NCJ Number
84898
Date Published
1969
Length
25 pages
Annotation
Types of evaluative research designs for corrections programs are described, and policy considerations in selecting research designs are considered.
Abstract
A number of research designs are available for use in evaluative studies. One of the most common forms of research is the survey, such as that used in compiling the Uniform Crime Reports. Questionnaires, psychological tests, and interviews have also been used to measure attitudes, abilities, and behavior. The Federal funding of programs of crime prevention is increasingly requiring evaluations of an experimental nature, which tend to use a classical research design that exposes one group (experimental) to the new treatment program, while another group (control) is not exposed to the new program. One of the best ways to ensure comparability of the two groups is to use random assignment procedures. Other methods of determining that observed changes in outcome are the results of a treatment program include (1) the time-series with a comparison group, (2) the multiply-interrupted time series, (3) the treatment group as its own control, and (4) statistical procedures of control. The primary policy consideration in evaluation reseach is to avoid unnecessary harm to the participants while using research procedures that minimize the ambiguity of inferences from the study. Much is now being done in laboratory settings and in pilot projects outside of corrections which may be applicable to correctional programs. Procedures such as operant or classical conditioning, sleep learning, perceptual restructuring, systematic desensitization, modeling, self-suggestion, and electronic monitoring or prompting are being studied. Forty-three footnotes are listed.