Self-reported Delinquency as an Evaluation Measure - Comparative Reliability and Validity of Alternative Weighting Schemes

• NCJ Number: 74683
• Journal: Criminal Justice and Behavior Volume: 7 Issue: 4 Dated: (December 1980) Pages: 369-386
• Author(s): C H Blakely; M G Kushler; J A Parisian; W S Davidson
• Date Published: 1980
• Length: 18 pages
• Grant Number(s): MH29160
• Sponsoring Agency: US Dept of Health, Education, and Welfare
• Publication Format: Article
• Publication Type: Report (Study/Research)
• Language: English
• Country: United States of America
• Note: Based on a paper presented at the 87th Annual Convention of the American Psychological Association, New York, New York, 1979.

Annotation

A self reporting instrument designed to measure delinquent behavior was evaluated through its administration to delinquency intervention project participants; the effect of weighting schemes on the applicability and usefulness of the instrument, was examined.

Abstract

The instrument's is 35 questions collected information for program evaluation purposes on a wide range of delinquent behaviors. The weighting schemes relate to incidence of behavior (one, two, or no occurrences) measured over three different time intervals (the last 6 weeks, the last year, or more than the last year). The self-reporting instrument was administered, as part of a process interview package, to 135 youth at 4 points in the 18-week intervention project. Archival records were used for validation. Little difference was noted in the weighting schemes which resulted from different combinations of incidence and time intervals. The frequency weighting scales were more reliable than a weighting scheme involving categorization of behaviors into property, force, school, and substance abuse subscales. The behavior categories were examined by multitrait-multimethod analysis with the substance abuse subscale eliminated. The resulting correlation matrices did not neatly fit the ideal convergent and discrimination property patterns, although they supported the suggestion of convergent validity. Significant relationships were found between all of the behavior subscales and court and police records. Results indicated that delinquency intervention programs need not utilize weighting schemes in evaluation efforts, because these schemes do not strengthen evaluation instruments. Categorical responses to self-reporting questionnaires provide sufficiently reliable information to make them satisfactory as a sole evaluation instrument. Notes, references, and tabular data are included.