The role of processing-by-attribute noncompensatory choice models

Research work on modeling consumer multiple choice problems using logit, regression, and probit is gaining more attention. However, in their work, Russ (1971), Tversky (1972), Newell and Simon (1972), Tversky and Sattath (1979), and Gensch and Svestka (1979; 1984) indicate that for many problems, choice behavior appears to be context dependent and hierarchical. With this specific issue in mind, this paper discusses a mathematical model which estimates threshold tolerances, eliminates nonchosen alternatives, provides choice probabilities and finally offers diagnostic information regarding the key attributes that are responsible for making a final decision. The use of other individual specific models such as: lexicographic, conjunctive, etc., have been briefly explicated.