Bayesian Latent Pattern Mixture Models for Handling Attrition in Panel Studies With Refreshment Samples

TL;DR

This paper introduces a Bayesian latent pattern mixture model that leverages refreshment samples in panel studies to correct for non-ignorable attrition bias, especially with categorical survey variables.

Contribution

The paper develops a flexible Bayesian latent class model that jointly models attrition and survey variables, improving bias correction in panel data with refreshment samples.

Findings

Simulation studies demonstrate the model's flexibility and effectiveness.

Application to election panel data shows bias correction benefits.

Model outperforms standard latent class approaches.

Abstract

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 non-ignorable 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.