NCJ Number
140143
Date Published
1991
Length
19 pages
Annotation
This discussion of social experiments and quasi- experiments in the field of drug prevention research identifies significant predesign issues, reviews concepts of causal inference and control, and outlines the basic types of validity and potential threats to validity. The chapter delineates selected experimental and quasi-experimental designs and discusses their advantages and drawbacks.
Abstract
The strength of any experimental or quasi-experimental prevention design depends on the careful consideration of several predesign issues: the problem to be prevented, the target population, the risk factors and associated mediating factors, the intervention, and the expected outcome and related evaluation criteria. Certain criteria must be met to infer the existence of a cause-effect relationship including, inter alia, covariation between the presumed cause and effect, relative temporal contiguity between cause and effect, and replicability. Four types of validity are used in connection with experimental and quasi-experimental research designs: statistical conclusion validity, internal validity, construct validity, and external validity. The authors identify six designs well suited for evaluating preventive interventions: (1) true experimental designs, (2) pretest/posttest control group design, (3) attention placebo six group design, (4) repeated measures design, (5) pretest/posttest nonequivalent control group design, and (6) simple and multiple time series design. 15 references