The cognitive processes are divided into dual-process item response proposed in this article. One is responsible to unconscious, rapid, and effortless process, the other is to conscious, slow, and effortful process. Three CMR models were proposed here. And the analysis of real data example is demonstrated, but no simulation studies involved.
1. The data of response time should be involved and modeled in CMR model, or the estimation error will give rise.
2. Only five items are used and analyzed in CMR model, which is a two-dimensional IRT model. It must suffer from large standard errors in estimates. Unexpectedly, the SEs in table 4 are averagely lower.
3. No time limit of administration involves in the data collection. By naïve intuition, the system 1 may be not very likely to happen because all the examinees have leisure time to finish the works. Except that those examinees who seems to use system 1 may be just reluctant to response, so they answered randomly.
4. The extension model B is looking reasonable. In contrast, I didn’t know is it important of model A?
5. A bit curious how to identify this model?
6. No sense with table3. What can we learn from it?
7. In table 4, it lists Est/SE ratio. What is its meanings?