|
American Statistical Association
|
The stratification score, which is the probability of being a case as a function of confounding variables, plays the same role in a case-control study as the propensity score does in a prospective study. Here we establish several novel features of the stratification score. We first show that the stratification score is a retrospective balancing score and use this property to show how to estimate the distribution of exposures among case and control participants that would have been observed if case and control participants had been sampled in such a way that they had the same distribution of confounding covariates. We additionally show that using the stratification score for matching is highly advantageous. We further show how the stratification score can be used to estimate the parameters in a (marginal) model for disease/exposure association that would have been obtained if case and control participants had the same distribution of confounding covariates. Finally, we consider additional features of the stratification score model that hold when the mean exposure is a linear function of potential confounders. We illustrate these features using a genome-wide association study of schizophrenia among African-Americans.
I attended Oberlin College, where I started as music composition major and ended up with a degree in Geology. I earned an M.A. in Physics and a Ph.D. in Chemical Physics at Harvard. Except for 2 years in the Peace Corps in Kenya and 2 years post-doctoral work in Biostatistics at UNC Chapel Hill with Larry Kupper, I have been at the CDC ever since. I worked for 10 years in HIV/AIDS and developed the method by which CDC monitors HIV incidence. For the past 10 years, I have worked in genetic epidemiology. I have coauthored over 100 papers, am a fellow of the ASA, and am a past chair of the ASA's Statistics in Epidemiology section and a past president of the Georgia chapter of the ASA. I am the father of three beautiful children ages 13, 14 and 15 and am an avid mountain biker.
| Date: | Thursday, December 9, 2010 |
|---|---|
| Time: | 4:00 - 5:00 P.M. |
| Location: |
Mailman School of Public Health
Department of Biostatistics 722 West 168th Street Biostatistics Computer Lab 6th Floor - Room 656 New York, New York |