Evaluation, as currently practiced in vocational education, may not contribute to program improvement; in fact, it may be counterproductive. Rather than asking how evaluations can be improved, it might be better to ask if a relationship does, in fact, exist between evaluation and program improvement and, if such a relationship exists, what new assumptions and approaches might improve that relationship. Current activities center around the provisions in the Vocational Education Amendments of 1976 and subsequent guidelines, which provided for the assessment of a sample of the students enrolled in vocational education in quantitative terms related to student achievement measures. Although the conventional notion is that if more resources were available for evaluation, the findings would be more interpretable for action alternatives, an alternative view suggests that the relationship of social problem solving to evaluation is not clear. Ordinary knowledge, social learning, and interactive problem solving (such as through public hearings) may, in reality, be the basis of social problem solving. Some alternative suggestions for the assumptions that underlie present vocational evaluation practices include (1) abandoning the pursuit of evaluation activities that are purported to contribute to improvement, but do not; and (2) focusing on content and learner decisions at the local level. Further educational research should be conducted apolitically, with the understanding that its findings will take at least a generation to be reflected in changed practices and should include much social research. (KC)