|
Activity Number:
|
315
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #303228 |
|
Title:
|
Interval Estimation of Risk Difference for Clustered Data
|
|
Author(s):
|
Tasneem Zaihra*+ and Sudhir R. Paul
|
|
Companies:
|
University of Windsor and University of Windsor
|
|
Address:
|
270 Patricia Road, Windsor, ON, N9B 3B9, Canada
|
|
Keywords:
|
Clustered binary data ; Epidemiological Index ; Risk difference ; Intracluster-correlation
|
|
Abstract:
|
Risk difference is one of the most popular epidemiological index used to compare the probability of developing a disease for individuals in an intervention group with that in a control group. There are varying cluster sizes in each group and the binary responses within each clusters are correlated. In our presentation we introduce 2 very simple methods under the cluster sampling scenario. One method is based on variance of a ratio estimator and the other is based on a sandwich estimator of the variance using the generalized estimating equations approach. These two methods are then compared, by simulation, in terms coverage probability, coverage length, symmetry and bias with other 4 methods for confidence interval construction discussed by Lui (Statistical Estimation of Epidemiological Risk, 2004).The method based on an estimate of the variance of a ratio estimator performs best overall.
|
