Linear rank tests for comparing treatments with a control when data are subject to unequal patterns of censorship

Abstract

A class of asymptotically distribution‐free tests is considered for comparing several treatments with a control when the, data are subject to unequal right‐censorship. A particular member of this class is proposed for use in practice and an illustrative numerical example is given. A general result for the Pitman efficacy of a test based on an asymptotically normal test statistic is proved for the multiparameter case and using this result the efficacy of the proposed class of tests is obtained under sequences of translation and proportional hazards alternatives. Simulation studies are conducted to compare the performance of the proposed test, in terms of power, with some other tests.