PERF's members are dedicated to addressing racially biased policing in all of its forms and they join the vast majority of police officers in this country who are dedicated, principled men and women committed to treating all citizens fairly and with dignity as we work to prevent, identify and eradicate racially biased policing.

PERF recently released a 160-page report on racially biased policing. (This report is available in its entirety on our website at www.policeforum.org.) It provides nearly 50 recommendations that are based on information culled from surveys returned by more than 1,000 law enforcement executives, materials from more than 250 police agencies, focus groups with citizens and police, a literature review, subject matter experts, national conference discussions, and advisory group input. The advisory group for this project was composed of a diverse group of law enforcement practitioners, community activists, civil rights leaders and academics. Much of the information we gained during this study has helped us better understand what S. 989 would mean for police and citizens.

Some General Principles
To put my remarks in context, I would like to mention a few principles that guide PERF's work in this area. First, we know that racially biased policing must be treated as a human rights issue. It has always been unconstitutional and there are remedies, such as civil litigation under section 1983 and U.S. Civil Rights investigations by the Department of Justice, and these should not be forgotten. We know that racial profiling has had devastating effects on citizens. It is a problem that should be addressed by police working in concert with community leaders, civil rights activists and other stakeholders toward a climate of mutual trust and respect.

Second, we know it is important to address the corrosive impact of both perceived and actual racial bias by police and that the concerns of both law enforcement and citizens go well beyond the standard, very narrow, definition of "racial profiling." We use the term "racially biased policing" rather than "racial profiling" because racial profiling has frequently been defined so restrictively that it does not capture the concerns of both police practitioners and citizens. The most common definition of "racial profiling" refers only to law enforcement activities (particularly vehicle stops) based solely on race. According to the PERF report, "racially biased policing" occurs when law enforcement inappropriately considers race or ethnicity in deciding with whom and how to intervene in an enforcement capacity. Compared with this definition, "racial profiling" targets a much narrower range of activity. These contrasting, but unspoken, definitions lead to confusion when assessing police or citizen perceptions of the problem and need to be clarified in any discussion about the range of activities of concern to police and the public.

The PERF report suggests that executives conceive of "interventions" broadly and consider the following areas in addressing racially biased policing:
1. accountability and supervision;
2. policies prohibiting biased policing;
3. recruitment and hiring;
4. education and training;
5. minority community outreach; and
6. data collection and analysis.

One of our concerns with S. 989 is its overemphasis on data collection as a remedy for this problem-reflecting the overly high expectations nationally for these efforts. We found during the course of our work that there is a widespread misunderstanding on the part of policy makers, agency executives and citizens of what data collection-that is, agencies' collection of race/ethnicity and other data on engagements with citizens-can and cannot do. After characterizing the constraints associated with data collection, I will link those concerns to several specific provisions in the bill.

The Limitations of Data Collection Efforts
In PERF's report, we caution against high expectations regarding the extent to which data can produce valid answers to the serious and legitimate questions an agency seeks to answer. Data collection can be a very positive agency response that can reflect the police executive's commitment to both the community and agency personnel that biased policing will not be tolerated. It can be used as a first step in internal assessments of particular enforcement tactics or individual officer performance when used responsibly. It can be problematic, however, if policy makers, including police executives, look to data collection efforts to provide a reliable and valid assessment of the nature and extent of racial profiling.

Social science is not capable of providing valid answers to every question posed. Indeed, there are many chiefs who sincerely would like to gauge whether or not their departments engage in racially biased policing, but have come to recognize that social science methods are not fully able to produce the information they seek. Specifically, while agencies can have reasonable confidence in the data they collect from their officers regarding whom they stop, there are legitimate questions as to whether there are, at present, cost-effective methods for interpreting those data to reach valid, meaningful conclusions. An agency can produce reasonably valid data to answer the question: "What percent of the people we stop in this jurisdiction for traffic violations are Hispanics?" The question we can't answer is "What percentage proves racial profiling?" That is, an agency might determine that 25 percent of its stops are of Hispanics but be unable (because of the limitations of social science) to draw any reliable conclusions regarding what this means. Some agencies might compare the percentage of Hispanics stopped to census data and yet the people who live in a particular area may be very different from who is traversing the roadways. A department would want to compare the demographics of those stopped with the demographics of those at risk of a stop, taking into consideration numerous factors, including, but not limited to, driving quantity, driving behavior, vehicle condition, and police presence. This information is not readily or easily available.

