The present research compared ANCOVA and two alternative analytic approaches (the LISREL V computer program and a factor-loglinear-regression approach) when applied to two data sets typical of criminal justice evaluations. One set was artificial, constructed from a larger data set generated during an evaluation of California youth service bureaus. The second data set came from an experimental study of an institutional program within the California Department of the Youth Authority. Outcome was a measure of subsequent official delinquency or crime. Background variables included measures of prior police contacts or arrests, demographic information, and scales measuring preexisting behavioral traits. The two alternative analytical approaches improved estimates slightly and provided information not obtained with ANCOVA. However, the three estimates did not differ substantially. The complexity or unavailability of the alternative methods make them less practical for most evaluation studies using criminal justice data. Data tables, figures, a list of 40 references, and appendixes presenting the questionnaires and related information are supplied.
Downloads
Similar Publications
- The Effects of a Co-Response Program on Patrol Call Volume for Mental Health Crisis-Related Calls: A Time Series Analysis
- The relative and joint effects of gunshot detection technology and video surveillance cameras on case clearance in Chicago
- Recreational Cannabis Legalization in Washington State: Residents' Opinions and Perceptions of Effects Five Years after Implementation