Class Size and Educational Achievement: a review of methodology with particular reference to study design[1]

Abstract
The article reviews research into class size effects from a methodological viewpoint, especially concentrating on the various strengths and weaknesses of randomised controlled trials and observational studies. It discusses population definitions, causation and generally sets out the criteria for valid inferences from such studies. For illustration it presents some new findings from a reanalysis of the large data set from the Tennessee STAR (Student Teacher Achievement Ratio) study.