A SINGLE MULTIDIMENSIONAL MODEL FOR DISCRIMINATION, IDENTIFICATION AND PREFERENTIAL CHOICE

Two forms of a mathematical model for tasks involving three alternatives are given. The model can be applied to triad discrimination, preferential choice and two-stimulus identification. In all cases multivariate normal probabilistic assumptions are made with a decision rule based on the smallest Euclidean distance between one of the alternatives and the other two. The model is presented as an indefinite quadratic form and as a weighted sum of central F distribution functions. The model should have numerous applications in marketing, sensory evaluation, experimental psychology and economics.