Alternative Measures of Crime - A Comparison of the Uniform Crime Reports and the National Crime Survey in the Twenty-Six American Cities (From Crime - A Spatial Perspective, P 77-92, 1980, Daniel E Georges-Abeyie and Keith D Harries, ed. - See NCJ-74011)

• NCJ Number: 74016
• Author(s): J F Nelson
• Date Published: 1980
• Length: 16 pages
• Grant Number(s): 75-SS-99-6029
• Sponsoring Agency: National Criminal Justice Information & Statistics Service; Hindelang Criminal Justice Research Ctr
• Sale Source: Columbia University Press
• Publication Type: Statistics
• Language: English
• Country: United States of America
• Note: Prepared for presentation at the annual meeting of the American Society of Criminology in Dallas, Texas, November 1978.

Annotation

The issue of whether the National Crime Survey (NCR) and the Uniform Crime Reports (UCR) measure similar amounts of criminal activity is examined through a comparison of the UCR and the NCS in 26 American cities.

Abstract

Both the NCS and the UCR estimate the number of criminal acts committed in different places and times. These estimates are based on methodologies designed for different purposes. The UCR is designed to measure the extent of reported crimes. The NCS is a multipurpose project that, among other things, is designed to produce time series data on crime. The question of whether the NCS and the UCR measure similar amounts of criminal activity needs to be answered, because NCS data are frequently used to estimate how demographic characteristics affect the chances of a person being victimized. If the UCR and the NCS measure the same underlying crimes, then patterns found in the NCS data should also underlie the UCR crime rates; however, if the UCR and the NCS rates do not measure the same basic criminal activity, then it would be inappropriate to use the NCS data to make inferences about UCR rates. In this case, the reliability of either or both the UCR and NCS rates would be uncertain. The analysis of how the UCR is related to the NCS begins by describing the NCS methodology. Differences between reporting procedures and sources of bias are discussed. NCS rates are then correlated with UCR rates for seven crimes. These correlations show that the NCS rates are closely related to UCR rates for theft, but not personal crimes. The analysis is concluded by regressing the UCR and NCS rates on ecological variables. These regressions show that the relationship of ecological variables to crime rates frequently depends on whether UCR or NCS rates are used to measure crime. Tabular data and 7 notes are provided. For related papers, see NCJ 74011. (Author abstract modified)