In 1988 a national shippers survey was carried out in France. This database has been used to estimate disaggregate mode choice models and to calculate values of time in freight transport. As a forerunner of a possible new national shippers survey, a revealed preference (RP) shippers survey was conducted in 1999 in the region Nord-Pas de Calais. This database contains about 650 shipments (of 200 shippers). Unlike the 1988 project, a number of the shippers have been re-contacted in 2000 for a stated preference (SP) survey. This resulted in an SP database with 150 shipments (of almost 100 shippers), with road (own account and hire and reward), train, combined and maritime transports. The SP survey contained both within-mode experiments (two alternatives are presented both referring to the same mode) and between-mode experiments (choice between two different modes). The attributes considered in these experiments include travel time, travel cost, reliability, probability of delay, availability of adapted logistic services, flexibility and frequency. Disaggregate logit models have been estimated on the SP data, both for the within mode data and the between mode experiments. An attempt has also been done on the RP data separately. All these data sources have finally been combined. A joint nested logit model has been estimated on the mixed revealed preference and stated preference information. This paper contains a description of the design and the main outcomes of the SP survey and the estimation results of models on the SP and RP data separately and on the joint SP/RP data.
Abstract