School cost functions: A meta-regression analysis

The education cost literature includes econometric studies attempting to determine economies of scale, or estimate an optimal school or district size. Not only do their results differ, but the studies use dissimilar data, techniques, and models. To derive value from these studies requires that the estimates be made comparable. One method to do this is meta-regression analysis (MRA) which was pioneered by Stanley and Jarrell [(1998). Gender wage discrimination bias? An MRA. Journal of Human Resources, 33, 947–973] as a result of similar frustrations with ‘omnipresent biases’ (p. 717). In this paper, the basic technique of MRA is described and then applied to 22 estimates of school costs, where an optimal school size (OSS) has or can be calculated. Results suggest an OSS of around 1543 students at the US secondary school level. Difficulties in interpreting the estimates are highlighted. The paper concludes with cautions in the use of MRA and opportunities for further research in this area.