Explanatory Item Response Models

Abstract

A promising approach to cognitive assessment is to make use of explanatory measurement, which utilizes item and person covariates to explain what is being measured and thus adds explanatory value to the measurements. The basic notions of explanatory item response models (EIRM) are explained in this chapter and three kinds of application are presented. All three applications are based on data from a spelling test and the lme4 code for model estimation is included for all three. The first application illustrates how EIRM can be used as a technology for test construction and item generation, starting from the explanation of item difficulties even when that explanation is far from perfect. The s..