Collecting statistics in response to racial profiling Allegations. (Perspective).

"This isn't a problem that can be quantified just in terms of statistics.... It's like the guy who is unemployed; the unemployment rate for him is 100 percent. For people who have been the victim of racial profiling, the statistic is 100 percent." (1)

U.S. Attorney General John Ashcroft 

The simple collection of data will neither prevent so-called "racial profiling" nor accurately document a law enforcement agency's activities as a means of protecting it from public criticism, scrutiny, and litigation. (2) Statistics represent meaningless numbers unless they are put in a relevant context or used as a legitimate means of comparison. (3) Standing alone, statistics, much like legal arguments, can be used to make or defend any position that someone may adopt on an issue. Law enforcement agencies, therefore, must take additional steps to ensure that the numbers they collect accurately reflect reality and support the positive enforcement and crime prevention efforts that they conduct. (4)

UNDERSTANDING THE TERMINOLOGY

Understanding the terminology of racial profiling constitutes the first step to gathering relevant statistical information. As some commentators have noted, the use of race for law enforcement purposes is unconstitutional when it is the factor used in selecting potential criminal suspects. (5) Other commentators have described the problem as a "targeting of individuals for police investigations based on their race alone." (6)

Because the Constitution requires that a law enforcement officer have "reasonable suspicion to stop and detain a person," (7) no one disputes that stopping or detaining individuals because of their race or ethnicity is unconstitutional. As two researchers describe it: "There is no one list of factors that gives rise to reasonable suspicion, as the varieties of suspicious behavior are as diverse as the types of activity punishable under the criminal law. However, reasonable suspicion may not be based on race alone." (8) No Fourth Amendment action, be it an investigative detention, an arrest, or a search, can be undertaken based solely upon race. Therefore, officers who base their Terry stops or other investigative activities on a suspect's race alone violate the law and are guilty of misconduct that very well may be sufficient to terminate their employment. (9) In short, no circumstances exist under which officers may stop citizens based solely on their race, sex, religion, national origin, or sexual orientat ion. Rather, officers must base stops on reasonable suspicion--facts and information that they can articulate, which, in turn, lead to their sense of suspicion that a violation of law has been or is about to be committed. (10)

The law enforcement community must strive to remind the public and the press of the correct definition of the problem, lest they come to believe that race has no importance in the investigation and prevention of crime or that such uses are improper.

For instance, the race of an at-large suspect often represents an important identifying characteristic that agencies must include in radio broadcasts. (11)

Similarly, law enforcement officials should remain alert to any inappropriate use of terminology that is not legally accurate or recognized, but appeals to the emotional or political aspect of this debate. For example, some legislative proposals use the terms consensual and nonconsensual when setting parameters for data collection relating to searches of motor vehicles. The term nonconsensual conveys that a search was conducted by use of force against the will of the individual, rather than on a recognized legal basis, such as incident to arrest. (12) Law enforcement authorities should avoid such misleading terminology because it creates mistaken and unfavorable impressions.

EXPLORING THE PROBLEM

Apparently, the public has come to believe that if the police are required to keep a record of their investigative activities, the "problem" (13) of racial profiling will be curtailed. (14) While accurate and meaningful data collection may have some social science and management value, it is fraught with pitfalls. In fact, the effort to collect statistics may further erode the public's trust in law enforcement and the morale among its members, as well as spend government dollars that could be better used on law enforcement training and education, community crime prevention, and drug treatment.

A more effective way to alleviate the problems caused by the perception that police engage in racial profiling rests with education--of the police and the public. Those officers who actually stop citizens only because of their race are either ignorant of the law or are unethical or immoral. Training in constitutional law and ethics, along with effective first-line supervision, would directly address this problem. It also may prove helpful for agencies to develop procedures for making more professional traffic stops that demonstrate sensitivity to this issue by including, for instance, informing the driver of the facts that led the officer to make the stop. Those facts would, and must, establish the reasonableness of the stop. (15)

Equally important are departmental policies, supervision, and discipline. Law enforcement agencies should send a clear message to all personnel that using race alone as the basis for any investigative stop is unacceptable conduct and can lead to termination from employment. In short, this is a "zero tolerance" issue. Officers who do not respond to training and discipline or appear simply immoral have no place in law enforcement.

