Theoretical and empirical comparison of Luce’s choice model and logistic Thurstone model of categorical judgment

Theoretical predictions of the two models are analyzed and shown to be different if the number of categories is three or more. Then two common methods of testing mathematical models empirically are examined: (1) direct testing of predictions embodied in constraint equations, and (2) minimum chi-square methods. In the experimental investigation, two undergraduate subjects performed an eight-stimulus, 8-response category judgment of loudness in a neutral and a payoff biased condition. There were no significant departures from the Thurstone model, but substantial and significant departure from the choice model, especially in the biased condition. Chi-square tests were more powerful than constraint tests, but less good at identifying the source of conflict between data and theory.