This project involves the development and application of new methodologies to epidemiologic research. (1) The use of randomized recruitment in case-control studies was further developed; (2) collaborative work continued toward developing methods for handling missing data in case-control studies; (3) a generalization of the Mantel-Haenszel estimator was developed that is applicable to data from 2xJ tables; (4) a method was developed to take geographic variability into account when computing standardized mortality ratios; (5) the effects of exposure measurement errors and population mobility were assessed in relation to designing studies of residential radon exposure and risk of lung cancer; (6) the effects of outcome misclassification (false positives and false negatives) in studies of early pregnancy loss were evaluated; (7) the effect of misclassification of time to pregnancy data on study power and bias was assessed by simulation; and (8) work continued in development of an algorithm for identifying the day of ovulation based on urinary levels of ovarian steroid hormones.
