REGRESSION AND SELECTION MODELS TO IMPROVE NONEXPERIMENTAL COMPARISONS (FROM EVALUATION AND EXPERIMENT, 1975, BY C A BENNETT AND A A LUMSDALE)

• NCJ Number: 36929
• Author(s): G G CAIN
• Date Published: 1975
• Length: 21 pages
• Corporate Author: Academic Press, Inc
• Publication Format: Document
• Language: English
• Country: United States of America

Annotation

THE AUTHOR ARGUES FOR THE USEFULNESS AND VALIDITY OF ECONOMETRIC AND RELATED NONEXPERIMENTAL APPROACHES FOR ASSESSING THE EFFECTS OF SOCIAL PROGRAMS.

Abstract

USING A REGRESSION APPROACH, THE AUTHOR PRESENTS A MODEL FOR PRODUCING UNBIASED TREATMENT EFFECTS IN EXPERIMENTAL SITUATIONS EVEN WHEN THE SELECTION PROCESS FOR TREATMENT IS NONRANDOM. THE AUTHOR CONTENDS THAT RANDOMIZATION IS NOT ESSENTIAL; INSTEAD, HE ARGUES THAT THE CRITICAL DIFFERENCE FOR AVOIDING BIAS IS NOT WHETHER THE THE EXPERIMENTAL ASSIGNMENTS ARE RANDOM OR NONRANDOM, BUT WHETHER THE INVESTIGATOR HAS KNOWLEDGE OF AND CAN MODEL THIS SELECTION PROCESS. STRATEGIES ARE THEN PRESENTED FOR OBTAINING UNBIASED ESTIMATES OF PARAMETERS OF INTERESTS FROM NONEXPERIMENTAL DATA.