The key point here is that, in the realm of "racial profiling," some benchmarks can help us show disparity (number of people stopped as compared with representation in the general community population, for example). We don't yet have the social science techniques that can help us to explain that disparity. That is, we cannot rule out all possible explanations for that disparity except race.

In conducting data collection, we are asking social science to determine whether there is a causal main effect between citizen race/ethnicity and police behavior. To show this causal effect we must, among other things, rule out all other possible explanations for that disparity. To draw definitive conclusions regarding stop data that indicate disproportionate engagement of racial/ethnic minorities, we would need to be able to identify and disentangle the impact of race from legitimate factors (such as driving behavior, driving frequency, status of equipment, hot spot enforcement) that might reasonably explain individual and aggregate decisions to stop, search and otherwise engage people. Social science cannot do this well. As stated in a U.S. General Accounting Office report (2000), because of methodological challenges, "we cannot determine whether the rate at which African-Americans or other minorities are stopped is disproportionate to the rate at which they commit violations that put them at risk of being stopped" (p. 18).

In an attempt to rule out alternative factors, agencies strive to develop comparison groups against which to evaluate their vehicle stop data. Specifically, agencies try to develop comparison groups that reflect the demographic makeup of groups at risk of being stopped by police in an unbiased world. We don't have this alternative unbiased world for purposes of comparison and, instead, social science only offers agencies alternatives for "benchmarking," which are wanting. (For instance, agencies are using census data, drivers' license data, accident data and so forth to develop standards for comparison.) In effect, the process of data collection is an effort to collect "circumstantial" evidence to tell us what is going on inside the heads of police officers when they make decisions. The methods we have are simply inadequate for the task.

PERF very much wants to increase the value of data collection efforts and, to that end, we are currently engaged in a federally funded project to develop guidance for police agencies in their analysis and interpretation of data. Our hope is to enhance the general understanding of the potential and constraints of these data and promote their responsible use. (We do not, however, advocate mandatory data collection.)

Indeed, let me emphasize that data collection can be one viable response to the issue of racially biased policing so long as the policy makers, including chiefs, understand what the data can and cannot do. The key is to examine the objectives one wants to achieve and have a full understanding of whether or not data collection can achieve those objectives. It can be a good starting point for assessing overall performance and be part of a comprehensive partnership effort with citizens in addressing the problem. The downside of data collection is misusing the findings because of a lack of understanding of its limitations. This can cause harm to agencies, to individual officers and to communities. This leads me to my specific comments regarding the bill.

Prohibitions Against Racial Profiling and Enforcement
Because of social science limitations, the PERF project team was disturbed during the course of the study to see policies that linked "racial profiling" results directly to officer discipline. This clearly demonstrated a misunderstanding of where these types of data might fall in terms of level of proof. It would be the equivalent of punishing someone based on hearsay evidence alone. The results of data collection efforts-whether at the individual or agency level-are not "proof" of misconduct. Thus, we were similarly disturbed that Title I, Section 102 provides that "[p]roof that the routine investigatory activities of law enforcement agents in a jurisdiction have had a disparate impact on racial or ethnic minorities shall constitute prima facie evidence of a violation of this title." As noted above, a finding of "disparate impact" is meaningless without proper analysis and interpretation, yet S. 989 uses it to shift the burden to police to prove they are not engaged in racial profiling. Data collection that results in a finding of disparate impact does not necessarily indicate that police are racially biased in their enforcement and investigatory decisions. For instance, if a minority community having a problem with a youth gang composed of mostly minority members asks for police action, there may be a disparate impact when "hot spot" enforcement is used, depending on the comparison group used to determine disparate impact.

