Enhancing Global Efficiency in CAT: A Maximin Information Criterion for Item Selection

• In Event: Computer Adaptive Testing Applications

Abstract

To enhance global test efficiency, this study applies maximin information criterion (MMI) for item selection. MMI tailors item selection to each examinee's provisional trait estimate, improving the efficiency of item administration. Simulation results indicate that MMI outperforms other ISRs by offering comparable RMSE and a more balanced item pool usage.

Authors

• Hsiu-Yi Chao, Soochow University

• Jyun-Hong Chen, National Cheng Kung University