On two-sample test for detecting differences in the IFR property of life distributions

Abstract

A test proposed for testing whether one distribution is more Increasing Failure Rate (IFR) than another, based on a measure of IFR is presented in this paper. The asymptotic normality of the proposed test statistic was also established. The asymptotic null variance from the data was estimated since the variance depends on the unknown distribution. The Pitman asymptotic efficacies of the proposed test statistic are computed for various alternative IFR distributions.   Key words: Two-sample test, IFR, Pitman efficacy.