Testing to Determine the Underlying Distribution Using Randomly Censored Data

Publisher Website
Google Scholar
Abstract
In the classical goodness-of-fit problem, a random sample of X from a population with distribution function F is observed. The null hypothesis, which is to be tested using the random sample, asserts that F is equal to G, where G is a completely specified distribution function. There is a need to generalize this problem to encompass censored data because in various situations, such as clinical trials, the X may represent times to the occurrence of an endpoint event and the data are usually analyzed before all [human] patients have experienced the event. For right-censored data, a goodness-of-fit procedure is developed for testing whether the under lying distribution is a specified function G. The test statistic C is the 1-sample limit of Efron''s (1967) 2-sample statistic W. The test based on C is compared with recently proposed competitors due to Koziol and Green (1976) and Hyde (1977). The comparisons are on the basis of applicability, the extent to which the censoring distribution can affect the inference and power. In certain situations the C test compares favorably with the tests of Koziol-Green and Hyde.

This publication has 0 references indexed in Scilit: