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

Predicting an Offender's Characteristics: An Evaluation of Statistical Modelling

NCJ Number
163681
Author(s)
C Aitken; T Connolly; A Gammerman; G Zhang; D Oldfield
Date Published
1995
Length
60 pages
Annotation
The potential of offender profiling to aid police investigations of serious crime was analyzed using data from a sample of 320 solved cases of sexually oriented homicides of children in Great Britain.
Abstract
The analysis revealed a clear tradeoff between the detail of predictions that can be made about an offender and the reliability of those predictions. Only general offender characteristics could be predicted independently with an acceptable level of reliability. More detailed predictions will probably be possible when larger data sets designed specifically for this type of crime analysis become available. The research demonstrated how statistical models can be developed from an analysis of solved cases to predict characteristics of the offender of a new case. Findings indicated that data need to be consistently coded, the sample needs to be sufficiently large and representative of the offender population of interest, and appropriate statistical techniques must be used. The police also need to understand the reliability of statistical models to decide how to respond to the predictions. Figure, tables, appended tables, and 30 references