Research
My dissertation research, directed by Jerry Reiter, was on the use of
synthetic data for statistical
disclosure limitation. It was funded by NSF Grant
ITR-0427889, Info Tech Challenges for Secure Access to Confidential
Social Science Data, under which my primary activity related to the
generation of partially synthetic public use files for the U.S. Census Longitudinal Business
Database at the Triangle
Census Research Data Center. In addition, I worked on
inferential methods for synthetic data, in particular multicomponent
tests for two-stage multiply imputed data where missingness and
disclosure limitation are handled simultaneously and
model selection for multiply imputed data. My preliminary exam project was on fixed and
random effects selection for logistic mixed effects models with David Dunson. I still am actively working on
several topics from my thesis as well as education statistics at NISS with NCES. In
general, my research interests include statistical disclosure limitation, Bayesian
statistics, model selection, and applications in education, energy and
environment, social science and policy.