A need also exists to educate the public. Agencies should convey to their communities that law enforcement is a difficult and dangerous profession and that most who enter it do so for only the best reasons. The public needs to know that the narrow scope of the racial profiling problem and the U.S. Constitution protect them from such abuses and that such abuses are not tolerated or advocated by law enforcement agencies.

COLLECTING THE STATISTICS

These thoughts aside, the perception appears to exist that the public wants law enforcement to do nothing more than "collect statistics," but with no apparent defined purpose. If agencies find themselves in a posture that requires this collection--either by legislation or some kind of executive order--or decides to do so voluntarily, they should bear in mind some complexities and sophisticated issues that generally are not acknowledged as problems with statistical collection of this kind. (16)

Methodology

The collection and application of statistical data is a scientific and academic exercise requiring a well-designed protocol. (17) Most mandated law enforcement data collection occurs as the result of a legislative compromise or an effort to appease particular interest groups and contains weak methodology. The protocols are determined not by academics or mathematicians but by legislators, governors, or other officials who may have no training or expertise in the field of statistics and, therefore, may pay little attention to the purpose, quantity, or nature of the data that they need to collect.

Experts in the field should design the data collection system and base it on a testable hypothesis. They should include protocols that will ensure that the process will collect empirical data for research purposes and not merely to provide evidence for a particular advocacy group or to falsely defend an errant officer. (18) These protocols must screen out other behavioral variables that may be culturally or geographically based or that include other relevant but nonracial factors. (19) Examples include highway safety studies that show African-American youth as 50 percent less likely to use seat belts than whites or Hispanics. (20) Similarly, traffic stops based on vehicle equipment violations may occur more often in some minority neighborhoods because owners may have fewer resources to devote to repairing their vehicles. (21)

The data collection methodology also should require that agencies collect sufficient data to allow for meaningful analysis. The notion of a "statistically significant number" is important but rarely mentioned when discussing data collection. Primarily, agencies must ensure that they collect enough data over a specific period of time to adequately appraise the circumstances surrounding the types of stops that their officers make.

Implementation

With all of these elements in mind, the collection of data must not burden the officers and agencies to the extent that it has a "chilling effect" on enforcement. The record keeping by officers may become quite time-consuming, interfere with their lawful discretion, and create animosity with the public. (22) There is even the potential for civil liability, depending on the time required and the method used. For example, if a citizen has to remain on the scene during the time an officer needs to fill out an "extra" form, the citizen may perceive this as an unlawful detention. Also, the collection of "racial" statistics may imply that race discrimination problems exist, thereby enhancing the public's negative perceptions at the cost of both the morale and effectiveness of officers. Finally, officers become burdened with attempting to identify the race of motorists by sight or by asking them outright, which may escalate tensions and create a dangerous, or at least uncomfortable, situation for both parties. (23)

The financial impact represents another consideration in the data collection process. Accurate and well-designed collection methods may require the purchase of new computer equipment, the hiring of persons to input data, and the employment of experts to analyze the meaning of the data. When a legislative body or governing executive requires the collection of statistics, appropriate funding rarely is attached to the mandate. (24) Agencies must consider whether it is in their communities' best interests to impose such a burden on their budgets, rather than seeing that money spent on crime prevention and enforcement efforts.

A data collection initiative also must take into account the differences between municipal and highway policing. The work of urban and suburban officers is responsive in nature--they engage in police work when and where they are summoned. Their enforcement pattern depends on the character of the neighborhoods that they serve. Highway policing is more self-initiated and, arguably, calls for more discretion. Conceivably then, municipal and highway collection models should be quite different from each other. Moreover, it may prove useful for the collection and analysis to distinguish between low and high discretion stops. (25)

Analysis

Finally, the data collection process must include a clear direction for the analysis of the data. If agencies compare the collected statistics to some demographic statistic, what is the relevant demography or benchmarks? In a highway patrol situation, this issue proves especially problematic because an interstate highway has neither a static "population" nor a counting mechanism, such as a census. To compare highway statistics to the local population is disingenuous because the highway is populated with interstate travelers. Similarly, when comparing urban data, agencies must establish a concomitant analysis of the local criminology and appropriate benchmarks or a comparative control group.