There are two related issues. One is that, chiefs may well be able to explain to a sufficient level of proof that they are not engaging in racial profiling, but they would still need to defend their actions in court based on this limited data. Upon passage of this legislation, there could conceivably be large numbers of police agencies in the country called into court to explain why the percentage of ethnic minority people arrested (per their Uniform Crime Report data) exceeds their representation in the population. The second issue is that we will be asking agencies to "prove a negative" with social science tools that are quite limited. Agency information that shows disparate impact can reasonably prompt additional investigation by police, analysts and the community to consider the context and myriad factors that went into the law enforcement decisions made. This is what is happening nationwide. But, this type of information does not warrant consideration as prima facie evidence of racial profiling.

The further harm of this provision is that it may have negative consequences for law enforcement's legitimate crime control and prevention activity in neighborhoods with largely minority residents. Several PERF members-including Chief Edward Flynn across the Potomac River in Arlington County, Virginia-shared recently that they have met with leaders of predominantly minority communities who have agreed that additional officer deployment in their neighborhoods is necessary to address crime and disorder. The chiefs' concern was that these deployments might increase allegations of racial profiling because appropriate benchmarks and safeguards against improper analysis and interpretation do not exist. They are concerned they will need to vest significant resources to explain their actions in court, under a cloud of alleged racial bias, even when their activities are legitimate. And, indeed, this provision in the bill appears to do just that.

Shifting the burden to police agencies to prove they are not engaged in racial profiling will only ensure that police agencies who are called into court to justify legitimate law enforcement activities that resulted in disparate impact, but no racial profiling, will have their budgets drained to pay for legal defense. Exacerbating this burden on police is the provision that law enforcement agencies that cannot prove the "racial profiling negative" may be required to pay "reasonable attorneys' fees as part of the costs, and may include expert fees as part of the attorney's fee." While prevailing plaintiffs may receive attorney's fees, even where police agencies prevail, they seldom, if ever, recover the costs associated with defending a claim.

At a time of stretched police agency budgets, it is the community who will suffer if public safety funds are diverted to pay for increased litigation against agencies based only on a finding of disparate impact. We would not want the bill to divert police focus or resources away from serving troubled communities.

PERF's report proposes that data collection should be considered by citizens and the police who serve them as one of many tools that can be used in a comprehensive approach to address racially biased policing, and its use should be determined by the priorities and resources of that community. There is an important place for data collection in the range of possible police-community responses: It just isn't appropriate to presume that disparate impact derived from data collection sufficiently indicates racial profiling to justify the financial burdens and stigma that is placed on police agencies even if they are able to explain the results as unbiased policing.

Loss of Critical Funding for Crime-Ridden Communities
Another troubling aspect of S. 989 is that police funding under the COPS program, Local Law Enforcement Block Grant Program and Byrne grants could be lost for failure to comply with such vague mandates as "cease existing practices that encourage racial profiling." Even if the proscribed activity is better defined, it seems contrary to the stated objective to take support away from citizens in violent neighborhoods under these federal grant programs because there is a police agency that is believed not to have "adequate policies and procedures designed to eliminate racial profiling" (emphasis added).

The funds from these grant programs are being used across the country to do tremendous things for communities. For instance, Local Law Enforcement Block Grants (LLEBG) funds are being spent to enhance domestic violence efforts, reduce gang crime, enhance security in and around schools, provide recreational activities and job training for "at risk" youths, support juvenile drug courts to name just a few uses. It would be tremendously unfortunate if the "at risk" kids, domestic violence victims, juveniles with drug problems and students lost their programs.

The objective should be to help all police agencies do the best possible job to address racially biased policing. If there are problems, we should work to solve them, not take away critical funding that will further impair police services to areas plagued by crime and disorder. It is important to remember that there are already in place powerful remedies for addressing "pattern and practice" as well as individual officer's misconduct that do not include taking funds that ultimately benefit citizens in areas overrun by criminal activity.

We want to be clear on this point: There is no legitimate purpose served by endangering critical public safety funds by adding requirements that grantees address the specific racial profiling mandates outlined by S. 989. Even those who argue that there are reasons to include such mandates must realize the limitations of the requirements that we discuss below.

