Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
The empirical applications of the metafrontier analysis have largely been popularized by the publications of Battese et al. (2004), and O’Donnell et al. (2008). Prior to the estimation of the metafrontier function, it is generally suggested to conduct a likelihood-ratio (LR) test to test if all group frontiers are the same. In this short note, we point out that although the null hypothesis of the suggested LR test restricts all frontiers to be the same for all groups, the heterogeneity in the error distribution may exist and should be considered. We propose a restricted LR test as an alternative procedure.