1. The meaning of composite score is absent. To understand the rationale of composite score in this article, one should refer to Yao (2010a).
2. The proposed method V2 suggested the weights derived from item information can be different among subjects. That is to say, the composite score is not an objective reality.
3. It seems a little tricky for me that a composite score is obtained from the estimated domain scores after taking the test. If the procedure is followed, how can we generate a true composite score before taking the test as well as the result of simulations?
4. If the rationality of composite score is questionable, then the proposed item selection method, minimize the error variance of composite score, is of foolishness.