class size study
Research approach and methods
Deciding that research was required and fundable, and deciding on the research aims, was a big first step but what kind of study would be most effective? In this chapter I describe the reasoning behind our research approach. As we shall see, this is informed by lessons that we drew from previous research on class size effects. I shall show how it differs from previous work and identify its main features. The chapter ends with a description of schools, classes and children involved in the study and the methods of data collection. While this chapter could be omitted by those readers who want to move immediately to the results, it will provide information to better understand and make judgements about the findings. At the very least the reader may want to refer back to this chapter to access information on the nature of the data we collected.
In several reviews my colleagues and I have critically examined previous research (Blatchford and Mortimore 1994; Blatchford and Martin 1998; Blatchford et al. 1998). Though rather simplistic, it is helpful to categorize previous research into two main types: cross-sectional, correlational studies and experimental studies.
Perhaps the most obvious research approach is to look at associations between a measure of class size or pupil–teacher ratios on the one hand and measures of pupil attainment on the other. If we find that children's attainment tends to be better when they have been educated in smaller classes this would seem to suggest that smaller classes have helped. Unfortunately, research that has used this approach, even when large numbers of schools are involved, has