NCJ Number
16430
Editor(s)
H W RIECKEN,
R F BORUCH
Date Published
1974
Length
357 pages
Annotation
THE USE OF RANDOMIZED, CONTROLLED EXPERIMENTS TO PLAN, DEVELOP, AND APPRAISE INNOVATIVE PROGRAMS.
Abstract
THE POSITION TAKEN IS THAT SYSTEMATIC EXPERIMENTAL TRIALS OF PROPOSED SOCIAL PROGRAMS HAVE CERTAIN IMPORTANT ADVANTAGES OVER OTHER WAYS OF LEARNING WHAT PROGRAMS (OR PROGRAM ELEMENTS) ARE EFFECTIVE, UNDER WHAT CIRCUMSTANCES, AND AT WHAT COST. THE IMPORTANCE OF RANDOM ASSIGNMENT OF STUDY SUBJECTS TO EXPERIMENTAL OR CONTROL GROUPS (TO PREVENT THE INTRODUCTION OF POSSIBLE BIAS) IS EMPHASIZED. THE USE OF A QUASI-EXPERIMENTAL DESIGN IS SUGGESTED WHEN EITHER RANDOMIZATION CAN NOT BE ACHIEVED OR WHEN SETTING UP A CONTROL GROUP IS NOT FEASIBLE. IN THIS WAY, THE EXPERIMENTER CAN APPROXIMATE EXPERIMENTAL PROCEDURES FOR COLLECTING DATA EVEN THOUGH HE LACKS FULL CONTROL OVER THE DELIVERY OF THE TREATMENT. THE ADVANTAGES AND DISADVANTAGES OF SOCIAL EXPERIMENTATION ARE ALSO CONSIDERED. THE APPENDIX LISTS REFERENCES TO AND ABSTRACTS OF ILLUSTRATIVE, RANDOMIZED EXPERIMENTS FOR APPRAISING THE EFFECTS OF SOCIAL PROGRAMS.