Informative IG Variance Priors for Bayesian Multilevel Item Response Model Precision

• In Event: Electronic Board Session I

Abstract

This study demonstrates how inverse gamma can be leveraged for informative variance priors in Bayesian multilevel item response models. Using a 20% subsample of 71 countries with PISA 2018 math data, credible intervals for posterior intercept estimates based on informative variance priors were more precise, while capturing full sample estimates.

Authors

• Liu Liu, University of Washington

• Elizabeth A. Sanders, University of Washington, Seattle