FORECASTING ISSUES IN STATED PREFERENCE SURVEY RESEARCH

摘要: The family of market research survey techniques termed stated preference methods (e.g conjoint analysis) has been used quite often in transportation planning over the past decade. While these have primarily to gather marketing information, such as willingness pay for service improvements, an important recent trend is use preferences forecasting. Such are now becoming seen a complement more traditional revealed cases where latter cannot provide full information necessary. Examples introduction qualitatively new transport mode or route (e.g. Channel Tunnel between France and England), impact policies electronic road pricing), reactions large changes travel conditions which be observed at present trip departure times due increasing peak hour traffic congestion). need accurately predict behaviour across entire population places special requirements on design, analysis application experiments. Many analogous those faced studies. There remain issues specific preference, however, because forecasting requires that built upon hypothetical choice context will yield realistic unbiased statements behaviour. As addressed, appear evolving into integrated part methodology. purpose this paper give overview evolution, terms study model estimation involved. discussion includes treatment various errors biases may arise opposed surveys, along with reference practical examples based projects Netherlands, Australia, UK.(A) For covering entry conference, see IRRD abstract no. 843297.