NCJ Number
75508
Date Published
1970
Length
49 pages
Annotation
The effectiveness of a prediction instrument, consisting of a multiple regression equation with seven variables, was tested in connection with recidivism rate projections for female parolees in the California Youth Authority.
Abstract
A large sample (2,001) of the female wards released to parole in 1963 and 1964 was used as a construction sample in predicting recidivism within 15 months of release, and another large sample of 1,182 female parolees released in 1965 was the validation sample. The variables applied were: age of admission to the Youth Authority; age at release on parole; admission status; number of cooffenders in the commitment offense; and number of foster home placements. Age at release on parole contributed the most weight to the predicted score. The relationship between age and recidivism rates could be interpreted as an indicator of early development of delinquent patterns. Due to limitations in accuracy and relevance of the prediction tables, the information they were expected to provide did not materialize. They did not provide the information about the correlates of recidivism which Youth Authority personnel hoped to use in assessing changes in the population with respect to recidivism-related variables. Even as a statistical control for program evaluation and decisionmaking concerning offenders, the prediction tables described here showed little, if any, usefulness. Tabular data are included in the text and in one appendix. A list of 28 references is appended. (Author abstract modified).