Reducing the Impact of Inappropriate Items on Reviewable Computerized Adaptive Testing

In a test, the testing score would be closer to examinee’s actual ability when careless mistakes were corrected. In CAT, however, changing the answer of one item in CAT might cause the following items no longer appropriate for estimating the examinee’s ability. These inappropriate items in a reviewable CAT might in turn introduce bias in ability estimation and decrease precision. An early proposed solution to this problem was rearrangement procedure. The purpose of this study was to implement the 4PL IRT model to reduce the estimation bias introduced by inappropriate items in reviewable CAT. The results of this study indicated that the 4PL IRT model could significantly lower the estimation bias for reviewable CAT, and, while it incorporates with rearrangement procedure, provide more accurate ability estimation. Also, the efficiency of reviewable CAT was promoted by introducing both 4PL IRT model and rearrangement procedure.