Reliability of Police Employee Counts: Comparing FBI and ICMA Data, 1954-2008

• NCJ Number: 236377
• Journal: Journal of Criminal Justice Volume: 39 Issue: 5 Dated: September/October 2011 Pages: 445-451
• Author(s): William R. King; Abdullah Cihan; Justin A. Heinonen
• Date Published: 2011
• Length: 7 pages
• Publication Format: Article
• Publication Type: Test/Measurement; Report (Study/Research)
• Language: English
• Country: Netherlands
• Note: A version of this paper was presented at the Midwestern Criminal Justice Association annual meeting, September 30, 2004.

Annotation

This study compared the reliability of annual counts of the number of police officers and civilians employed by police agencies as reported by the FBI's Police Employees data and the International City/County Management Association's (ICMA) Municipal Year Book reported for 38 large city police agencies between 1954 and 2008.

Abstract

The Police Employees (PE) data are agency-level data collected by the FBI as part of the Uniform Crime Reports (UCR) data collection. Such data have been collected annually from U.S. police agencies since 1930. Since 1936, the ICMA has surveyed municipalities with more than 10,000 inhabitants about various components of local government, including finances; government type; and information related to police, fire, and refuse departments. Results are published annually in the Municipal Year Book. Overall, this study found that correlations for each type of employee count were statistically significant across all but two cities, suggesting that the PE and Year Book data are remarkably reliable. A closer examination of data from the two cities that are the exception (Jersey City and Pittsburgh) shows temporally bounded reporting irregularities in employee counts that may account for their statistically insignificant correlations; however, occasional data irregularities were found even in cities that report reliable employee counts. The authors conclude that researchers planning to analyze employee counts in either the PE or ICMA data can proceed fairly confidently, but should still consider the reliability of data for certain years and cities in each data set, respectively. For analyses that use a large number of agencies in a comparative framework, the few reporting errors are unlikely to be significant. The study used bivariate correlations and visual analyses of line charts in order to demonstrate the reliability of these two datasets, particularly examining situations in which inaccuracies emerged. 2 tables, 3 figures, 7 notes, and 41 references