Bayesian sensitivity analysis of incomplete data: bridging pattern-mixture and selection models
- 24 September 2014
- journal article
- research article
- Published by Wiley in Statistics in Medicine
- Vol. 33 (27) , 4841-4857
Abstract
Pattern‐mixture models (PMM) and selection models (SM) are alternative approaches for statistical analysis when faced with incomplete data and a nonignorable missing‐data mechanism. Both models make empirically unverifiable assumptions and need additional constraints to identify the parameters. Here, we first introduce intuitive parameterizations to identify PMM for different types of outcome with distribution in the exponential family; then we translate these to their equivalent SM approach. This provides a unified framework for performing sensitivity analysis under either setting. These new parameterizations are transparent, easy‐to‐use, and provide dual interpretation from both the PMM and SM perspectives. A Bayesian approach is used to perform sensitivity analysis, deriving inferences using informative prior distributions on the sensitivity parameters. These models can be fitted using software that implements Gibbs sampling. Copyright © 2014 John Wiley & Sons, Ltd.Keywords
Funding Information
- National Institutes of Health (5P01HD039386-10)
This publication has 24 references indexed in Scilit:
- A Bayesian model for time-to-event data with informative censoringBiostatistics, 2012
- A Semiparametric Estimation of Mean Functionals With Nonignorable Missing DataJournal of the American Statistical Association, 2011
- Last Observation Carried Forward: A Crystal Ball?Journal of Biopharmaceutical Statistics, 2009
- A Bayesian sensitivity model for intention‐to‐treat analysis on binary outcomes with dropoutsStatistics in Medicine, 2008
- A Bayesian Approach for Clustered Longitudinal Ordinal Outcome With Nonignorable Missing DataJournal of the American Statistical Association, 2006
- Adjusting for Nonignorable Drop-Out Using Semiparametric Nonresponse ModelsJournal of the American Statistical Association, 1999
- Semiparametric Regression for Repeated Outcomes with Nonignorable NonresponseJournal of the American Statistical Association, 1998
- Modeling the Drop-Out Mechanism in Repeated-Measures StudiesJournal of the American Statistical Association, 1995
- Double Exponential Families and Their Use in Generalized Linear RegressionJournal of the American Statistical Association, 1986
- Formalizing Subjective Notions about the Effect of Nonrespondents in Sample SurveysJournal of the American Statistical Association, 1977