- Does the racial distribution of the pedestrian stops by New York Police Department officers suggest racial bias?
- Do certain officers disproportionately stop nonwhites?
- Are there racial differences after the stops?
In 2006, the New York City Police Department (NYPD) stopped a half-million pedestrians for suspected criminal involvement. Raw statistics for these encounters suggest large racial disparities — 89 percent of the stops involved nonwhites. Do these statistics point to racial bias in police officers’ decisions to stop particular pedestrians? Do they indicate that officers are particularly intrusive when stopping nonwhites? The NYPD asked the RAND Center on Quality Policing (CQP) to help it understand this issue and identify recommendations for addressing potential problems. CQP researchers analyzed data on all street encounters between NYPD officers and pedestrians in 2006. They compared the racial distribution of stops to external benchmarks, attempts to construct what the racial distribution of the stopped pedestrians would have been if officers’ stop decisions had been racially unbiased. Then they compared each officer’s stopping patterns with an internal benchmark constructed from stops in similar circumstances made by other officers. Finally, they examined stop outcomes, assessing whether stopped white and nonwhite suspects have different rates of frisk, search, use of force, and arrest. They found small racial differences in these rates and make communication, recordkeeping, and training recommendations to the NYPD for improving police-pedestrian interactions.
External Benchmark Analyses
- Evaluating racial disparities in pedestrian stops using external benchmarks is highly sensitive to the choice of benchmark. Therefore, analyses based on any of the external benchmarks developed to date are questionable.
- Benchmarks based on crime-suspect descriptions may provide a good measure of the rates of participation in certain types of crimes by race, but being a valid benchmark requires that suspects, regardless of race, are equally exposed to police officers.
- We found that black pedestrians were stopped at a rate that is 20 to 30 percent lower than their representation in crime-suspect descriptions. Hispanic pedestrians were stopped disproportionately more, by 5 to 10 percent, than their representation among crime-suspect descriptions would predict.
Internal Benchmark Analyses
- We compared the racial distribution of each officer's stops to a benchmark racial distribution constructed to match the officer's stops on time, place, and several other stop features.
- This analysis identified 15 officers who stopped more blacks and hispanics than their colleagues, while 14 officers stopped fewer. This means 0.5 percent of the 2,756 NYPD officers most active in pedestrian-stop activity were flagged as having stop patterns warranting further investigation. Those 2,756 most active officers accounted for 54 percent of the total number of 2006 stops. The remaining stops were made by another 15,855 officers, for whom an accurate internal benchmark could not be constructed, mostly because they conducted too few stops.
Differences in stop outcomes experienced by different race groups are substantially reduced when we ensure that comparisons of outcomes are made only between stops that are truly comparable in terms of, for instance, the time of day and location of the stop.
- Officers frisked white suspects slightly less frequently than they did similarly situated nonwhites (29 percent of stops versus 33 percent of stops).
- Black suspects are slightly more likely to have been frisked than white suspects stopped in circumstances similar to the black suspects (46 percent versus 42 percent).
- The rates of searches were nearly equal across racial groups, at between 6 and 7 percent. However, in Staten Island, the rate of searching nonwhite suspects was significantly greater than that of searching white suspects.
- White suspects were slightly more likely to be issued summons than were similarly situated nonwhite suspects (5.7 percent versus 5.2 percent). On the other hand, arrest rates for white suspects were slightly lower than those for similarly situated nonwhites (4.8 percent versus 5.1 percent).
The NYPD should review the boroughs with the largest racial disparities in stop outcomes.
The NYPD should identify, flag, and investigate officers with unusual stop patterns.
All officers should explain to pedestrians why they are being stopped.
New officers should be fully conversant with stop-question-frisk (SQF) documentation.
The UF250 form, which must be filled out after an SQF, should be revised to capture data on use of force.
The NYPD should consider modifying the audits of the UF250 form.
Table of Contents
Introduction: Review of the New York City Police Department’s Stop, Question, and Frisk Policy and Practices
Description of the 2006 Stop, Question, and Frisk Data
External Benchmarking for the Decision to Stop
Internal Benchmarking for the Decision to Stop
Analysis of Post-Stop Outcomes
Conclusions and Recommendations
Details of Statistical Models Used in the External-Benchmark Analysis
Details of Propensity-Score Weighting
Estimating False Discovery Rates
Unified Form 250: Stop, Question, and Frisk Report Worksheet
The research described in this report was supported by the New York City Police Foundation and was conducted under the auspices of the Center on Quality Policing (CQP), part of the Safety and Justice Program within RAND Infrastructure, Safety, and Environment (ISE).
This report is part of the RAND Corporation Technical report series. RAND technical reports may include research findings on a specific topic that is limited in scope or intended for a narrow audience; present discussions of the methodology employed in research; provide literature reviews, survey instruments, modeling exercises, guidelines for practitioners and research professionals, and supporting documentation; or deliver preliminary findings. All RAND reports undergo rigorous peer review to ensure that they meet high standards for research quality and objectivity.
This document and trademark(s) contained herein are protected by law. This representation of RAND intellectual property is provided for noncommercial use only. Unauthorized posting of this publication online is prohibited; linking directly to this product page is encouraged. Permission is required from RAND to reproduce, or reuse in another form, any of its research documents for commercial purposes. For information on reprint and reuse permissions, please visit www.rand.org/pubs/permissions.
The RAND Corporation is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.