Intervention research: GAO experiences.

This paper describes tools of program evaluation that may prove useful in conducting research on occupational health and safety interventions. It presents examples of three studies conducted by the U.S. General Accounting Office that illustrate a variety of techniques for collecting and analyzing data on program interventions, including analysis of extant data, synthesis of results of existing studies, and combining data from administrative files with survey results. At the same time, it stresses the importance and difficulty of constructing an adequate "theory" of how the intervention is expected to affect outcomes, both for guiding data collection and for allowing adequate interpretation of results.