Abstract
In this paper we introduce the sequential precedence test (SPT) for testing the equality of two continuous distribution functions, against the stochastically ordered alternative. This procedure replaces the classical precedence test (PT), used in life-testing experiments, by a sequence of tests which are applied at the failure times of one of the samples. This allows the possibility of stopping the experiment earlier than the PT. By means of extensive Monte Carlo simulations and real data, we show that the proposed methodology results in substantial saving of experimental time and cost, without compromising in power. Algorithms for the implementation of the SPT are included.
Acknowledgements
We thank two anonymous referees for their careful reading of the manuscript and for suggestions that helped us to improve the paper. This research was supported by PFB-03-CMM project, Fondecyt grant 1120408 and project MTM2010-15972 of MICINN and FEDER. The last three authors are members of the research group Modelos Estocásticos (DGA).