Bureau of Labor Statistics Conference Center, Rooms 1-3
Postal Square Building
2 Massachusetts Avenue, NE
Washington, DC 20212
This two-and-a-half day workshop, held at the Bureau of Labor Statistics in Washington, DC, involved roughly 12-15 invited talks along with formal panel discussions, a poster session, and time available for informal collaboration-building discussions.
The workshop was part of SAMSI’s 2013-14 program on Computational Methods in Social Sciences (CMSS). For many years, practical work with censuses and surveys has involved complex methodological issues. Standard approaches have addressed some of these issues in a reasonably comprehensive form, while leaving other issues with unsatisfactory or incomplete solutions. In addition, large-scale statistical organizations are now encountering important new methodological opportunities and challenges arising from prospective new data sources; and from changes in salient features of the data–collection environment, resource constraints and cost structures. Addressing these issues requires modern statistical methodology and novel computational approaches.
This workshop, in conjunction with other CMSS activities, brought together researchers and practitioners from academia, statistical agencies, and survey organizations to discuss recent research advances and needs related to the above-mentioned challenges and opportunities.
On January 7, 2014, Stephen Fienberg of Carnegie Mellon University was the featured speaker at the 23rd Morris Hansen Lecture that took place at the US Department of Agriculture. Further details are here.
Schedule and Supporting Media
Wednesday, January 8, 2014
|Session 1: Models for Longitudinal Surveys||Chair: Steve Fienberg|
|9:15-9:45||A Flexible Bayesian Approach to Longitudinal Studies with (Monotone) Nonignorable Missing Data with Extensions to Surveys||Mike Daniels, University of Texas|
|9:45-10:15||Mixed Membership Trajectory Models for Longitudinal Survey Data on Disability||Daniel Manrique-Vallier, Indiana University|
|10:45-11:15||On the Shop Floor: Issues and Questions for Computational Methodologists from the SIPP Program. – Perspectives from a Large-Scale Longitudinal Survey||Jason Fields, U.S. Census Bureau|
|11:45-1:30||Lunch on your own|
|Session 2: Imputation in Complex Data||Chair: Richard Smith|
|1:30-2:00||Bayesian Nonparametric Models for Heterogenous Data||Jared Murray, Duke University|
|2:00-2:30||Bayesian Automatic Editing||Hang Joon Kim, NISS and Duke University|
|2:30-3:00||Application of Administrative Records Usage for the Nonresponse Followup Operation in the Decennial Census||Thomas Mule, U.S. Census Bureau|
Thursday, January 9, 2014
|Session 3: Integrated Data from Multiple Sources||Chair: Jerry Reiter|
|9:15-9:45||Spatial Fay-Herriot Models for Small Area Estimation with Functional Covariates||Scott Holan, University of Missouri|
|9:45-10:15||Combining Information from Multiple Sources in Bayesian Modeling||Tracy Schifeling, Duke University|
|10:45-11:15||Combining Information from Multiple Data Systems to Enhance Analyses Related to Health: Examples and Lessons Learned||Nat Schenker, National Center for Health Statistics|
|11:45-1:30||Lunch on your own|
|Session 4: Record Linkage||Chair: Connie Citro|
|1:30-2:00||Full Bayesian Procedure for File Linking to Analyze End-of-Life Medical Costs||Roee Gutman, Brown University|
|2:00-2:30||Clustering Approaches to Human Rights Violations in Syria||Rebecca C. Steorts, Carnegie Mellon University|
|3:00-3:30||A Bayesian Framework for Duplicate Detection, Record Linkage, and Subsequent Inference with Linked Files||Mauricio Sadinle, Carnegie Mellon University|
|3:30-4:00||Quality and Analysis of Sets of National Files||William Winkler, U.S. Census Bureau|
Friday, January 10, 2014
|9:00-9:30||Working Groups Organize|