General Departmental Seminar Series
Exact Two-Sample Inference with Missing Data
Ken Cheung, PhD,
Dept. of Biostatistics, Columbia University
November 5, 2004, 12 - 1 pm in room G5/113 Clinical Sciences Center (600 Highland Ave.)
When comparing follow-up measurements from two independent populations, missing records may arise due to censoring by events whose occurrence is associated with baseline covariates. In these situations, inferences based only on the completely followed observations may be biased if the follow-up measurements and the covariates are correlated. In this talk, I will discuss exact inference for a class of modified U-statistics under covariate-dependent dropouts. The method involves weighing each permutation according to the retention probabilities, and thus requires estimation of the missing data mechanism. The proposed procedure is nonparametric in that no distributional assumption is necessary for the outcome variables and the missingness patterns. A connection between the procedure and the bootstrap is noted. The Gibbs sampler is used to sample from the null permutation distribution, and is shown to be fast and accurate via simulation. A small data set with one follow-up time point is given as illustrative example, although the method can be easily applied to longitudinal data with arbitrary measurement times. Several possible extensions of the method will be discussed.
Return to seminar list