Where does good evidence come from?

Abstract

This article started as a debate between the two authors. Both authors present a series of propositions about quality standards in education research. Cook's propositions, as might be expected, not only concern the importance of experimental trials for establishing the security of causal evidence, but also include some important practical and acceptable alternatives such as regression-discontinuity analysis. Gorard's propositions, again as might be expected, tend to place experimental trials within a larger mixed-methods sequence of research activities, treating them as important but without giving them primacy. The article concludes with a synthesis of these ideas, summarising the many areas of agreement and clarifying the few areas of disagreement. The latter include what proportion of available research funds should be devoted to trials, how urgent the need for more trials is and whether the call for more truly mixed-methods work requires a major shift in the community.