process must be ongoing, and will require a change in attitude on the part of police.

Rather than teach officers to follow procedures set in stone, police must be encouraged to experiment with policy change and to respond rapidly to new information. While many police departments are beginning to do this with new technology, this work is usually limited to supervisors. Police officers are not encouraged to be creative; they are taught to follow rules and orders, not to think about them. Individual officers should be encouraged to use their skills and knowledge to address problems and should be rewarded for doing so. In part, innovation and change may itself begin to unsettle hierarchy and traditional ways of making decisions; police will be forced to confront and question data that is collected, which may help lead to more long term strategic thinking. As the crime-fighting advantages of flexible, rapid response become apparent, these reforms may take hold more quickly.

Instead of taking orders, police must become used to open-forum debate and deliberation. Police use of standing committees or project-based work teams will, in effect, flatten the police hierarchy and lead to more flexible definitions of people's jobs, so that officers can question supervisors. As one commentator noted, "[R]igid supervisory control should give way to a more respectful, more collegial, more participative style of management." These kinds of changes may engender resentment and fierce opposition from police unions in particular, as reforms may undermine hard-won work rules by threatening the seniority system, offering rewards for creativity and problem solving, requiring flexibility and ambiguity in job descriptions, and broadening the police mission to include community needs. Thus, any reform effort along these lines must involve unions in the planning process and make clear that officers will be given more responsibility, at the same time that they will be able to develop their skills and contribute more than ever before. If departments learn to institutionalize responsiveness to the information that they gather, they will prove more able to adapt to problems like racial profiling. Measuring the effectiveness of policies is not easy and requires asking questions about what the goals of policing should be. By focusing on effectiveness, police will be able to assess policies and ask whether discretionary stops lead to sufficient arrests, seizures, or other benefits to the community so as to justify the harm that such discretion may cause. Indeed, evidence increasingly shows that police should consider alternatives to deterrence-oriented policing--in which officers engage in large numbers of pretextual stops--because, rather than leading to less crime, or more arrests or seizures, these methods instead lead to shoddy police work that requires prosecutors and judges to dismiss more cases. The information gathering techniques described suggest a new model for problem-oriented policing. While problem-oriented police departments have engaged in some of these practices, they have only engaged community groups informally and have shared limited information when they desired to do so. Even though community policing has become a mantra adopted by most police departments, few actively engage the community, and even when they do, community policing has generally occurred on the police's terms and without full disclosure of information to participants. Creating new information systems allows outsiders to help evaluate the success of police work in a systematic way. Combining information sharing and pooling with problem-oriented policing, as in the examples described, can itself institutionalize problem solving. Future remedies should move beyond both the consent decrees and old models for problem-oriented policing by combining problem solving with information gathering. They should create incentives for problem-oriented policing by encouraging supervisors to spend time evaluating information and questioning practices in a regular forum. This kind of change is feasible. Something as simple as weekly meetings to discuss data and suggest responses could suffice. Establishing a standing commission to give this body some permanence and an institutional place within the police department would also help. Police also could be given the message that attending the meetings and making helpful suggestions on how to address problems suggested by the data will be considered for promotions. Openness of leadership and continuing support will be critical to maintain problem solving.

