The present study focused on the stopping rule for the CAT applications to psychiatric assessing where test length is short and where reduction of testing burden is important. Three rules were compared in terms of the efficiency and precision: (1) The minimum standard error stopping rule, which an adaptive test was terminated when a predetermined SE has been reached, (2)the modified version of minimum information stopping rule, which the CAT is terminated when there is no more items to provide a predetermined minimum level of information, and (3) the newly proposed predicted standard error reduction (PSER) stopping rule, which two additional threshold parameters (hyper and hypo parameters) are in conjunction with the minimum SE stopping rule.
I have two questions:
(1) The threshold values are specified in the current study. What are the reasons to specify the value of hyper and hypo parameters to be 0.03 and 0.01, respectively? How to overcome the limitation by systermatically varying threshold values (as the authors suggested) in the future study?
(2)Why use the modified minimum information stopping rule?