Item Exposure Control in Variable-Length Cognitive Diagnostic Computerized Adaptive Testing (CD-CAT)

• In Event: New Developments and Applications in Cognitive Diagnosis Models

Abstract

Cognitive diagnostic computerized adaptive testing (CD-CAT) aims to obtain diagnostic information in an efficient manner by taking advantage of CAT. Most of the current CD-CAT research, however, focuses on fixed-length CD-CAT and a more efficient counterpart is variable-length CD-CAT. This study attempts to address the item exposure control issue in variable-length CD-CAT. The only feasible method, SHTVOR, is difficult to understand and implement. This study proposes several much simpler alternatives: the modified restrictive progressive method (MRP), the modified restrictive threshold method (MRT) and the binary pattern stratification method (BPS) based on the binary searching in computer science. These new methods can make CD-CAT more accessible to education practitioners and students.

Authors

• Chanjin Zheng, Jiangxi Normal University

• Hua-Hua Chang, University of Illinois at Urbana-Champaign