Even the U.S. Department of Justice acknowledges that jurisdictions need assistance "in designing statistical benchmarks and determining comparative populations" (26) and that "the characteristics of a traffic stop are difficult to interpret." (27) Likewise, the author's research has yet to find an analytical model that has established a meaningful benchmark for comparison or a sufficiently sophisticated method of interpretation. (28)

With respect to these benchmarks, another neglected issue arises in the discussion concerning the analysis of the data collected. A racial disparity, if one can be established, is only legally significant under the Equal Protection Clause of the Fourteenth Amendment if individuals stopped solely on the basis of their race were similarly situated to others who were not stopped. (29) In other words, the correct comparison is not to the people living in the neighborhood or driving on the highway who did not engage in the same conduct as the person stopped. Rather, the complainant must show that those who did engage in the same conduct were not stopped because they were not of a minority race. Therefore, to demonstrate racial bias, the statistics used by the complainant must show that officers had an equal opportunity to stop others who were similarly situated to the suspect person but who they did not stop solely because they were not of a minority race. (30)

CONCLUSION

The collection of data concerning the issue of racial profiling cannot be taken lightly. Superficially, it may seem like a relatively painless way to appease public and political concerns, especially to those law enforcement executives who are confident that their departments are functioning properly. However, significant pitfalls exist. If agencies do not base their collection methodology on a well-designed scientific model, the resulting statistics can be manipulated easily to serve as a sword, rather than a shield. To then defend against this sword, agencies may be forced to attack the basis of their own statistical collection efforts, which ultimately may be perceived as an effort to engage in a cover-up.

There is no question that agencies must address the public concerns about racial profiling and the related issues of public confidence and respect for the law enforcement community. However, it remains naive to believe that data collection is the sole answer. Indeed, if agencies do not properly conduct the data collection process, it may only serve to exacerbate the problem and undermine legitimate crime-fighting efforts.

The opinions expressed in this article are the author's alone and not those of the Many/and Attorney General or his staff.

Endnotes

(1.) "Bush and Ashcroft Announce Racial Profiling Initiative," Criminal Justice Newsletter 31, no. 8 (2001): 2-4.

(2.) For an overview of racial profiling see, Richard G. Schott. "The Role of Race in Law Enforcement," FBI Law Enforcement Bulletin, November 2001, 24-32.

(3.) See, for example, T. Ginsberg, "High Probability of Befuddlement," Baltimore Sun, July 15, 2001, sec. C, p. 5, noting that "[e]ven if statistical findings are repudiated, they can still perpetuate statistical myths"; and H. MacDonald, "The Myth of Racial Profiling," City Journal 11, no. 2 (spring 2001), "Their alleged evidence for racial profiling comes in two varieties: anecdotal, which is of limited value, and statistical, which on examination proves entirely worthless."

(4.) For information on racial profiling data collection systems, see Deborah Ramirez, Jack McDevitt, and Amy Farrell, Northeastern University, A Resource Guide on Racial Profiling Data Collection Systems: Promising Practices and Lessons Learned, NCJ 184768 (U.S. Department of Justice, November 2000).

(5.) Grady Carrick, "Professional Police Traffic Stops: Strategies to Address Racial Profiling," FBI Law Enforcement Bulletin, November 2000, 8-10.

(6.) Abraham Abramovsky and Jonathan I. Edelstein, "Pretext Stops and Racial Profiling After Whren v. United States: The New York and New Jersey Responses Compared," 63 Alb. Law Rev. 725, 729 (2000). See U.S. Department of Justice, A Resource Guide on Racial Profiling Data Collection Systems: Promising Practices and Lessons Learned (Washington, DC, 2001), 3, defining racial profiling as "police-initiated action that relies on...race..."; and Earl M. Sweeney, "Ohio's Statewide Effort to End Profiling," Police Chief July 2001, 16.

(7.) Terry v. Ohio, 392 U.S. 1 (1968). See David Rudovsky, "Law Enforcement by Stereotypes and Serendipity: Racial Profiling and Stops and Searches Without Cause," 3: 1 Journal of Const. Law 296, 306 (2001).

(8.) 63 Alb. Law Rev. 725, 729; and supra note 7 (Rudovsky).

