Since the mid-1960's, and the work of Quarmby (1967), Lee and Dalvi (1969) and others, research into the value of travel time savings (VTTS) has explored a variety of types of data set, using a corresponding variety of analytic techniques. The data sets are almost all cross-sectional, and can almost without exception be classified as either Revealed Preference, or Stated Preference, or Transfer Price. The analysis of the RP and SP data sets has been of discrete choice form (usually logit) while the analysis of the TP data has been by regression (usually simple linear). These methods have been applied to travellers on own-business by various modes, and to freight shippers and hauliers (de Jong et al. 1992). Something of an 'orthodoxy' has now built up around the use of SP data sets, with an attempt to maximise the 'realism' of the hypothetical context of the trading by linking the context of the questions to current behaviour. The purpose of this paper is to look a little closer at this orthodox approach, prompted by two findings from recent research. The first of these findings was reported in the appendices to a report from Accent and Hague Consulting (AHCG) Group, "The value of travel time on UK roads", published in 2000, and describes evidence that there are different 'latent classes' in the population, having characteristically different time and cost sensitivities (Gopinath, 1995). The second finding is presented in recent papers from ITS Leeds which suggest that in trading situations presented in SP VTTS studies like the AHCG one, the `status quo' has itself a much increased intrinsic attractiveness to the respondents in some circumstances. These are when not only the context of the experiment is in terms of a past trip, but the time and cost of that trip are actually as experienced. The work reported here, which is ongoing, explores the use of responses to a Transfer Price question included in the 1994 British study as a means to isolate and model separately sub-groups of respondents, on the basis of willingness to pay for time savings on their existing journey. In addition, the ITS analysis of `inertia' is extended using the full models fitted in the AHCG study to examine the ITS conclusions in more detail.
Abstract