Files
Abstract
Most recent educational assessments were developed under an item response theory (IRT) framework and were designed to scale examinees on an ability continuum. Because there is an increasing desire for more formative feedback about mastery of specific skills, researchers and educators have recently been exploring diagnostic classification models (DCMs), which have potential to improve educational assessment and research due to the multidimensional nature of their feedback. However, because DCMs have only been introduced in recent years, they are not yet widely implemented in analysis. This study seeks to examine the effects of retrofitting, or retroactively fitting a DCM to item responses from an assessment not intended for classification. Using simulations and empirical analysis, I explore item parameter translation and model fit performance within a retrofitting context.