ON THE DANGERS OF AVERAGING ACROSS SUBJECTS WHEN USING MULTIDIMENSIONAL-SCALING OR THE SIMILARITY-CHOICE MODEL

When ratings of judged similarity or frequencies of stimulus identific ation are averaged across subjects, the psychological structure of the data is fundamentally changed. Regardless of the structure of the ind ividual-subject data, the averaged similarity data will likely be well fit by a standard multidimensional scaling model, and the averaged id entification data will likely be well fit by the similarity-choice mod el. In fact, both models often provide excellent fits to averaged data , even if they fail to fit the data of each in&vidual subject. Thus, a good fit of either model to averaged data cannot be taken as evidence that the model describes the psychological structure that characteriz es in&vidual subjects. We hypothesize that these effects are due to th e increased symmetry that is a mathematical consequence of the averagi ng operation.