VALIDATION OF RAIL DEMAND FORECASTS GENERATED BY COMPETING STATED PREFERENCE AND REVEALED PREFERENCE MODELS: OBSERVATIONS BASED ON LOW COST PANEL & CONVENTIONAL SURVEYS

The city of Belfast has recently witnessed the opening of two new rail projects including a new city centre station. The decision to proceed with these projects was informed by two modelling exercises undertaken in the late 1980s involving separate execution of revealed preference and stated preference procedures. The University of Ulster's Transport Research Group are currently evaluating the impacts of these schemes both related to travel behaviour and to other issues. Part of the exercise involves an attempt to validate the modal forecasts with particular influence to the SP procedures. These results will be available in the Spring of 1997. The process of data collection associated with the impact study included developments of panel data sets as well as a variety of other surveys. Low cost data procedures dictated by funding have been employed. The quota based panel surveys of over 1000 households were undertaken both before and after the opening of the rail facilities in a selection of catchment areas deemed affected by the developments. The resulting sample exhibited bias in terms of age and car ownership. Corrective weighting is being applied. On board surveys to identify how the development directly affected rail users given the small subsample of rail passengers the overall panel produced. The paper considers not only travel behaviour change but reviews the degree of accuracy in the original level of service assumptions as well as background conditions. It also reviews issues of sample design and bias. For the covering abstract, see IRRD E101013.