Combining Information from Multiple Sources in Bayesian Modeling

• Activity Number: 283
• Type: Topic Contributed
• Date/Time: Tuesday, August 5, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Survey Research Methods Section
• Author(s): Tracy Schifeling*+ and Jerome P. Reiter
• Companies: Duke University and Duke University
• Keywords: missing data ; latent class ; Bayesian

Abstract

In this talk, we will discuss combining information from multiple sources. In particular, we will consider combining a categorical survey with marginal information from another source. In practice we may have information on the margins of certain variables, such as gender, educational status, or age group from the Current Population Survey. We will incorporate such marginal information using a Bayesian latent class model that can capture complex interactions. We will present empirical examples and show how our method can apply to multiply imputing missing data caused by nonignorable unit nonresponse.