Requirements for Grant Funding
Policies
S. 989 requires that agencies must have a policy that states a prohibition of racial profiling. We agree that agencies should have a policy, but do not think this should be linked to federal funding. PERF has developed a recommended policy that is already being considered seriously by police agencies across the nation despite the fact that it was released just two weeks ago. It has already been adopted by the National Oceanic and Atmospheric Administration's (NOAA's) Office for Law Enforcement with some modifications.

While our policy is similar in substance to that contained in S. 989, we believe our policy is clearer and, arguably, more encompassing (covering all law enforcement decisions). (It is not clear, for instance, whether the policy set forth in S. 989 encompasses police decisions to arrest or use deadly force.) The PERF Model Policy has been reviewed by law enforcement lawyers and constitutional scholars and builds upon both the 4th and 14th Amendments of the Constitution. In the 4th Amendment realm, our policy makes it very clear that race can never be used as the sole factor to establish either reasonable suspicion or probable cause.

Data Collection
The second requirement for receiving covered grants is "the collection of data on routine investigatory activities… to determine if law enforcement agents are engaged in racial profiling." We do not support mandatory data collection because social science does not currently provide us with the benchmarks and other tools necessary to determine if law enforcement agencies are engaged in racial profiling. This position has also been taken by the Major Cities Chiefs for the same reasons.

And, as mentioned earlier, we think the decision whether or not to collect data is best made by citizens and the police who serve them. We think it is reasonable for these partners to decide that the money to be expended to address this critical issue might be more reasonably and effectively spent on other activities, such as academy and in-service training programs designed to reduce racially biased behaviors, adoption of the PERF policy with associated training, enhanced supervision techniques, concerted minority hiring efforts, purchase of in-car videos, and enhanced outreach to minority communities.

The data collection requirement in S. 989 includes all "routine investigatory activities," which include traffic stops, pedestrian stops, frisks and other types of body searches, and consensual or nonconsensual searches. In the PERF report, we discuss balancing the need for information on high-discretion/low visibility stops against considerations of response times, officer safety, community priorities, resources, etc. Another consideration is that the benchmarking challenges increase exponentially with every additional law enforcement activity targeted (e.g., adding pedestrian stops to traffic stops). Reflecting this balance, we propose that agencies choosing to collect data, or mandated by state law, target vehicle stops (that is, traffic and investigatory stops of motorists). Further, we acknowledge that some agencies may choose to institute data collection in stages-adding additional categories of activities as the system is developed. The agencies collecting vehicle stop data or adding activities incrementally to produce a sound system would be ineligible for covered program grant funds.

Citizen Complaint and Officer Disciplinary Procedures
S. 989 also requires that grant applicants have adequate citizen complaint and officer disciplinary procedures to address racial profiling. The PERF report supports procedures that will increase the transparency and integrity of the citizen complaint process. We have recommended, "the public complaint management system include a separate category to permit clear and accurate monitoring of complaints of biased policing, with the capacity to identify patterns and practices inimical to equal treatment of citizens." We recommend audits of the complaint system with spot checks to evaluate effectiveness. We have also recommended other means for ensuring that complainants are not intimidated, discouraged or coerced in any way and understand the process, among others. With that said, these recommendations should never be a sanctioned requirement for federal grants listed under covered programs in the bill.

As to oversight requirements for grantees, there is no single model of police oversight that will work in every jurisdiction. The term "independent" used to describe the required complaint procedures is not defined and will lead to myriad interpretations. The independent complaint procedures required by the bill may exclude review boards that effectively use both police and citizen input, if police involvement means the process is no longer independent. The federal government should not dictate to state and local authorities what type of review board they should have, so long as the process is effective, efficient and fair. In fact, the reason that PERF has opposed the federal Police Officer Bill of Rights legislation called the State and Local Law Enforcement Discipline, Accountability and Due Process Act of 2001 (H.R. 1626 and S. 840) is precisely because those federal mandates would undermine the state and local controls and investigatory processes that effectively hold officers accountable for misconduct. The oversight provision is an inappropriate requisite for funding.

Attorney General's Policies
It is unclear why there is a need for an additional policy requirement that law enforcement agencies must meet "such other policies or procedures that the Attorney General deems necessary to eliminate racial profiling" to retain and gain federal grant funding. We are unaware of why there would be a need to legislate such authority.

