Open Access
March 2016 Bayesian latent pattern mixture models for handling attrition in panel studies with refreshment samples
Yajuan Si, Jerome P. Reiter, D. Sunshine Hillygus
Ann. Appl. Stat. 10(1): 118-143 (March 2016). DOI: 10.1214/15-AOAS876


Many panel studies collect refreshment samples—new, randomly sampled respondents who complete the questionnaire at the same time as a subsequent wave of the panel. With appropriate modeling, these samples can be leveraged to correct inferences for biases caused by nonignorable attrition. We present such a model when the panel includes many categorical survey variables. The model relies on a Bayesian latent pattern mixture model, in which an indicator for attrition and the survey variables are modeled jointly via a latent class model. We allow the multinomial probabilities within classes to depend on the attrition indicator, which offers additional flexibility over standard applications of latent class models. We present results of simulation studies that illustrate the benefits of this flexibility. We apply the model to correct attrition bias in an analysis of data from the 2007–2008 Associated Press/Yahoo News election panel study.


Download Citation

Yajuan Si. Jerome P. Reiter. D. Sunshine Hillygus. "Bayesian latent pattern mixture models for handling attrition in panel studies with refreshment samples." Ann. Appl. Stat. 10 (1) 118 - 143, March 2016.


Received: 1 December 2014; Revised: 1 August 2015; Published: March 2016
First available in Project Euclid: 25 March 2016

zbMATH: 06586139
MathSciNet: MR3480490
Digital Object Identifier: 10.1214/15-AOAS876

Keywords: categorical , Dirichlet process , multiple imputation , nonignorable , Panel attrition , refreshment sample

Rights: Copyright © 2016 Institute of Mathematical Statistics

Vol.10 • No. 1 • March 2016
Back to Top