Tests of independence for bivariate data with random censoring a contingency table approach
Akritas, M.G.; Clogg, C.C.
Biometrics 47(4): 1339-1354
ISSN/ISBN: 0006-341X DOI: 10.2307/2532390
Procedures are proposed for testing the hypothesis of independence between two discrete or discretized random variables, one or both of which may be randomly censored. The censoring mechanisms can be either independent (bivariate censoring) or identical (univariate censoring). The proposed tests reduce to goodness-of-fit tests for log-linear models applied to either complete or incomplete contingency tables. Three examples are analyzed for illustrative purposes.