Longitudinal Perspective on Patterns of Crime

• NCJ Number: 79377
• Journal: Criminology Volume: 19 Issue: 2 Dated: (August 1981) Pages: 211-218
• Author(s): J McCord
• Date Published: 1981
• Length: 8 pages
• Grant Number(s): R01 MH26779
• Sponsoring Agency: US Dept of Health, Education, and Welfare
• Corporate Author: Massachusetts Commissioner of Probation
• Publication Format: Article
• Language: English
• Country: United States of America
• Note: An earlier version of this article was presented at the meeting of the American Society of Criminology, November 7-10, 1979.

Annotation

This article addresses the possibility of prospectively identifying chronic criminal offenders using identifiers available to the criminal justice system.

Abstract

Data were collected as part of a longitudinal study of the lives of men who, between 1939 and 1942, had been included in a delinquency prevention program in eastern Massachusetts. Known as the Cambridge-Somerville Youth Study, the program included boys born between 1925 and 1934 who lived in factory-dominated areas; boys had been assigned by chance to either a treatment or control group. The present followup began in 1975. The names and pseudonyms of all the men were checked through the Massachusetts Department of Probation's centralized records and by 1979, 98 percent of the men had been located. After elimination of some subjects for several reasons, 208 men were left to provide the basis for testing the possibility for prospectively identifying recidivists. Discriminant analysis was used to test the adequacy of two discriminating variables as a basis for prospective identification. By forming a linear combination of the variables, using weighting coefficients, the discriminant analysis maximized separation between recidivists and nonrecidivists along a dimension which combined age at first conviction and type of crime for which a man was convicted. Among the 111 former members of the treatment group who had been convicted for nontraffic crimes, 51 were recidivists. Thus, recidivists constituted 46 percent of the sample. The linearized discriminant function correctly identified 63 percent of the 51 recidivists and 70 percent of the desistors. Random predictions based on prior probabilities would be expected to identify 50 percent correctly as recidivists or desistors. Other results of the analysis are also discussed. The article concludes that although prospective identification of recidivists may be possible, legal action based on predictions would be wrong. Six references are provided.