The effectiveness and equity of NYPD stop and frisk policy, 2003–2014

Abstract

NYPD stop and frisk policy, one of the more contentious policing issues of the past decade, has essentially come to a formal end. Initial data are overviewed relative to stops, frisks, and crime outcomes over time at the city level between 2003 and 2014. The stop and frisk policy literature relative to effectiveness and equity is also reviewed to provide context. Using structural equation modeling to build and test a theoretical model of policy effectiveness, the data did not fit the model well. Some evidence is found that the policy was not effective in reducing crime at the precinct level year over year (just three of twenty relationships tested between police actions and five serious crime types reached significance with a negative sign). Equity was analyzed using decision tree regression/classification, While race was found to be a significant and important variable in the frisk decision (though less so over time), it was not in either the decision to arrest or issue a summons. This analysis of the stop and frisk outcomes supports previous research using new techniques, further calling into question the balance between the inevitable equity conflict of the policy with its effectiveness in reducing crime.

Keywords:

• NYPD stop and frisk
• effectiveness
• equity
• SEM
• decision tree classification

Notes

1. This case was appealed successfully, and the department settled afterward, including the use of a court-appointed NYPD monitor.