Racial Profiling - The John F. Finn Institute for Public Safety

Racial Profiling

Stops by Syracuse Police, 2006 – 2009

Robert E. Worden, Sarah J. McLean, and Andrew P. Wheeler (June, 2010)


The City of Syracuse (NY) adopted legislation in 2001 mandating data collection on stops by police. The Institute analyzed data on stops by Syracuse police between 2006 and 2009. Applying the “veil-of-darkness” method devised by researchers at the RAND Corporation, our analysis found hardly any evidence of racial bias in vehicle stops. We conclude that further data collection and analysis is warranted, to demonstrate the SPD’s responsiveness to public concerns, but more importantly, steps that address the perceptions of profiling, and that may improve police-community relations, could also be taken.

Stops by the Syracuse Police

Presentation to the Syracuse Common Council Public Safety Committee
November 15, 2010

Syracuse Police Department website, racial profiling studies

Comment on the 2010 City of Syracuse Police-Citizen Encounter Study

The study commissioned by the Common Council chose to take an “outcome-based approach” that is rooted in economics. The approach was formalized by a group of economists (Knowles, Persico, and Todd) in 2001. The logic is fairly simple and intuitive: if in initiating discretionary searches the police apply the same standard of suspicion to people of different races or ethnicities, then the rate at which police discover contraband and take legal action (the “hit rate”) will be the same across racial and ethnic groups. Any group with a lower “hit rate” may have been subjected to unwarranted police authority. However, this outcome test rests on some key assumptions about the motives and behavior both of police and of offenders, and at best it applies only to discretionary police searches. It is not appropriate for the detection of bias in stops by police, which would require assumptions about police decision-making that are contradicted by 40 years of research on police behavior. The flaws in this study included:

  • “Hit rates” of arrests and tickets (combined) were computed for all stops of each race/ethnicity category, an analysis that rests on untenable assumptions about police officers’ behavior and that even the economists who formalized the approach acknowledge would be inappropriate. Comparisons of these hit rates are ill-advised.
  • The data on ‘on-view’ (non-dispatched) arrests do not include information on searches by police. But because department policy requires that officers conduct a search incident to any arrest, the analysis stipulates that a search was conducted in every search. This is an error: these are not discretionary searches, and moreover, the arrest precedes rather than follows the arrest; the search was an outcome of the arrest, not vice-versa. The arrest records do not indicate whether any other (discretionary) search was conducted, and that might have formed the basis for the arrest. This is the reason that the hit rates are outlandishly (and artifactually) high – over 90 percent, and the comparisons are not informative.
  • For stops that did not culminate in arrests, the data indicate only whether a search OR frisk was conducted. Frisks are not the equivalent of searches; the former are performed for the purpose of officer safety, and not to detect any form of contraband (other than a weapon), so the assumptions of the outcome test do not hold. Moreover, we would expect that frisks are more common in high-crime areas, which also tend to be areas with disproportionately large minority populations, and this would artificially depress the rates of minorities more than those of whites.
  • Many on-view arrests are not stops, as such, but are based on follow-up investigation and the development of probable cause prior to the contact with the citizen. These arrests should not be analyzed as stops.

The first two problems are, by themselves, fatal.

Other Resources

Web sites

Northeastern University, Racial Profiling Data Collection Resource Center

Reports and Articles

Amy Farrell, Jana Rumminger, and Jack McDevitt, New Challenges in Confronting Racial Profiling in the 21st Century: Learning from Research and Practice (Boston: Northeastern University), available at http://www.racialprofilinganalysis.neu.edu/IRJ_docs/Report_NewChallenges21.pdf.

Lorie A. Fridell, By the Numbers: A Guide for Analyzing Race Data from Vehicle Stops (Washington: Police Executive Research Forum, 2004), available at http://www.policeforum.org/library/racially-biased-policing/by-the-numbers/BytheNumbers%5B1%5D.pdf.

Greg Ridgeway, Cincinnati Police Department Traffic Stops: Applying RAND’s Framework to Analyze Racial Disparities (Santa Monica: RAND Corporation, 2009), available at http://www.rand.org/pubs/monographs/MG914/.

Robin Engel, James Frank, Rob Tillyer, and Charles Klahm, Cleveland Division of Police Traffic Stop Data Study: Final Report (Cincinnati: University of Cincinnati, 2006), available at http://www.uc.edu/ccjr/Reports/ProjectReports/Cleveland_Traffic_Stop_Study.pdf.