Marginal models for the analysis of longitudinal measurements with nonignorable non-monotone missing data

We propose methods for the analysis of continuous responses subject to nonignorable non-monotone missing data. We form a pseudolikelihood by naively assuming independence over time and using a product of marginal likelihoods at each time point, and we obtain consistent and asymptotically normal esti...

Full description

Saved in:
Bibliographic Details
Published in:Biometrika 1998-09, Vol.85 (3), p.661-672
Main Authors: TROXEL, ANDREA B., LIPSITZ, STUART R., HARRINGTON, DAVID P.
Format: Article
Language:English
Subjects:
Consistent estimators
Estimating equation
Estimation methods
Estimators
Exact sciences and technology
Independence
Logistics
Marginal model
Mathematics
Missing data
Modeling
Multivariate analysis
Non-monotone pattern
Nonignorable missing data
Parametric models
Probability and statistics
Pseudolikelihood
Sciences and techniques of general use
Simulations
Statistical variance
Statistics
Unbiased estimators
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We propose methods for the analysis of continuous responses subject to nonignorable non-monotone missing data. We form a pseudolikelihood by naively assuming independence over time and using a product of marginal likelihoods at each time point, and we obtain consistent and asymptotically normal estimators of the mean and missingness parameters. Our primary interest is in estimating the parameters of the marginal model at each time point, and we make no assumption about the correlation structure. An approximate variance is found for the parameter estimates using a sandwich estimate. We provide an example, and present a simulation study that assesses the performance of the model.
ISSN:0006-3444
1464-3510
DOI:10.1093/biomet/85.3.661