Results of discrete choice models for long-distance travel based on the DATELINE survey: conference paper STRC 2006

Abstract

This working paper describes the enrichment of the Dateline long-distance travel dataset with synthetic mode and destination choice sets and the results of logit modelling. Dateline is a revealed preference survey of travel of 55,000 residents of 16 European countries in 2001. It contains detailed trip stage and sociodemographic information. Synthetic mode choice sets were constructed by merging the attributes of non-chosen road, train, and air alternatives to the geocoded origins and destinations. The synthetic destination choice sets consist of the chosen destination plus 9 randomly chosen destination alternatives at NUTS3 level. The attributes of the destinations are taken from the Eurostat geographic databases GISCO and Regio/New Cronos. Other Pan-European discrete choice models at this resolution are not known. It is the first known application of discrete choice estimation to the Dateline dataset. This paper presents the construction of the dataset and its quality, the model specification, and the model results. The models show a significant interactive influence of gender and trip purpose on mode choice, with trip distance attenuating the travel time. Gross domestic product and purchasing power parity at regional scale, and the number of hotel beds, significantly explain destination choice.