Mixture-Distribution and HYBRID Rasch Models

Abstract

This chapter provides an overview of mixture-distribution Rasch models (RMs) and HYBRID RMs and their extensions. Discrete mixture-distribution IRT models assume that the observed data were drawn from an unobservable mixture of populations. Within each of these populations, a different item response model may hold (HYBRID models), or models with different sets of item parameters and different ability distributions may hold (mixture Rasch models, or more generally, mixture IRT models).