Missing Data and Imputation in the Uniform Crime Reports and the Effects on National Estimates

• NCJ Number: 221801
• Journal: Journal of Contemporary Criminal Justice Volume: 24 Issue: 1 Dated: February 2008 Pages: 69-85
• Author(s): Jame P. Lynch; John P. Jarvis
• Date Published: February 2008
• Length: 17 pages
• Publication Format: Article
• Publication Type: Research (Applied/Empirical)
• Language: English
• Country: United States of America

Annotation

This article complements previous research on the Uniform Crime Reporting (UCR) program by assessing the magnitude and distribution of missing data nationally and their effect on national-level and change estimates.

Abstract

In theory, the UCR program is based on a census of all police organizations, but in reality, not all police department submits data and in some areas, the submitted data is substantially incomplete. This analysis examined data submitted in Return A, Offenses Known to the Police data that is frequently used to describe the overall crime picture and includes the number of criminal offenses reported, the number of crimes unfounded, and the number "cleared" by arrest. The analysis found that the proportion of agencies reporting a complete 12 months of data decreased from 78 percent in 1992 to 65.5 percent by 2003. Most of the missing data resulted from agencies not reporting any data at all, as compared to failing to report data for a month or two. In addition, it was found that the missing data were not randomly distributed across the population of State and local police agencies. In cities with populations greater than 250,000, 95 percent of the agencies reported complete 12 month Return A data, in contract to cities with populations less than 25,000 where just 57.2 percent of the agencies reported complete 12 month data. It was also found that the extent of the missing data varied by type of police organization, and according to the social organization of the UCR program in the jurisdiction. Thus, if agencies that do not report are substantially different in important respects from agencies that do report, then missing data are more likely to affect the national-level and change estimates in important ways. The UCR program employs an imputation procedure to estimate the missing data; this procedure is described in detail in the article. Suggestions for future research on imputation methods are presented. Tables, notes, and references