An Empirical Study of Applying Person-Fit Indices to Detect Answer Copying on Multiple-Choice Examinations

• In Event: Item Response Theory: Applications and Extensions

Abstract

This study compared the performance of several person-fit indices at detecting copiers under various copying conditions. It was found that the detection rates for all indices were affected by copying conditions, such as copying strategies and copying from sources at various ability levels. The IRT-based fit indices performed better than the CTT-based fit indices when using longer tests. Caution must be excised when applying the CTT-based indices to examinees with extremely high or low scores. To assist researchers with using the CTT based fit indices to identify aberrance, their critical values were proposed by simulation.

Authors

• Aolin Xie, University of Minnesota

• Troy T. Chen, ACT, Inc.

• Chi-Yu Huang, ACT, Inc.

• Deborah J. Harris, ACT, Inc.