ABSTRACT
An outlier response in dual scaling is defined as a response that exerts an excessive influence on the outcome of quantification. This excessive influence shows up in the form of spurious results on weights or other statistics. In dual scaling, such outlier responses are observed almost exclusively in multiple-choice data. Therefore, our discussion of outliers is restricted to this data type. In the first section, we see a case in which a fraction of the data set exerts such a strong effect that it determines the characteristics of the first solution. In other words, the first solution is determined by only a few outliers in a large data set.
