A Simple and More Efficient New Approach to Life Testing

摘要

Testing of various classes of life distributions has long been a subject of investigation. Most of these procedures, however, are not easy to grasp by non-specialized users of statistics. Additionally, not all of these procedures have good efficiencies. In the present investigation, a new and much simpler approach is presented for testing exponenti-ality (the common null hypothesis in life testing) against two perennial life classes, namely the class of increasing failure rate (IFR) distributions, the class of new better than used (NBU) distributions and the class of new better than used of specified age (NBU - t_s). This new approach is very simple and is also more efficient for some commonly used alternatives than older and more involved methodologies in the sense of Pitman's asymptotic relative efficiency.