Support for a Best Practices Grant Program
PERF agrees that an appropriate federal role in addressing racially biased policing is providing funds to enhance or create local police efforts to address racially biased policing.

It is our understanding from Hill staff that the Section 302 Best Practices Grant Program does not require compliance with the mandates set out in section 301. If this is the case, PERF would support such a grant program and is pleased that the list of supported activities reflects some of the many suggestions in the PERF report about effective responses to racially biased policing. (We should note, however, that some PERF members voiced concern that the grant program appropriations would be made from existing police funding programs such as COPS, Byrne and Local Law Enforcement Block Grants. PERF would not support draining existing law enforcement funds to support the new grant program.)

The grant program is consistent with PERF's objectives to see more work done in training and education, for example. Our report includes recommendations for topics and methods of delivery for both training and education. Police agencies should consider integrating education and training related to racial bias in a wide range of curricula, although a single course of instruction may suit immediate needs. The funds for training outlined in the bill could also be used to train officers in policies-such as the model developed by PERF-that clearly and tightly restrict officers' use of race or ethnicity to make law enforcement decisions. Expansive reforms to training and education would benefit from federal funding support.

PERF believes that support for data collection technology will aid those agencies that choose to collect data or are mandated by state law to do so, but believes that the use of those funds as specified in the bill be expanded further to help underwrite costs associated with changing forms, increased personnel time and other associated budget items.

Many PERF members have also struggled with finding funding for in-car cameras and portable computer systems and would welcome grants to support these efforts.

The best practices grants would include support for early warning systems tailored to the needs of each jurisdiction, which many PERF members have considered adopting in identifying "bad apples" as well as institutional policies and/or procedures that may have the unintended consequence of contributing to racially biased policing.

In terms of accountability and supervision, PERF's report focuses on the need for an assessment of the organizational culture; quality assurance methods for all operations; an integrated approach for encouraging police awareness and appreciation of racial/ethnic diversity and cultural differences; and regular reviews of the complaint reception and management/monitoring processes. Funding in these areas would be an important investment in these critical processes.

According to the PERF report, supervisors should look for evidence of improper practices and patterns and should be responsible for ensuring that citizen complaints of biased policing are given a formal and respectful hearing, and that complaints are documented in accordance with agency policy. The ranking police representative should ensure that complainants are not subjected to any form of discouragement, intimidation or coercion in filing their complaints at the police station or in bringing their complaints to the attention of any officer. They must also provide the complainant with information on how the department deals with complaints, and with the name of the office responsible for handling them. Many of these recommended measures would benefit from financial commitment to citizen complaint systems and community education efforts that many agencies will struggle to find in current budgets.

PERF believes additional allowable uses for grant funds might support efforts to increase minority representation on our police forces or support trust-building partnerships between agencies and their minority communities. Proactive efforts to recruit and keep a diverse police force must be realized. PERF also recommends periodic audits of selection processes to ensure that qualifications and standards are valid and fair to all applicants and that neither the sequencing of the testing stages nor the length of the process hinders minority hiring. Once selected, police executives should determine whether minority recruits are disproportionately dismissed from the agency during recruit training, field training and probationary periods, and if so, determine why and seek ways to reduce that disparate impact.

Finally, there are many innovative minority outreach and partnership efforts being conducted across the nation (and which are described in the PERF report) that can be replicated and tailored to other jurisdictions, which would also benefit from funding support.

Summary
In closing, we can appreciate that the sponsors of S. 989 want to eliminate racially biased policing. PERF members join them in that goal. We support the provisions in the bill that would provide federal grants to support training, education, data collection for those who choose it or are mandated by state law, new technologies such as in-car cameras, and others. We should not, however, be holding back federal grants that benefit citizens in crime-ridden communities if police are unable to meet standards or requirements related to racially biased policing, when other enforcement measures exist.

We hope that we can continue to work with staff and Members on these and other law enforcement issues and that we will find constructive grant programs and other measures that will help the vast majority of police professionals in this country who are eager to prevent and address racially biased policing.

Thank you.