Diagnostic Classification Models

The term diagnostic classification models (DCMs) refers to a family of psychometric models that are used in education to provide statistically driven classification of examinees according to mastery levels of a predefined set of knowledge components, skills, or abilities. The knowledge components, skills, or abilities are typically called attributes. Attributes comprise the construct of interest for a diagnostic assessment—they are the latent variables that the assessment is designed to measure. The distinguishing feature of DCMs from other latent variable models—for example, item response theory or factor analysis—is that the latent attributes are assumed to have a categorical distribution instead of a continuous distribution. As a result, DCMs classify examinees into groups instead of scaling examinees along a continuum. This entry describes the theoretical underpinnings of ...