The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.
Online Program Home
Abstract Details
Activity Number:
|
180
|
Type:
|
Contributed
|
Date/Time:
|
Monday, July 30, 2012 : 10:30 AM to 12:20 PM
|
Sponsor:
|
IMS
|
Abstract - #303992 |
Title:
|
Comparison of Imputation Methods for Missing Values in AR(1) Longitudinal Studies
|
Author(s):
|
Michikazu Nakai*+ and Yoshihiro Miyamoto and Kunihiro Nishimura
|
Companies:
|
National Cerebral and Cardiovascular Center, Japan and National Cerebral and Cardiovascular Center, Japan and National Cerebral and Cardiovascular Center, Japan
|
Address:
|
, , _, ,
|
Keywords:
|
missing data ;
imputation ;
complete case ;
LOCF ;
longitudinal data ;
AR1
|
Abstract:
|
Missing values often occur in longitudinal studies. It is known that Complete Case Method gives some biases and may cause incorrect analysis. In this presentation, we investigate the efficiency of several imputation methods in longitudinal studies with AR(1) covariance structure with four imputation methods including the Complete Case Method, Mean Imputation Method, Last Observation Carried Forward (LOCF) Method, and Multiple Imputation Method. Simulation studies are conducted with repeated measurements Yit (i=1 to 100;t=1 to 5) generated from a multivariate normal distribution with mean response E(Yit)=ß0+ß1t where ß0= intercept, ß1= slope and p|s-t| =correlation for p=0. The variance at each occasion is assumed to be constant over time, while the correlations have a first-order autoregressive (AR(1)) pattern with positive coefficient. Assuming that the first occasion is fully observed, simple random sampling without replacement is used to make MCAR (Missing Completely at Random) datasets and to test three different cases.Normality Shapiro-Wilk test is performed to each imputation method at each time point. At last, two different slope values (0.1 and 2) are tested.
|
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 2012 program
|
2012 JSM Online Program Home
For information, contact jsm@amstat.org or phone (888) 231-3473.
If you have questions about the Continuing Education program, please contact the Education Department.