The value of response times in item response modeling

A new and very interesting approach to the analysis of responses and response times is proposed by Goldhammer (this issue). In this approach, differences in the speed-ability compromise within respondents are considered to confound the differences in ability between respondents. These confounding effects of speed on the inferences about ability can be controlled for in experimental settings. As a result, the data for psychological or educational inferences consists of the response vectors only. The response time vectors are redundant as these are equal for all respondents (at least in the response signal paradigm as preferred by Goldhammer, this issue, and Goldhammer & Kroehne, 2014). To assess the merit of this promising approach by Goldhammer, a straightforward question that arises is: Why are we interested in differences in response times? Below, I will argue that the natural variability in response times can give valuable information for psychological and educational inferences about response processes and solution strategies but that the approach by Goldhammer is very valuable if a single process or strategy needs to be measured in isolation.