Stated preference data is used frequently in transport planning to aid in service or policy design, or to forecast travel demand. In the latter case, the validity of such data with respect to actual travel behaviour has often been questioned. Recently, techniques have been developed to use stated preference information together with actual choice (revealed preference) information. Estimation methods have been developed which use various types of SP and RP data simultaneously, taking advantage of the strengths of each type. It is possible to perform such estimation with existing statistical packages. In this paper the strengths and weaknesses of different types of data-in terms of possible errors and biases - is discussed. Then different ways of combining different data types in analysis, including a simultaneous "treelogit" estimation technique are introduced. Two recent stated preference / revealed preference studies from the Netherlands are used as examples.
Abstract