Questions and Comments:
(1) The theta was estimated using the first stage response and then fixed as "true" in the final-stage ability. Will the theta accurate since the fake responses were involved? Why not use the responses which have no erasures in the first stage to estimate the theta?
(2) I don’t believe that ‘no suspicious patterns in the blocks of adjacent items with large positive residuals’ in the plots Figures indicated that no cheating happened. In generally, students got wrong in some items. If the cheating happened, only those wrong items will be erased and replaced by correct response. Hence, cheating is unlikely to happen in block. In other words, the analysis with the form of a plot of the estimated WR residuals for the flagged test takers is realistic.