This article proposed an approach to estimate participants’ trait when extreme response style (ERS) is present. The methodology was to simultaneously analyze data from the targeted scale and the other scale, and it stated that the results supported the use. However, some issues are needed to be clarified and described in the followings:
(1) The detection of ERS problems. Are there any criteria for detecting the ERS problems? How could we know that responses to a certain scale were usual and caught researchers' attention? Do we use the information, such as the proportions of participants responding extremely negatively or positively consistently across all items within a scale? If yes, how large is the proportion that should be recognized as an ERS problem?
(2) The selection of the other scale. Are there any standards for selecting a scale to combine with the targeted measure? Do we select scales assessing similar constructs? In addition, because individual study should start with a comprehensive plan for data collection, are there any prerequisites a scale should meet to avoid ERS problems?
(3) Is it possible to analyze more than 2 scales simultaneously that might yield more accurate estimates of participants’ abilities compared with analyzing 2 scales?
(4) The extreme positive and negative responses were rescaled as the same ratings. However, positive response and negative responses might reflect different response patterns and how to deal with this situation?
(5) In table 9, the patterns of responses to scale 1 and scale 2 were different for the 4 participants. Also, the estimates of latent ability 2 were different from the generating traits in joint analysis of scale 1 and scale 2 responses. These findings did not provide strong evidence to support the conclusion.