Toward Large-Scale Item Response Theory Factor Analysis of Polytomous Data

• In Event: Marriott Roundtable Session 29
  In Roundtable Session: Estimation and Reporting of Multidimensional Item Response Theory Models

Abstract

The estimation of item and person parameters in the framework of item response theory (IRT) is a common practice with regard to educational and psychological assessments. More recently, there has been great interest in estimation procedures for multidimensional IRT (MIRT). The goal of this paper is to compare two different estimation procedure for IRT factor analysis of polytomous data. The first is the estimation procedure is Metropolis-Hastings Robbins-Monro algorithm for factor analysis (Cai, 2010) which is the core function of the MHRM estimation option in the software flexMIRT (Houts & Cai, 2016). The second is the SAEM algorithm adapted for factor analysis (Camilli & Geis, under review; Camilli & Fox, 2015).

Authors

• Eugene J Geis, Deluxe Marketing Solutions

• Gregory Camilli, Rutgers University - New Brunswick/Piscataway