(9.) A recent legislative proposal (2000) in Maryland offered that an officer who engaged in racial profiling be fined $1,000 and be required to attend training. The Maryland legislators appeared surprised when several chiefs of police informed them that they would fire any officer they found to have committed racial profiling.

(10.) Terry v. Ohio, 392 U.S. at 30-31.

(11.) See Brown v. City of Oneonta, 221 F.3d 329 (2d Cir. 2000); and supra note 3 (MacDonald) noting that DEA has identified common characteristics of drug couriers that may include "the ethnic makeup of drug-trafficking organizations." MacDonald also asserts that "[t]he antiprofiling crusade thrives on ignorance of policing and a willful blindness to the demographics of crime." See also Hoover, "Why the Resistance to Collecting Race Data on Police Traffic Stops," Police Labor Monthly, July 2000, 5.

(12.) See, for example, New York v. Belton, 453 U.S. 454, 460 (1981).

(13.) It continues to be difficult to discern how prevalent the practice of racial profiling--using race as the sole reason for initiating an investigative stop--really is. Many anecdotal accounts are based only on the perception of the individuals who were stopped and who may or may not be aware of other factors that officers considered in determining that they had reasonable suspicion to stop these individuals. Others argue that statistical evidence exists; however, that evidence often seems to have been collected under questionable circumstances and by nonobjective parties. Those statistical studies also may suffer from some of the shortcomings that this article discusses, see supra note 3 (MacDonald). Other commentators pointedly disagree, see supra note 7 (Rudovsky), 298-303.

(14.) See supra note 6 (U.S. Department of Justice), 13, asserting that data collection can identify potential police misconduct and deter it.

(15.) Supra note 6 (Sweeney), 18.

(16.) At present, a growing recognition exists that the attendant issues to data collection are unknown, uncertain, or unresolved, as demonstrated by the references noted in this article. See, in particular, Fridell, Luaney, Diamond, Kubu, Scott, and Laing, Racially Biased Policing: A Principled Response, Police Executive Research Forum, July 2001.

(17.) The U.S. Department of Justice has suggested a "core set of data" to be collected, consisting of nine elements. U.S. Department of Justice, Traffic Stops and Data Collection Analyzing and Using the Data (Washington, DC, February 2000), 5-6. See also supra note 6 (U.S. Department of Justice) and supra note 16 (Fridell et al), 126-128.

(18.) On this issue, the author gratefully acknowledges the insights provided by Carl Milazzo and Aimee B. Anderson during their lecture, "Current Legal Issues in Racial Profiling: 2000 Update," presented at the annual conference of the International Association of Chiefs of Police in November 2000.

(19.) Supra note 7 (Rudovsky), 365.

(20.) Benson, Crawford, and Mitchell, 1998, cited in U.S. Department of Transportation, Blue Ribbon Panel to Increase Sear Belt Use Among African-Americans: A Report to the Nation (Washington, DC, 2000).

(21.) Supra note 11 (Hoover).

(22.) U.S. Department of Justice, Traffic Stops and Data Collection: Analyzing and Using the Data (Washington, DC, February 2000), 9-10.

(23.) Jerry A. Oliver and Alicia R. Zatcoff, "Lessons Learned: Collecting Data on Officer Traffic Stops," Police Chief July 2001, 23; supra note 11 (Hoover), 2; and supra note 16 (Fridell et al.), 129.

(24.) See generally supra note 6 (U.S. Department of Justice).

(25.) Supra note 6 (U.S. Department of Justice), 9-10.

(26.) Supra note 6 (U.S. Department of Justice), 56.

(27.) Supra note 6 (U.S. Department of Justice), 53.

(28.) See supra note 3 (MacDonald) noting that "no traffic study to date comes near the requisite sophistication"; and supra note 16 (Fridell et al.), 137.

(29.) Yick Wo v. Hopkins, 118 U.S. 356 (1886). See also supra note 16 (Fridell et al.), 133-134, noting that "these data cannot prove causation--only correlation."

(30.) See supra note 7 (Rudovsky), 322; and Chavez v. Illinois State Police, 251 F.3d 612, 645-48, (7th Cir. 2001), "without comparative racial information, plaintiffs cannot prove that they were stopped, detained, or searched, when similarly situated whites were not."