INTRODUCTIONThis is primarily a paper on methodology and, as such, it will not discuss substantive findings of research on teaching styles. However, research methodologists are frequently accused of being more concerned with tidiness of design and sophistication of analysis than with the real problem under study. Such criticisms are not without foundation. In some cases, it has been claimed (Westbury, 1970) that methodologists have even changed the nature of problems by imposing sophisticated analyses on data sets. Whether by 'design' or 'accident', this level of intrusion by methodology needs to be resisted by researchers. The research problem must remain in central focus; prescriptions for design and analysis should serve to make that focus as sharp and distinct as possible.For this reason, I should like to begin this paper by describing the kind of problem I have in mind when I speak of research on teaching styles. With this problem as focus, I should then like to examine the strengths and weaknesses of possible research designs and analyses.The discussion presented here is restricted to those studies with a commitment to measurement (and hence quantification) of the educational environment, process, and outcomes. While acknowledging the legitimacy of other modes of inquiry applied to a study of teaching, this paper does not address those who, whether by persuasion or through frustration with empirical methods, prefer such approaches.
BACKGROUND TO THE PROBLEM