Prof. Micha Mandel
Cross-sectional Sampling, Bias and Dependence
von 11:30 bis 13:00
|Wo||Eckerstr.1, Room 404|
Cross-sectional sampling leads to truncated and length-biased data. Methods for dealing with a random, possibly censored, sample of length-biased survival times are well developed. However, cross-sectioning often result in complicated bias function and dependence among observed survival times, the latter has been overlooked and in part of the literature independence was assumed without a clear justification. It is therefore important to understand and to account for the bias and dependence when estimating the distribution of lifetimes in the population.
In this talk, we consider inference from a cross-sectional sample of a population that can be joined at a known sequence of discrete times. We calculate the bias function and study conditions under which observed sojourn times are independent and conditions under which treating observations as independent, using the product of marginals in spite of dependence, results in proper inference. We provide conditions for consistency, and further asymptotic properties, including normal and non-normal distributional limits of estimators. The proposed method will be illustrated using data on hospitalization time after bowel and hernia surgeries collected by a cross-sectional design.
Joint with Yosi Rinott, Ronen Fluss, Laurence Freedman and the Department of Health Services Research, The Ministry of Health, Israel.