Trends in Evaluation

GAO examined some of the progress made in the research of evaluation methodology, which appeared in the GAO Review, Vol. 16, Issue 3, Summer 1981. Much work remains to be done in the estimating of program and project cost. However, any particular method of evaluation is subject to pitfalls and biases which may be checked by using alternative methodologies. The use of alternative methodologies appears to offer great promise as guidance for future data systems design, particularly if evaluators would also report more specifically the cost of implementing various alternatives. Such reporting would give more precise estimates of cost/effectiveness of alternatives in the planning stage. It is also important to improve evaluation methodology applied to nonexperimental environments. Costly evaluations have repeatedly failed in these environments to prove whether the program in question had any significant effect. One reason may be that the studies have been conducted as independent efforts without incorporating prior knowledge into the design. Determined efforts are being made to construct flexible and adaptive procedures for designing and carrying out evaluations in statistical settings. Adaptive methods allow for the reallocation of evaluation resources to conserve resources and best meet the objectives of the study and are more suitable to the nature of management and policymaking in a political environment.