Overview
My primary areas of research include methods for preserving data confidentiality, for handling missing values, for integrating information across multiple sources, and for the analysis of surveys and causal studies. I enjoy collaborating on data analyses with researchers who are not statisticians, particularly in the social sciences and public policy.
Current Duke Appointments & Affiliations
Henry W. Newson Distinguished Professor of Statistical Science
·
2026 - Present
Statistical Science,
Trinity College of Arts & Sciences
Professor of Statistical Science
·
2013 - Present
Statistical Science,
Trinity College of Arts & Sciences
Faculty Research Scholar of DuPRI's Population Research Center
·
2010 - Present
Duke Population Research Center,
Duke Population Research Institute
Affiliate Faculty Member, Duke-Margolis Institute for Health Policy
·
2024 - Present
Duke-Margolis Institute for Health Policy,
University Institutes and Centers
Bass Fellow
·
2015 - Present
Statistical Science,
Trinity College of Arts & Sciences
Recent Scholarly Works
Regression analysis after bipartite Bayesian record linkage
Journal article Computational Statistics and Data Analysis · January 1, 2027 In many settings, a data curator or data analyst links records from two files to produce an integrated dataset. These linked data are then used to estimate regression models of interest. This two-stage approach does not necessarily account for the uncertai ... Full text CiteOutcome-Assisted Multiple Imputation of Missing Treatments.
Journal article Observational studies · January 2026 We provide guidance on multiple imputation of missing at random treatments in observational studies. Specifically, analysts should account for both covariates and outcomes, i.e., not just use propensity scores, when imputing the missing treatments. To do s ... Full text CiteMultiple imputation for nonresponse in surveys using design weights and auxiliary margins
Journal article Survey Methodology · January 1, 2026 Survey data typically have missing values due to unit and item nonresponse. Sometimes, survey organizations know the marginal distributions of certain categorical variables in the target population. As shown in previous work, survey organizations can lever ... CiteRecent Grants
Synthetic Data for the National Center for Health Statistics
ResearchPrincipal Investigator · Awarded by Georgia Institute of Technology · 2023 - 2025Enhancing Synthetic Data Techniques for Practical Applications
ResearchPrincipal Investigator · Awarded by National Science Foundation · 2022 - 2025Addressing Bias from Missing Data in EHR Based Studies of CVD
ResearchCollaborator · Awarded by National Institutes of Health · 2018 - 2023View All Grants
Education
Harvard University ·
1999
Ph.D.
Duke University ·
1992
B.S.