Organizational Research Methods

 

Advanced Search

Journal Navigation

Journal Home

Subscriptions

Archive

Contact Us

Table of Contents

Register here to gain access to SAGE's 500+ Journals Online

SAGETRACK

Sign In to gain access to subscriptions and/or personal tools.
This Article
Right arrow Full Text (OnlineFirst PDF)
Right arrow All Versions of this Article:
1094428106295499v1
10/4/609    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Add to Saved Citations
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Request Reprints
Right arrow Add to My Marked Citations
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Cheung, M. W.-L.
Right arrow Search for Related Content
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?
First published on July 23, 2007, doi:10.1177/1094428106295499

Organizational Research Methods 2007;10:609.

A more recent version of this article appeared on October 1, 2007

Article

Comparison of Methods of Handling Missing Time-invariant Covariates in Latent Growth Models under the Assumption of Missing Completely at Random

Mike W.-L. Cheung*

National University of Singapore

* To whom correspondence should be addressed. E-mail: mikewlcheung{at}nus.edu.sg.


  Abstract
Latent growth models implemented in multilevel models (MLM) or structural equation models (SEM) may be used to analyze longitudinal data with an emphasis on interindividual and intraindividual differences. The main objective of this study is to compare methods of handling missing time-invariant data under the assumption of missing completely at random. Listwise deletion (LD), mean substitution (MS), the expectation-maximization (EM) algorithm, multiple imputation (MI), and full information maximum likelihood (FIML) are compared via a computer simulation study. The findings show that FIML and LD generally perform best, whereas the standard errors of EM and MI are usually underestimated. The results on MS are generally acceptable except when the percentage of missingness is large. Practical implications and directions for future research are discussed.
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati    What's this?