The proposition that a more educated person is more productive is so well entrenched that few have sought to question it. This paper synthesises IDS research in Mexico, Sri Lanka and Ghana which has attempted to test the proposition, and draws on research from elsewhere. The IDS data suggests that differences in educational level (across a span of five years of education) appear to be unrelated to supervisors' ratings of individual productivity among groups of people doing the same job. The treatment of the education variable in this research was conventional: education was measured in terms of years of schooling. A more quantitative approach to the educational variable in future research of this kind is recommended. Such an approach might have the effect of turning the attention of manpower and educational planners towards qualitative improvement of their systems.