ABSTRACT
Comparative life test experiments are of paramount importance when the object of a study is to ascertain the relative merits of two or more competing products in regard to the duration of their service life. This chapter presents methods for comparing the parameters of k exponential distributions based on complete or censored samples from each population. It focuses on comparisons which, in the setting of two-sample problems, involve testing the null hypothesis that there are no differences between the parameters of the two exponential distributions which model the life lengths of the two products under study. The chapter discusses the related issues of setting confidence intervals for the ratio of the scale parameters as well as for the difference between the location parameters. It examines the treatment of inference procedures when each sample is either fully observed or is Type II censored.
