One of the key parameters used in transport planning has traditionally been the value of travel time. A number of studies have compared the values of in-vehicle time (IVT) obtained from behavioural choice models calibrated to Revealed Preference (RP) and Stated Preference (SP) data and there has generally been an encouraging degree of similarity (Accent and Hague Consulting Group, 1999; Bradley and Gunn, 1990; Oscar Faber TPA, 1993; TPA, 1992; Wardman, 1986, 1988, 2001b). Given that the RP models were generally estimated to large samples drawn from contexts where real choices existed, we can therefore have some confidence that the SP models provide a reasonable account of travellers' actual preferences amongst IVT and cost. However, demand forecasting and appraisal for public transport modes requires that a broader range of attributes are considered. These include walk time, wait time, headway, interchange and integration, ticketing issues and information provision, on-board and at-terminal facilities, vehicle type and overcrowding. The valuation of these attributes has been much more reliant on SP methods and as a result far less attention has been paid to validating the estimated values against the findings of corresponding RP models. The main aim of this paper is to provide evidence on the extent to which the values of non-IVT attributes obtained from SP models correspond with those obtained from RP models. Implicit in this discussion is the assumption that an RP choice model based on a well defined choice context and producing sufficiently precise estimates provides an accurate account of the relative importance individuals attach to different travel attributes. The evidence the authors cite covers review material and the results of fresh empirical research involving a series of within-mode comparisons of RP and SP models. The emphasis is placed on evidence obtained in the British context, largely relating to the penalty involved in having to interchange and the values of walking and waiting time.
Abstract