Linking Serious Sexual Assaults Through Behaviour

• NCJ Number: 190290
• Author(s): Don Grubin; Peter Kelly; Chris Brunsdon
• Date Published: 2001
• Length: 80 pages
• Sale Source: Great Britain Home Office Research Development and Statistics Directorate
• Publication Format: Book (Softbound)
• Publication Type: Report (Study/Research)
• Language: English
• Country: United Kingdom
• Note: Home Office Research Study 215.

Annotation

This report presents the methodology and findings of a British study that examined the extent to which behavioral consistency was a regular feature of serial sex offending against females unknown to their attackers, with attention to how such an analysis of behavior might be used to link crimes.

Abstract

Data were obtained from a British database of serial stranger sexual assault as well as a similar Canadian database. The study broke down each rape offense into four domains: the offender's behavior in controlling the victim, the sexual behavior, the offender's concerns about escape, and the personal style of the attacker. Using cluster analysis, each of these domains was in turn categorized into four behavior types. Every offense was categorized according to its domain types, and the offenses of each serial offender were examined for their consistency. The study found that of the serial offenders in the British database, 83 percent were consistent in at least one of the four domains throughout their series, and 26 percent had at least two offenses within their series that matched across all four domains. Although it was not possible to make a direct comparison, the results for the Canadian cases were better than for the British data. On further examination, the researchers found that serial offenders were most consistent in the "control" domain, with the same domain type occurring 68 percent of the time in consecutive serial offenses. The "style" domain was the least stable. The study concluded that the findings would contribute to the development of a screening tool to help in the identification of similar cases, but such a tool would not provide a means of confirming linkage (or non-linkage) of offenses. The study recommended further research to determine those behaviors that best link offenses, as well as the interactions between behaviors. 23 tables, 4 figures, 32 references, and appended review of previous research, conventions to deal with data problems, and domain variables for British databases