Allowing for Heterogeneous Decision Rules in Discrete Choice Models: an Approach and Four Case Studies
S Hess, ITS, University of Leeds, UK; A Stathopoulos, University of Trieste, IT; A Daly, ITS, University of Leeds/RAND Europe, UK
This paper discusses a modelling concept which allows groups of respondents in the sample population to make use of different decision rules
The study of respondent heterogeneity has been one of the main areas of research in the field of choice modelling in recent years. The emphasis has been on variations across respondents in the parameters used in the utility function while maintaining the assumption that the actual utility specification is generic across respondents. Recent work has moved on from this by allowing for differences in the utility specification across respondents in terms of inclusion or otherwise of specific attributes, in the context of work looking at heterogeneous information processing strategies. The underlying assumption that all respondents employ the same choice paradigm remains, despite evidence in the literature that different paradigms work differently well on given datasets. In this paper, we go one step further by presenting a latent class framework in which the model at the sample population level is a mixture of different individual models. We present applications on three different datasets, showing mixtures between "standard" random utility maximisation models and lexicography based models, models with multiple reference points, elimination by aspects models and random regret minimisation models. In each of the case studies, the behavioural mixing model obtains significant gains in fit over the base structure where all respondents are hypothesised to use the same rule. The findings offer important further insights into the behavioural patterns of respondents. There is also evidence that what is retrieved as taste heterogeneity in standard models may in fact be heterogeneity in decision rules. The resulting patterns of heterogeneity are in many cases more behaviourally plausible than those obtained when imposing a single Mixed Logit class on the entire sample.
Association for European Transport