3. Partnership

Participation by key actors in the community has proved critical to successful problem-oriented policing efforts. Where police have reached out to solve quality of life problems by working with schools, landlords, neighborhood watch groups, businesses, and others, the results have been positive. As Susan Sturm describes, "more interactive forms of participation are both possible and crucial to achieving remedial legitimacy and success." Participation provides police with information and expertise, and helps build relationships and support for their initiatives. Moreover, considerable research suggests that, in general, high quality problem solving and analysis results from participation of people with diverse backgrounds. As the recent Police Executive Research Forum (PERF) report states, since "conducting analysis wholly internal to the agency can make the results appear suspect," not only should outside analysts be employed, but "[c]itizens . . . should be represented in developing and implementing the data collection and analysis system." Participation can begin with a simple invitation to join in problem- solving efforts. Both the Chicago and Philadelphia examples show how despite a history of mistrust between community groups and the police, extending a simple but genuine invitation to work on the problem was enough to begin a meaningful relationship. Police-community meetings have often led to serious efforts to remedy racial profiling. Next, institutionalizing participation, so that regular opportunities exist for police to interact with others, formalizes the relationship and gives outside actors assurance that they will not be shut out, abandoned, or blamed for failures. This can be achieved by convening a committee, as in Philadelphia, on which experts from interested groups are invited to collaborate in solving a problem, not just to serve as outside monitors. Bringing outside groups into the decision-making process also avoids the problems that plague external monitors; excluded from police culture, these actors have been dismissed as out of touch or lacking a stake in the outcome or real access to police information.Alternatively, police can identify the fora in which they already have regular contact with outside groups or the community. Beat meetings, cop watch programs, and even city council meetings could be used as initial ways to disclose data and invite public discussion about a problem. Once there is some dialogue, police can then invite closer participation of interested groups in the actual problem- solving effort.

Information gathering distinguishes this approach from past attempts to start discussion about race and policing. Simply convening a group with a vague mission to discuss race, or the problem of racial profiling, might degenerate into a shouting match. While dialogue alone may help create trust, this lack of definition may be a failing of efforts to hold discussion forums, community meetings, or initial attempts to solicit feedback. The key to formalizing the relationship is that all sides have a project to work on--designing procedures for data collection, developing protocols for stops and the use of race in these encounters, interpreting the information, and reevaluating the procedures.

Data collection, if nothing else, gives everyone something discrete to work with and evaluate. One criticism is that data collection is only a partial solution. Where departments need to rethink training, supervision, hiring, and general policies and procedures, why focus on data collection? Data collection is a point of entry into those larger efforts. By looking at the data collection process, and even deciding what to include in the form that officers fill out, groups will be forced to think about what happens during an encounter, what officers should do and say during a stop, and what elements of this interaction should be scrutinized most closely. In designing data collection, police and citizens will have a chance to evaluate their priorities and begin to think about how to approach racially biased policing. In turn, the data itself will help evaluate the success of other reforms.

Advantages of outside participation itself run in several directions. Bringing in outsiders, by its nature, introduces a new democratic element of shared decision-making in the police department itself, and thus helps break down the insular and hierarchical processes that themselves create barriers for effective problem solving. As Professor Waldeck explains:

By encouraging alliances between the police and the community, the strategy breaks through the "us and them" mentality. Moreover, when the emphasis is on partnerships, aggressive and quasi- militaristic attitudes that risk alienating significant segments of the community are counterproductive. In sum, community policing has the ability to alter not only the external face of policing--with the introduction of permanent beat officers, neighborhood meetings and the like--but also the internal dynamics and values of a police department itself. This challenge to traditional internal decision-making processes can be viewed as a threat, and poses many of the same challenges that effective problem solving poses.

Outsiders have access to different information, and working partnerships allow everyone to 'pool' their data and expertise. The process of institutional reflection requires questioning, which can be better informed by others who share a commitment to a problem, but have a different perspective, or access to different information. Thus, making police information public is a precondition to any meaningful problem solving. Participation requires evaluating police information, comparing it with information groups may have, and deciding whether new decisions should be made.

If the controversy surrounding the way that statistics are gathered teaches us anything, it is that accuracy and completeness can only be gauged by the purpose of the data, which is value laden, requiring the input of a diverse group of outsiders. By defining the racial profiling 'problem' as a numerical disparity, only the race of those stopped would be collected and data collection would become a litmus test. The consent decrees recognize that many more complex factors are at work. As discussed above, the way stops are conducted, the way police talk, body language, cultural differences, whether force is used, and the number of people watching all affect perception about whether a stop is unjust, and those perceptions prove critical to the ways that groups define the problem. Thus, even for sound data gathering, outside groups will be needed to supply that kind of information about perceptions. Police observations about a stop are one-sided and may not capture the full problem. Day-to-day observation may help interpret data, as in the Maltz project in Chicago, and community concerns may help to improve the quality of the data and uncover new problems. Participation may also suggest new ways of collecting data and suggest new sources of information.

