A Framework to Specify a Cognitive Diagnosis Component in ILEs

Abstract : This article presents a framework for the cognitive diagnosis of learners' errors in an interactive learning activity occurring in an intelligent learning environment. The proposed framework supports the implementation of an authoring tool. This tool helps instructional designers to specify the features of a component for cognitive diagnosis. Two key issues are addressed. First, the cognitive diagnosis process should promote reflection which may in turn enhance the accuracy of the diagnosis hypotheses: (1) reflection leads learners to contemplate their own understanding of the cognitive skills required by a learning situation; (2) when the pedagogical context is devoid of the physical presence of human tutors, reflection also allows for a more reliable representation of the learner's needs. Second, the pedagogical context in which the learner is diagnosed must be taken into consideration: the parameters of this context affect what is observed and diagnosed during the learning experience. First, this view of cognitive diagnosis is formalized into a framework. Second, the article presents the main functions of CD-SPECIES, a framework-based authoring tool which supports the specification of a cognitive diagnosis component in intelligent learning environments.