Recent advances in cognitive diagnosis computerized adaptive testing

Abstract

In recent years, cognitive diagnosis models (CDMs) have gained increasing popularity because of their potential to provide finer-grained inferences that can inform learning and teaching. To further capitalize on the advantages of CDMs and make diagnostic testing more efficient, cognitive diagnosis computerized adaptive testing (CD-CAT) has been proposed. This presentation will compare various item selection indices that have been used in CD-CAT with respect to attribute classification accuracy, item usage, and implementation time. To examine the benefits of CD-CAT with real data, its attribute classification accuracy vis-à-vis a proportional reasoning test will be compared with those of the original and optimally designed paper-and-pencil versions of the test. The presentation will also discuss other recent advances in the area, which include item selection indices for different item responses and test formats, nonparametric CD-CAT, and CD-CAT implementation when the number of attributes is large.