Statistical hypothesis test for the difference between Hirsch indices of two pareto-distributed random samples

Abstract

In this paper we discuss the construction of a new parametric statistical hypothesis test for the equality of probability distributions. The test bases on the difference between Hirsch's h-indices of two equal-length i.i.d. random samples. For the sake of illustration, we analyze its power in case of Pareto-distributed input data. It turns out that the test is very conservative and has wide acceptance regions, which puts in question the appropriateness of the h-index usage in scientific quality control and decision making. © 2013 Springer-Verlag.