Measuring Racial Profiling by Police


Jul 6, 2004

This commentary originally appeared in Law Enforcement News on July 6, 2004.

Racial profiling by police is a complex phenomenon that is extremely hard to measure. We all want complete answers on whether such profiling is taking place in communities, but we won’t get those answers from incomplete data. Unless we take the time to ask the right questions and get the right answers, the truth about how much racial profiling is really going on will remain unknown.

Racial profiling is frequently defined as law enforcement activities — such as traffic stops, arrests and searches — that are initiated solely on the basis of race, unless police are looking for a specific suspect who has been described by race. Police departments around the country have been accused by critics of using profiling to unfairly target law-abiding members of minority groups for closer scrutiny.

Many racial profiling studies have drawn criticism from police in the past for lacking valid statistical backing. But critics of police departments have countered that police don’t want to face up to the problem of racial profiling and thus are trying to do all they can to discredit the unfavorable studies simply to defend their own misconduct.

The latest racial profiling study to come under police attack was issued by Northeastern University in Boston in May. It concluded that most police departments in Massachusetts racially profile minority drivers.

The Northeastern study said minority drivers in Massachusetts get a disproportionate number of traffic tickets and are searched more often than whites. In contrast, the study found that whites stopped by police get off with only warnings more often than do minorities

An open-and-shut case of racial discrimination by police, right? Well, not necessarily.

Our analysis of the Northeastern study of traffic stops by 366 police departments shows that the study simply did not collect enough data to support its claim of racial profiling by 249 of the departments. That doesn’t mean the study’s conclusions are wrong. It simply means that we can’t say one way or the other what the report really found without examining additional information.

At first glance, it seems hard to understand how a study that analyzed 1.6 million traffic citations issued from April 1, 2001, to June 30, 2003, might not be thorough enough. But on closer examination, it is clear that key deficiencies in the data make the findings of the Northeastern report highly questionable.

To determine whether there are racial differences in traffic ticket rates for residents of a given city, the authors of the Northeastern study compared the racial distribution of the citations with the racial distribution of the residential census population. But that’s the wrong way of proceeding. The residential census has been widely discredited as a benchmark because driving patterns may well differ by race.

Blacks and whites, for example, may use public transportation at different rates — a difference that would affect exposure to traffic citation enforcement.

In addition, the residential analysis did not collect information on the location of the traffic stops and therefore is not adjusted for the distribution of law enforcement activities. Out on the streets, police may devote more manpower to law enforcement in minority communities. Whether these deployment patterns are appropriate is a legitimate subject of community debate. But without information on driving patterns, location of stops and law enforcement deployment, the Northeastern results do not tell us much.

The Northeastern report also compared the rates at which white and non-white non-residents were given traffic citations. A non-resident is someone cited in one city but living in another. Here the authors developed a Driving Population Estimate that analyzed factors (jobs, entertainment, etc.) that might push or draw drivers through communities where they do not live.

This analysis suffers from the same limitation that the resident analysis does. Without information on the location of the stop and the deployment pattern, we can’t conclude much about the disparities in the stop patterns. Indeed, as the authors report, “disparities may be attributable to officer bias, institutional bias, or differential law enforcement…in response to crime control problems.” This variety of possible causes is so broad that it amounts to an admission that the Northeastern study raises more questions than it answers.

The Northeastern study concludes that the “report will be extremely important as departments develop strategies to reduce the disparities that have been identified.” However, the disparities identified in this report could easily be reduced or disappear once stop location and deployment patterns are accounted for. Similar problems exist for the search analysis.

Despite the shortcomings of the Northeastern study, it will have one beneficial outcome. Massachusetts has ordered the 249 police departments accused of racial profiling by the study to collect data for another year on all traffic stops. A full and proper analysis of all that new information will lead to a better understanding of the extent of racial profiling being carried out by police in the state.

It is always tempting to look for easy answers to hard questions and rush to judgment on issues that take a long time to figure out. We should not make that mistake when examining racial profiling by police. Wrong measurements of the problem lead to wrong conclusions and wrong remedies. Communities need to take the time and invest the effort to find the right answers.

Jack Riley is director of the Public Safety and Justice unit for the RAND Corporation, a nonprofit research organization. Greg Ridgeway is a statistician in the unit.

More About This Commentary

Commentary gives RAND researchers a platform to convey insights based on their professional expertise and often on their peer-reviewed research and analysis.