The Development of Computerized Adaptive Testing with Cognitive Diagnosis for an English Achievement Test in China.

Cognitive Diagnostic Computerized Adaptive Testing (CD-CAT) is an innovative development in psychological and educational measurement. The CDCAT is designed to select items sequentially during the process of assessment to provide accurate information about examinees' cognitive strengths and weaknesses. This paper introduces the development and implementation of a web based CD-CAT program for a large-scale English test in China, the Level 2 English Achievement. The paper is organized as follows. First, the main components of the CD-CAT design will be introduced, such as identification of attributes, Q-matrix formation, choice of CD model, item bank and calibration, item selection strategy, the parameter estimation, and design of CAT delivery system. Next, a study investigating an adequacy of model fitting for a large-scale pretest is described. This is followed by a classification accuracy study using the item parameters calibrated from the pretest. An ensuing section presents the results of a field CD-CAT test with 582 students in Beijing. A validity study of the computer-generated diagnostic results was also conducted. The final section discusses directions for future work. [ABSTRACT FROM AUTHOR]