NCJ Number
231895
Journal
Journal of Quantitative Criminology Volume: 26 Issue: 3 Dated: September 2010 Pages: 333-349
Date Published
September 2010
Length
17 pages
Annotation
This study investigated the extent to which crime statistics conform to the Benford law (distribution of leading digits in data of an anomalous nature will conform to a formula of logarithmic intervals).
Abstract
Benford's law suggests that the distribution of leading (leftmost) digits in data of an anomalous nature (i.e., without relationship) will conform to a formula of logarithmic intervals known as the Benford distribution. Forensic auditors have successfully used digital analysis vis-a-vis the Benford distribution to detect financial fraud, while government investigators have used a corollary of the distribution (focused on trailing digits) to detect scientific fraud in medical research. This study explored whether crime statistics are Benford distributed. The authors examined crime statistics at the National, State, and local level in order to test for conformity to the Benford distribution, and found that National- and State-level summary UCR data conform to Benford's law. When National data were disaggregated by offense type we found varying degrees of conformity, with murder, rape, and robbery indicating less conformity than other offense types. Some tentative implications of these findings are discussed, as are areas for further research. Tables, figures, and references (Published Abstract)