Abstract
A framework for the concepts “model”, “theory” and “data” within the context of probabilistic choice models is presented. What is commonly called a test of a model comes down to an assessment of the goodness-of-fit relation between a model of a theory and a model of data. A distinction is made between theories that lead to ordinal restrictions on choice probabilities (Type I), and theories that specify a functional relation between utilities and choice probabilities (Type II). The study of probabilistic choice behavior could benefit from an assessment of the goodness-of-fit of Type I theories with statistical methods usually reserved for Type II theories.
Original language | English |
---|---|
Pages (from-to) | 207-219 |
Number of pages | 13 |
Journal | Advances in Psychology |
Volume | 60 |
Issue number | C |
DOIs | |
Publication status | Published - 1 Jan 1989 |