Abstract
A class of non-parametric tests suggested by martingale theory is proposed for testing the equality of failure rates in a competing risks model with k independent risks, when data are the causes of failure and the observed times of failure. It generalizes the result of Yip and Lam (1992) and includes some existing tests when an appropriate weight function is chosen. Choices of the weight functions giving optimal local asymptotic power are discussed with examples. We also consider the performance of the test statistics by simulation for different weight functions. An example is given to demonstrate the sensitivity of the tests to the choice of the weight function.

This publication has 8 references indexed in Scilit: