VALIDATION OF STATED PREFERENCE FORECASTING: A CASE STUDY INVOLVING ANGLO-CONTINENTAL FREIGHT

As part of an ESRC funded study, Before and After surveys were conducted in the UK with manufacturers, freight forwarders and hauliers regarding Anglo-Continental freight. The freight involved was mostly non-bulk exports from the UK crossing the Straits of Dover. As part of the Before Survey, conducted in 1993, a Stated Preference experiment was conducted using the LASP method. This enabled rough estimates of the relative valuations of cost, scheduled journey time, risk of delay, and mode specific constraints for Eurotunnel's Le Shuttle and the through rail services (relative to the existing ferry service) to be derived. These could then be pooled to provide more reliable estimates of these relative valuations for different subsets of the sample. In 1995 the After Survey was conducted with the same firms. In all 32 observed choices of mode (with the tunnel open) were matched with Before interviews, each with its own estimated model coefficients, from which Stated Preference forecasts can be derived. This allows a validation of those forecasts, subject to the usual provisos that all else was not held constant. This paper describes and discusses this validation exercise, presenting results for this case study and drawing wider lessons regarding the validation of Stated Preference forecasts. For the covering abstract, see IRRD E101013.