Bayesian Nonparametric Analysis of Multi-Rater Ordinal Data, with Application to Prioritizing Research Goals for Suicide Prevention

• Activity Number: 60
• Type: Topic Contributed
• Date/Time: Sunday, August 3, 2014 : 4:00 PM to 5:50 PM
• Sponsor: Section on Statistical Computing
• Author(s): Terrance Savitsky*+
• Companies: Bureau of Labor Statistics
• Keywords: Bayesian hierarchical models ; Ordinal data ; latent models ; Poisson-Dirichlet process ; Markov Chain Monte Carlo

Abstract

Our application data are produced from a scalable, on-line expert elicitation process that incorporates hundreds of participating raters to score the importance of research goals for the prevention of suicide with the purpose to inform policy-making. We develop a Bayesian formulation for analysis of ordinal multi-rater data motivated by our application. Our model employs a non-parametric mixture distribution over rater-indexed parameters for a latent continuous response under a Poisson-Dirichlet process mixing measure that allows inference about distinct rater behavioral and learning typologies from realized clusters.