1. Since PR-SE is appropriate to be implement to the variable-length CAT, Different stopping rule should be compared. However, PR-SE should be revised on the different stopping rule. It means PR can be combined with any weight function under the different stopping indices. Therefore, the PR combined with the different stopping rule will lead to different items exposure rate in the end of test.
2. it is odd to me that the BIAS in the low ability persons is negative and in the high ability person is positive. The author said that he estimated ability by EAP approach and the prior is standard normal distribution. I expected it should lead the variance of estimation smaller. That is, the BIAS for negative ability person should be positive and for positive ability person should be negative. However, the results showed in this paper completely was opposite to my expected.