As a recently rising field of psychometrics, cognitive diagnostic model (CDM), has drawn attention of various practitioners in the field, such as classroom teachers, test services providers, curriculum experts and certificate licensing authorizes, etc. Computer adaptive testing (CAT), on the other hand, is relatively a ‘traditional’ field of psychometrics where its theory and practice have been growing at great speed with support of advancement of computing power and complexity which produced numerous discoveries in the last 20 years. As both CDM and CAT are growing it is inevitable but also interesting to witness their relating to each other: CD-CAT.
‘Item selection’ and its strategy is one of the most fundamental components that make CAT work. Researchers have created different item selection rules based on different mechanisms. Basically, there are two popular streams of strategies, one is based on maximizing the Kullback-Leibler (KL) information, and the other is based on minimizing the Shannon Entropy (SHE). The current article focuses on advancing the KL method by introducing two new item selection methods. One problem every selection method has to address is to control item exposure while maintain good accuracy and efficiency. The proposed methods, subject to the authors’ simulations, are created to balance the accuracy-exposure dilemma. As restrictive progressive method, these model allowing flexible size of information component and stochastic component. By considering the relative size of the components during the test, these methods act as an initial step for its applications on real applications. To achieve this, the authors should consider further accessing resources of item banks where item were calibrated based on the CDM since such items enable them to test the methods based on accurate item information. It is worth notice that the authors in the paper have manipulated the exposure so that the maximum exposure rate is always lower than the upper bound; however, the same results may not replicate when some items are being over-exposed in other occasions.