Chapter 10: The Mixture General Diagnostic Model

Rule space methodology (Tatsuoka, 1983) and latent structure models with multiple latent classifications (Goodman, 1974a, 1974b; Haberman, 1979; Haertel, 1989; Maris, 1999) represent the most well-known early attempts at diagnostic modeling. The noisy-input deterministic-and (NIDA; Junker & Sijtsma, 2001; Maris 1999) is an example of a recently discussed diagnostic model. Similarly, the deterministic-input noisy-and (DINA) model, which is a constrained (multiple classification) latent class model has been discussed by several authors (Haertel, 1989; Junker & Sijtsma, 2001; Macready & Dayton, 1977). More recently, the unified model (DiBello, Stout, & Roussos, 1995), which lacks identifiability in its original parameterization, underwent modification and was recast as the reparameterized unified model (RUM; also referred to as the fusion model or the ARPEGGIO system; Hartz, Roussos, & Stout, 2002).