Challenges and Strategies in Administrative Data Analysis

• Activity Number: 208
• Type: Invited
• Date/Time: Monday, August 4, 2014 : 2:00 PM to 3:50 PM
• Sponsor: ENAR
• Author(s): Joan X. Hu and Rhonda J. Rosychuk*+
• Companies: Simon Fraser University and University of Alberta
• Keywords: Censoring/Truncation ; Incomplete data ; Supplementary Information

Abstract

A typical administrative database collects information over time from a target population starting on a calendar date. Scientifically meaningful analyses, however, often use an individual time of the study subjects, the elapsed time since an individual-specific event such as age. Moreover, many situations demand inference on a population larger than the one from which the data are collected. These, together with other practical constraints, result in various types of incomplete data. This talk begins with a description of three administrative databases to introduce challenges inherent in the data analyses. We then consider a uniform framework to bridge several incomplete data structures, and explore usefulness of supplementary information in attempt to improve inference efficiency and robustness and to reduce computational intensity.