摘要：Analyses of data from random utility models of choice data have typically used fixed parameter representations, with consumer heterogeneity introduced by including factors such as the age, gender etc of the respondent. However, there is a class of models that assume that the underlying parameters of the estimated model (and hence preferences) are different for each individual within the sample, and that choices can be explained by identifying the parameters of the distribution from which they are drawn. Such a random parameter model is applied to stated choice data from the UK, and the results compared with standard fixed parameter models. The results provide new evidence of preferences for various aspects of the UK food system, particularly in relation to GM food but other environmental and technical aspects also. Indications of how random parameter models might be developed further are discussed on the basis of these results.
