Assessing the Extent of Nonresponse Bias on NIBRS Estimates of Violent Crime

• NCJ Number: 221799
• Journal: Journal of Contemporary Criminal Justice Volume: 24 Issue: 1 Dated: February 2008 Pages: 32-49
• Author(s): Lynn A. Addington
• Date Published: February 2008
• Length: 18 pages
• Publication Format: Article
• Publication Type: Research (Applied/Empirical)
• Language: English
• Country: United States of America

Annotation

This study used NIBRS data from 2002 and 2003 to assess the amount of nonresponse in NIBRS, and to asses the effect of nonresponse in NIBRS on estimating violent crime rates, generating change estimates for violent crime, and using these data for analytical purposes.

Abstract

This study found that the overall response rate of agencies reporting NIBRS data was 46 percent, with the rates varying from 52 percent for cities with populations of less than 10,000 to 16 percent for cities with populations greater than 250,000. The highest response rate was observed among the smallest cities and the suburban and rural counties, which is consistent with the overrepresentation of small agencies in NIBRS. The study found a possible bias for the small response rate of larger cities due to the small sample size, as few law enforcement agencies serve cities with populations of more than 250,000. The effect of nonresponse was also examined and was found to have less of an effect on annual violent crime rates as compared to the change estimates. The findings show that for annual violent crime rates, the low response rates did not generate a consistent amount of bias but rather variations across crime types and subnational population groups. The study also found that the nonresponse in NIBRS introduced bias for generating change estimates, especially with regard to the magnitude of the change. As for using the data for analytical purposes, the study found that the agencies reporting to NIBRS were relatively similar to agencies reporting to the full Uniform Crime Report with regard to population characteristics related to crime. Additional research is needed in this area, especially at the county level. This study used NIBRS data from 2002 and 2003 on murder, forcible rape, and aggravated assault to examine the amount of nonresponse in NIBRS and to determine the biases generated by the level of nonresponse. To measure nonresponse and bias, the study compared national and subnational population groups. Tables, notes, and references