ERIC Number: ED196803
Record Type: RIE
Publication Date: 1980-May
Reference Count: 0
Social Experimentation: Some Whys and Hows.
Archibald, Rae W.; Newhouse, Joseph P.
This report consists of three parts: when to conduct a social experiment (and when not to), how to manage a social experiment, and some practical advice to experimenters. Social experimentation compares people who receive a certain program (experimental treatment) with similar people who do not or who receive some other treatment. Some form of randomization is typically used to maximize the likelihood that the various groups are in fact similar. The experimental method is the best available for establishing that a certain program or intervention actually caused a given result or set of results. Other methods often encounter difficult problems which are discussed in the report. Obstacles may also impede or even preclude social experimentation. Perhaps, most important, it is usually more expensive and time consuming to collect and analyze experimental data than it is to use existing data. Consequently, the experimenter should be sure that existing data will not suffice for the purpose before deciding to go ahead with an experiment. In the management of an experiment, planning and cost estimation are both important skills. Early and frequent interaction between research teams is necessary for success. Also, the data processing group should participate in the early stages of design if its operations are to be efficient and effective. Several tips are given to the prospective experimenter. Create a pilot sample of people with whom to pretest the operational feasiblity of the experiment. Build into the design an ability to measure effects that are an artifact of the experiment. Construct the experiment to keep refusal and attrition at low levels. (Author/RM)
Rand Corporation, 1700 Main Street, Santa Monica, CA 90406 ($3.00).
Publication Type: Opinion Papers
Education Level: N/A
Sponsor: Department of Health, Education, and Welfare, Washington, DC.
Authoring Institution: Rand Corp., Santa Monica, CA.