Evaluating educational changes: A statistical perspective

Abstract

The paper explores some of the issues involved in evaluating educational policy initiatives. It gives examples of how research findings can be evaluated and draws lessons for the ways in which policymakers can interact usefully with researchers. It argues that while central government's use of research evidence is often highly selective and concerned with its own perceived short term interests, a broader view of the research process is more productive and beneficial. The issues of class size, school league tables and the effects of homework are studied in detail and the often provisional nature of research evidence is emphasised as well as the uncertainty surrounding the findings of individual studies.