U.S. flag

An official website of the United States government, Department of Justice.

NCJRS Virtual Library

The Virtual Library houses over 235,000 criminal justice resources, including all known OJP works.
Click here to search the NCJRS Virtual Library

Use of Multiple Indicators to Estimate Crime Trends in American Cities

NCJ Number
145939
Journal
Journal of Criminal Justice Volume: 21 Issue: 5 Dated: (1993) Pages: 509- 516
Author(s)
R J Bursik Jr; H G Grasmick
Date Published
1993
Length
8 pages
Annotation
This note argues that the relatively recent availability of calls-for-service data in many jurisdictions provides an opportunity for a multiple-indicator estimation of these trends at the local level; this approach is illustrated with data collected over a 22-month period in Oklahoma City.
Abstract
Generally, when a researcher cannot measure a concept directly with a relatively high degree of reliability, multiple indicators of that concept are incorporated into the analysis. This approach attempts to minimize the effects of measurement error. Researchers have used this method in the attempt to identify U.S. national crime trends through the simultaneous analysis of the Uniform Crime Reports (UCR) and National Crime Survey (NCS) statistics. The findings of these multiple-indicator studies, however, have little relevance for those criminal justice personnel who are charged with estimating crime trends at the local level. The records of calls for service maintained by many metropolitan police departments are alternative indicators of the crime rates. Calls-for-service data can be used in conjunction with UCR statistics to estimate local crime trends based on multiple indicators. Calls-for-service data have several advantages over the UCR. First, the "gatekeeping" procedures used by police departments in recording citizen complaints are largely bypassed, since computers maintain running records of the number and nature of all calls. Second, the interviewer and memory effects endemic to the NCS are also absent. Finally, since all calls received by a police department become part of the dataset, sampling effects do not exist. 4 tables and 23 references