The specific aims will include: 1. Analysis of Censored Data. The paradigm of multiple imputation-estimation will be used to address: a) the estimation of the hazard function in the presence of right and/or left censoring, truncation and interval censoring; b) the analysis of time to onset of disease detectable only at death in survival/sacrifice experiments; c) an approach for treating ties due to grouping in the Cox model; and d) the regression analysis of lifetime data under a variety of censoring schemes when one does not assume any particular parametric family of survival distributions. 2. A. Calculation of the Posterior Distribution (Missing Data). Tanner and Wong introduced the IE algorithm for the calculation of the posterior distribution in the presence of missing data.
This specific aim will be devoted to improve the efficiency of the IE algorithm and to apply the algorithm to various statistical problems which will arise in collaboration with investigators at the WCCC. One important application is the analysis of designed experiments with missing data. B. Calculation of the Posterior Distribution (No Missing Data). Tanner and Wong introduced a rejection algorithm for the calculation of posterior distributions in the multinomial case.
This specific aim will be devoted to extend the applicability of the algorithm, to improve the efficiency of the algorithm and to apply the algorithm to various statistical problems in cancer research. 3. Modeling Agreement/Disagreement Among Raters. Tanner and Young introduced an approach for modeling agreement/disagreement among raters. In the proposed research, model-based measures of agreement will be developed and investigated.
