A multicriteria analysis of stated preferences among freight transport alternatives

Stated preferences data can be of different types: choice data, rankings or ratings. In all cases, these data can be used in different ways as inputs of econometric discrete choice models. This allows to estimate the weights of the different attributes characterizing an alternative. For freight transport, an alternative's attributes would be, for example, reliability, safety, frequency, etc., besides time and cost. Depending on the data sample, number of alternatives and number of attributes, it is possible to proceed to an analysis of individual data or of aggregated data. In case one is interested to analyze individual behaviors in depth, the option exists to rely on some kind of multicriteria analysis for deriving individual utility functions (actually, decision functions) rather than on a classic discrete choice model. Such a procedure also can be useful for deriving individual utilities as input in a hybrid model combining individual utilities with group data. Such a multicriteria approach is envisaged in the context of a stated preference experiment that is currently applied to freight shippers in Belgium. The data in this case are rankings of alternatives, and there is multicritera method that is particularly well adapted for such data: the UTA models developed by Jacquet-Lagrèze and Siskos. It is based on the specification of an additive utility made of non-linear partial utility functions that are piecewise linear. This allows the convenient set-up of a linear goal programming problem which estimates all the functions and their weights. The paper intends to present the ranking experiment, and to use some of the preliminary interviews to illustrate this UTA methodology. Also, it will be shown how it can be used to derive equivalent money values for each attributes on the basis of the cost attribute, and how to distinguish valuations in terms of willingness to pay and willingness to accept a compensation.