Nonparametric Inference Based on Sampled Minima

Ibrahim A. Ahmad

A sampling plan that may find applications in economics, biomedical research, reliability and life testing consists of putting kn units on test in groups of k each and observing the minimum value of each group. Thus, the observations are sampled minima and they are to be used to do inference about the original distribution. In this investigation, we present properties of the empirical distribution and sample moments and show how these estimates can be used to do goodness of fit testing about the underlying distribution.