How Reliable is Stated Preference?

Agencies involved with transport operations have a need for traveller valuations of the various attributes of the transport alternatives. Observation of revealed market place behaviour in order to make Revealed Preference estimates of these valuations have great limitations. As an alternative, analysis of travellers' Stated Preferences has gained in popularity, alheit with worries as to its reliability. This paper seeks to make the case that Stated Preference surveys can be reliable if carried out in a reasonably competent way. Validation against Revealed - Preference methods is discussed, followed by an explanation of why Stated Preference can be better than Revealed Preference. The importance of testing designs is stressed. A method which provides consistency checks and independently derived estimates of attribute valuations is presented. Orthogonality of designs is discussed. The Scale Factor problem and Taste Variation are discussed, with advice as to how to overcome any difficulties.