Assessing the “Statistical Accuracy” of the National Incident-Based Reporting System Hate Crime Data

Abstract

The current study introduces a method to assess hate crime classification error in a state Incident-Based Reporting System. The study identifies and quantifies the “statistical accuracy” of aggregate hate crime data and provides insight from frontline officers about thought processes involved with classifying bias offenses. Random samples of records from two city and two county agencies provided data for the study. A systematic review of official case narratives determined hate crime classification error using state and federal definitions. A focus group sought to inquire about officers’ handling of hate crimes. Undercounting of hate crimes in official data was evident. When error rates were extrapolated, National Incident-Based Reporting System Group A hate crimes were undercounted by 67%. Officers’ responses validated complications involved with classifying hate crimes, particularly, incidents motivated “in part” by bias. Classification errors in reporting hate crimes have an impact on the statistical accuracy of official hate crime statistics. Officers’ offense descriptions provided greater awareness to issues with accurately interpreting and classifying hate crimes. The results yield useful information for officer training, understanding the true magnitude of these crimes, and a precursor for adjusting crime statistics to better estimate the “true” number of hate crimes in the population.