Second, outside expertise may be needed. Many police departments, unsure of what the law or the public requires, and facing allegationsof racial profiling, have turned first to the community to decide how to address the problem. They have held meetings and open discussions on the problem. Outsiders also may help open up these discussions and permit a more experimental approach. In Philadelphia, legal groups used their unique perspective to question policies and to revise legal standards that the department had relied on. Furthermore, community groups, as in Chicago, may provide the data itself as well as political support. Other police departments facing similar problems may be interested in cooperating, as may other government institutions. In addition, researchers and information systems experts may prove helpful. The role of the press also proves important as many lawsuits and investigations into racial profiling began through press investigations and studies. Police have found outside participation necessary because solving a problem requires building new relationships and enlisting the support of others. Collaboration with key actors in Philadelphia provided a coalition of support for the changes the police began to make and helped them respond to anger over past wrongs. Political legitimacy flowed from the participation of individual neighborhood police watch groups in Chicago. Part of the racial profiling problem is perception of inappropriate behavior, and police must open up dialogue with outsiders if they are to understand and allay that perception and to gauge improvement.

Trust is particularly critical where police are accused of racialdiscrimination. As the Terry Court stated, "in many communities, field interrogations are a major source of friction between the police and minority groups." Race and police stops raise deep political issues that require political dialogue. For example, Boston police in the early 1990sused a racially charged profile when investigating a murder, causing a rift between police and the community. They have made great strides since then to involve the community in all decisions, making special efforts to build relationships with a Ten-Point Coalition of key community leaders. In 1995, crisis loomed again when a white Assistant Attorney General, Paul McLaughlin, was shot, apparently because of his work to combat gangs. A new murder investigation began with the use of a sparse, racially charged suspect description: "black male, about 14 or 15 years old, 5 foot 7, wearing a hooded sweatshirt and baggy jeans." However, because of their new partnership, "Boston law enforcement officials and Ten- Point Coalition ministers used their response to the McLaughlin murder to solidify and publicly display their new-found cooperation." The situation was handled sensitively by all sides and has only deepened cooperation by solidifying the degree to which police and the community rely on one another to solve problems. Cooperation can gradually foster trust of advocates and community groups. If the public is involved and understands why decisions are being made, the discussion can move beyond divisiveness to issues of how to reform training, allocation of resources, or patrol methods. Being involved in data collection may help allay a perception of discriminatory policing. For example, in the Maltz study, word of mouth would sometimes amplify one incident into a perception of a crime wave, so that collaboration reduced community fears. Trust moves in two directions, so cooperation also may help police trust the community. Individual officers may abandon stereotypes once they get to know and work with community members and become more familiar with a neighborhood. Inviting participation seems natural where policing is no longer seen as specialized work, but instead a social service that has become part of an interlocking web of services. Police regularly cooperate with schools, social services, domestic violence advocates, legal groups, and others. In the past, police have violently and bitterly resented civilian involvement in work that they feel only police officers can understand. However, police have increasingly found such participation useful and have granted broader access to outsiders in ways that are now starting to occur with racial profiling, often using information systems to facilitate communication. In turn, advocacy groups and experts are finding police increasingly receptive to what they have to offer, and have begun to cooperate by providing services and helping to reform police practices.

This move to wider participation and joint problem-solving reflects a shift in the structure of local government. Public-private partnerships, created to address perceived failings in social services, increasingly replace traditional city government bureaucracies. As more community groups provide services in cooperation with government and others, they blur the lines between the government and community. Structural reform remedies can benefit from these new partnerships that are developed precisely because neither community groups nor government think that they can solve the problems alone. Some groups have done innovative work to change the nature of the services delivered, combining social services and providing health, housing and legal help under one roof. While blending private and public can mean an abdication of government responsibility, partnership need not take that form. To the contrary, the idea of government as separate from the community implies hierarchy or isolation; community partnership can make government more accountable. Community policing suggests ways that community members can directly participate in police decisions, allowing them to share in the responsibility for decisions made while police retain accountability