|
Activity Number:
|
196
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, August 3, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
WNAR
|
| Abstract - #303187 |
|
Title:
|
Outcome-Dependent Sampling Studies for Longitudinal Binary Response Data with Time-Varying Covariates
|
|
Author(s):
|
Patrick Heagerty*+ and Jonathan Schildcrout
|
|
Companies:
|
University of Washington and Vanderbilt University
|
|
Address:
|
Department of Biostatistics, Seattle, WA, 98195,
|
|
Keywords:
|
|
|
Abstract:
|
The wide availability of longitudinal data through ongoing cohort studies and other resources permits examination of any number of novel hypotheses. Often, we have all information available to conduct analyses except for a key exposure or confounding variable. If ascertainment costs are high, we must be judicious about who is sampled, and in such circumstances, outcome dependent sampling designs permit efficient parameter estimation. We introduce a class of designs that sample with probability related to whether or not subject-specific response variability was observed, and we propose maximum conditional likelihood for estimation and inference. We will compare estimation efficiency among designs in this class and will discuss the extent to which the target covariate distribution must be considered during the planning phase.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2009 program |
