The irrelevance of distributional assumptions on reaction times in multidimensional scaling of same/different judgment tasks

摘要: Takane and Sergent developed a model (MAXRT) for scaling same/different judgments response times (RTs) simultaneously. The assumes that RTs are distributed lognormally. Our experiment showed the RT distribution of might be task dependent. It is shown lognormal provide far better fit than exponential, normal, Pareto (with same means variances), but final parameter estimates from data set with hardly differ alternatively RTs. Finally, despite robustness distributional assumption respect to estimates, it have an informational value not contained in alone.