Toward an Integration of Item-Response Theory and Cognitive Error Diagnosis.

Abstract : In this study, a new methodology that is capable of diagnosing cognitive errors and analyzing different methods for solving problems will be introduced, illustrated with fraction subtraction problems. The new approach called Rule space integrates Item Response Theory (IRT) and the algebraic theory of databases (Lee, 1983). The rule space model is general enough to apply to any domain of interest, where classifications and selections of students, diagnoses of misconceptions are done by Bayes' decision rules for minimum errors (Fukunagawa, 1972) or any other equivalent decision rules with respect to a set of systematic errors determined prior to analyses. The first section discusses important objectives which the construction of cognitive diagnostic tests must follow. Then the introduction of the rule space model starts with its brief concept, connection to a distribution theory, construction of a bug library and the rule space, and finally introduction of the operational classification scheme.