Several New Approaches to Life Testing Ibrahim A. Ahmad University of Central Florida Abstract: The classical approach to test that a life is ageless (exponential) against that it belongs to one of the many ageing classes in life testing depends on the empirical distribution and often leads to a form of U-statistics. These U-statistics are often not easy to calculate by hand or have low efficiencies or power or both. Hence few of these procedures found their way to textbooks in nonparametric statistics. In this talk, three new approaches are illustrated that would simplify the structure of the test statistics or improve the efficiency or the power. These approaches are called: (i) goodness of fit, (ii) moments inequalities, and (iii) kernel methods. Some new avenues of research are also illustrated.