|
Activity Number:
|
130
- Time Series Data, Trend Analysis, and Repeated Measures
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2018 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Government Statistics Section
|
|
Abstract #327149
|
Presentation
|
|
Title:
|
Pseduolikelihood for Clustered Time-To-Event Outcomes from Complex Surveys
|
|
Author(s):
|
Jing Wang*
|
|
Companies:
|
|
|
Keywords:
|
Frailty model;
sampling weight;
informative sampling;
partial likelihood;
proportional hazards;
Gaussian-Hermit Quadrature
|
|
Abstract:
|
A pseudo-maximum likelihood approach is employed to clustered time-to-event outcomes from complex surveys to account for design effects including sampling weights under informative sampling. The cluster effects are incorporated into the frailty model and assumed to be independent and identically distributed. The likelihood function is based on the partial likelihood for proportional hazards and is approximated by using Gaussian-Hermit Quadrature. Standard errors are estimated using the Sandwich estimator. Sampling weights are rescaled to reduce bias. Asymptotic properties of estimators are studied as the sizes of level-1 and level-2 units increase. The proposed approach is illustrated using a Monte Carlo simulation study and the Early Childhood Longitudinal Study (ECLS-K) on time to obesity for children in kindergartens between the fall of 1998 and spring of 1999.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2018 program
|
