Meta-regression analysis of higher education productivity growth studies.

Abstract

This study provides a meta-regression analysis (MRA) of studies measuring productivity in higher education. We consider a total of 31 published and unpublished studies (working papers, theses e.t.c), written or published until April 2013, from which we finally use 28, yielding a total of 1018 observations. We examine the impact of different study characteristics on the estimated rate of productivity change and in addition we check the robustness of our findings using three alternative estimation methods for MRA. The first uses equally weighted observations, implicitly giving more weight to studies with multiple TFP estimates, the second uses per paper weights, resulting in equal weight in cross-paper observations, and the third uses one observation per paper. The results indicate that there are actual differences between the three different weighing methods, in the MRA estimates.