LAUSR

Abstract Meta-analysis is a statistical method used to make a systematic review of the literature to integrate the results of a series of studies. It is increasingly adopted in social sciences but according to our best knowledge used for the first time to aggregate and contrast findings on rail transport efficiency. The experiment adopted a permutation test to evaluate the influence of variables discussed in the literature in the mean efficiency scores. The results suggest that railways located in Japan and in the US have characteristics that push them toward increasing efficiency. The passenger rail systems reached significantly higher estimates than conventional cargo systems. Estimates from parametric and nonparametric models showed significant difference, while from nonparametric models including Data Envelopment Analysis (DEA) and from Network DEA did not. The number of variables and the ratio between the number of decision making units and the number of variables employed significantly influenced the scores. Unexpectedly, different data structures did not. Validation methods are presented. Public policies based on the empirical results are commented.