This study developed a hybrid of item response theory (IRT) models and latent class models, which combined the strengths of each type of model. The primary motivation for developing the new model is to describe characteristics of examinees' knowledge at the time of the examination. Hence, the application of the model lies mainly in so-called achievement, diagnostic, and mastery testing environments. The treatment of multidimensionality is central to the hybrid model. The new model's essential characteristics involve its abilities to represent a certain cognitive structure, to incorporate test users' informal task and error analysis as well as formal exhaustive analyses, and to relate to IRT in order to facilitate assessment of inter-item equivalency and performance of differential item analysis. Parameter estimation is central to the attempt to develop the hybrid model. An EM algorithm is used to compute maximum-likelihood estimates from incomplete data. Two experiments were undertaken to assess the model. Overall, results indicate that the hybrid model provides unusual freedom based on its unification of continuous and discrete item response models. By varying group and state probabilities, the model can range from one extreme of IRT-only to the other extreme of latent-classes-only models; any combination of these two models can also be arranged. These latent classes do not have to be ordered in any manner; hence, they may be able to accommodate individual differences of the order of learning materials. A 56-item list of references and 12 data tables are included. (TJH)