The Gini Test for Survival Data in Presence of Small and Unbalanced Groups

Abstract

The aim of this note is to study the performance of the Gini concentration test for survival data in presence of unbalanced and small samples. We compared the performance of the asymptotic test with an alternative permutation distribution test, illustrating by simulation that if groups are very small the latter test should be used. Also, we show how the definition of the length of time considered in the construction of the test statistic can be chosen to improve the performance of the test.