Field Testing : Methodological Considerations and a Specific Example.
The "classical" techniques of experimental design as taught at graduate school usually have to be seriously modified in the environment of industry or government. This paper reports on the successful application of a relatively well-controlled experimental design in an operating Air Force environment and elaborates on the modifications that were necessary. Attention is focused on considerations of decision alternatives and their associated errors, and the consequent problem of sample size. The tradeoffs between classical experimental design and more loosely controlled designs, and the requirements for each are explored. A discussion is included of approaches to decisionmaking other than that allowed by classical experimental design. 18 pp.