|
Activity Number:
|
312
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #304699 |
|
Title:
|
Flexible Bayesian Human Fecundability Models with Generalized t-Link
|
|
Author(s):
|
Sungduk Kim*+ and Rajeshwari Sundaram and Germaine M. Louis
|
|
Companies:
|
National Institute of Child Health and Human Development and Eunice Kennedy Shriver National Institute of Child Health and Human Development, NIH and Eunice Kennedy Shriver National Institute of Child Health and Human Development, NIH
|
|
Address:
|
Division of Epidemiology, Rockville, MD, 20852,
|
|
Keywords:
|
Conception ; Fecundity ; Generalized t-Link ; Intercourse ; Menstrual Cycle ; Posterior distribution
|
|
Abstract:
|
There has been considerable statistical literature developing methods to better model probabilities of conception by incorporating various biological and behavioral aspects starting with the work of Barrett and Marshall (1969) and, recently, extensively by Dunson and others (ex. Dunson and Stanford, 2005). The logit link and complementary log-log link models (Dunson and Stanford, 2005) for the day-specific conception probabilities have been the popular choices in models for the human fecundity data. However, these popular links may not always provide the best fit for a given data set. We propose a new Bayesian conception model based on the generalized t-link. Various theoretical properties and attractive features of the proposed links are examined and explored in detail. A detailed analysis of the Oxford Pregnancy Study is presented to illustrate the proposed methodology.
|
