Context-Induced Parameter Instability in a Disaggregate-Stochastic Model of Store Choice

A model of individual consumer choice is described which posits that the relative weights consumers associate with product attributes in forming preferences covary with the amount of variability on the attributes among choice alternatives in particular contexts. The model is derived in a logit-like form, and has advantages over previously used nested models in terms of not requiring an analyst to prespecify a decision hierarchy and being relatively simple computationally. A binomial form of the model is illustrated by an analysis of a set of experimental store choice data.