Downloads & Free Reading Options - Results

A version of sequential mastery testing is studied in which response behavior is modeled by an item response theory (IRT) model. First, a general theoretical framework is sketched that is based on a combination of Bayesian sequential decision theory and item response theory. A discussion follows on how IRT based sequential mastery testing can be generalized to adaptive item and testlet selection rules; i.e., to a situation in which the choice of the next item or testlet to be administered is optimized using the information from previous responses. The performance of IRT based sequential and adaptive sequential mastery testing is studied in a number of simulations using the Rasch model. The possibilities and difficulties of application of the approach in the framework of the two-parameter logistic and three-parameter logistic models is also discussed. (Contains 19 tables and 16 references.) (Author/SLD)