Bayesian Competing Factor Models for Bi-directional Latent Predictors

• Activity Number: 102
• Type: Topic Contributed
• Date/Time: Monday, August 9, 2004 : 10:30 AM to 12:20 PM
• Sponsor: ENAR
• Author(s): Amy H. Herring*+ and David B. Dunson and Nancy Dole
• Companies: University of North Carolina, Chapel Hill and National Institute of Environmental Health Sciences and Carolina Population Center
• Keywords: Bayes ; categorical data ; discrete choice model ; joint modeling ; latent variables ; stress

Abstract

Researchers often rely on questionnaire data to study the health effects of stress. A subject indicates the occurrence of potentially stressful events and quantifies the strength of reaction to these events, ranging from strongly negative to strongly positive. These data are used to obtain measures of levels of negatively and positively perceived stress. Motivated by such data, we propose a latent variable model characterized by event-specific negative and positive reaction scores. If the positive reaction score dominates the negative reaction score for an event, the individual's reported response to that event will be positive, with an ordinal ranking determined by the value of the score. Measures of overall positive and negative stress are obtained by summing the reaction scores across the events reported by an individual. By incorporating these measures as predictors in a regression model and fitting the stress and outcome models jointly using Bayesian methods, inferences are conducted without assuming known weights for the different events. We propose an MCMC algorithm for posterior computation and apply the approach to study the effects of stress on pre-term delivery.