RECIDIVISM COMPARISONS ACROSS GROUPS - METHODS OF ESTIMATION AND TESTS OF SIGNIFICANCE FOR RECIDIVISM RATES AND ASYMPTOTES

• NCJ Number: 55196
• Journal: EVALAUTION QUARTERLY Volume: 3 Issue: 1 Dated: (FEBRUARY 1979) Pages: 105-117
• Author(s): M R LLOYD; G W JOE
• Date Published: 1979
• Length: 13 pages
• Corporate Author: Sage Publications, Inc
• Publication Format: Article
• Publication Type: Program/Project Evaluation
• Language: English
• Country: United States of America

Annotation

DATA FROM A FOLLOWUP STUDY OF DRUG TREATMENT PROGRAM CLIENTS ARE USED TO DEVELOP A METHOD FOR ESTIMATING RECIDIVISM.

Abstract

THE METHOD BUILDS ON A MODEL BY MALTZ AND MCCLEARY THAT MEASURES TWO DIMENSIONS OF RECIDIVISM: THE RECIDIVISM RATE WITHIN A FIXED TIME PERIOD, AND THE ULTIMATE PROPORTION OF CLIENTS WHO WILL RECIDIVATE. TO OBTAIN MAXIMUM LIKELIHOOD ESTIMATES AS DESCRIBED IN THE MODEL FOR BOTH THESE DIMENSIONS, DATA ON 1,523 MALE CLIENTS IN FIVE TREATMENT GROUPS OF A TEXAS DRUG PROGRAM ARE ANALYZED. RECIDIVISM WAS DEFINED AS RETURN TO DAILY DRUG USE OR TREATMENT. INITIAL RECIDIVISM RATES AT 6-MONTH FOLLOWUP AND SUBSEQUENT RATES AT 6-MONTH INTERVALS OVER A 3-YEAR PERIOD ARE USED TO CONSTRUCT CURVES FROM WHICH MAXIMUM LIKELIHOOD ESTIMATES OF THE ULTIMATE PROPORTION OF FAILURES FOR EACH TREATMENT GROUP CAN BE DETERMINED. THE ADVANTAGES OF THE PROPOSED METHOD ARE THAT IT MAKES USE OF ALL AVAILABLE DATA, DOES NOT REQUIRE UNIFORM FOLLOWUP PERIOD FOR ALL CLIENTS, AND USES THE UNIQUE FEATURES OF MAXIMUM LIKELIHOOD ESTIMATES TO MAKE INTERGROUP COMPARISONS. FULL DETAILS OF THE METHOD, INCLUDING MATHEMATICAL FORMULATIONS AND TESTS OF SIGNIFICANCE, ARE INCLUDED, TOGETHER WITH A LIST OF REFERENCES. (LKM)