The goal of this project is to develop improved statistical methods for human fertility studies. Work has progressed in two major areas:(1) developing biologically-motivated models that distinguish between the different factors underlying fertility; and (2) developing methods that account for missing and mismeasured data. In the first area, we developed new Bayesian fertility models for distinguishing between factors associated with sterility, multiple ovulation, reduced cycle viability, and embryo loss. These models can also be used to predict biologically meaningful parameters, such as the frequencies of dizygotic twin implantation and birth. In the second area, we developed two new methods for adjusting for bias in the fertility parameters caused by measurement errors in identifying the day of ovulation. In addition to correcting for bias, these methods can be used to estimate the magnitude of error in commonly used markers of ovulation. Nonreporting of intercourse can also produce bias and inefficiency in estimating fertility parameters, and methods were developed to correct for this problem.
Showing the most recent 10 out of 25 publications
