NCJ Number
101799
Journal
Journal of Quantitative Criminology Volume: 2 Issue: 1 Dated: (March 1986) Pages: 69-80
Date Published
1986
Length
12 pages
Annotation
This unsuccessful attempt to replicate M. Harvey Brenner's influential 1984 study of the relationship between homicide rate and economic conditions reveals a number of methodological problems with the study.
Abstract
Using data for 1951 to 1980, Brenner reported a positive relationship between the homicide rate and the ratio of between the homicide rate and the ratio of unemployment among 16- to 24- year-old males to the national unemployment rate and a negative relationship between homicide rate and annual change in disposable personal income. Although not obvious from his write-ups, the 1984 variables and the results differ substantially from those of Brenner's earlier 1976 study. Further, using the same data and procedures (as far as could be ascertained) as reported by Brenner, the present calculations show that economic indicators have little or no influence on homicide rate. Finally, there are a number of methodological problems with Brenner's regression specification. First, only two of seven variables are economic indicators: the remaining five are measures of drug and alcohol use. Second, neither of the economic indicators has a close statistical relationship to the unemployment rate. Finally, the specification selection procedure employed by Brenner has the potential of ruling out specifications simply because they do not confirm the author's prior beliefs about how economic conditions affect homicide rates. It is concluded that it would be inappropriate to base economic policy on Brenner's reported results. 2 tables and 11 references.