Stated Preference (SP) methods have been used extensively in transport research and elsewhere both for demand forecasting purposes and to value theimportance attached to different product features and travel attributes. Alongside the broader acceptance and wider application of SP methods, somepractitioners have argued for greater openness in discussing what they see as significant concerns surrounding SP. A particular problem is the upward biasing of willingness to pay in situations where respondents have little reason to expect that they would actually have to pay. This study looksat the case of improved rolling stock (railway coaches) in a city in northern England. An SP experiment was used with the two alternatives being two stock types, illustrated and reasonably well known in the area. It is generally accepted that there is a big difference in quality of experience between travel in the two stock types. It is less clear if short distance travellers are willing to pay extra to enjoy superior rolling stock. Importantly, respondents were also asked if they thought it likely that fares would actually rise if the better quality rolling stock was introduced to their route. This application formed a good basis for the exploration of theeffect of such perceptions due to its size and rigorous design. Target sample size was determined after careful simulation work looked into how important differences in valuations could be spotted with a reasonable statistical confidence. Responses were obtained from 1222 inbound morning peak hour travellers into central Manchester, yielding 10998 preference observations. Two possible methods of influencing bias in the SP responses were considered: (a) making the experiment more complex in the hope of masking its purpose, and (b) by adding a written warning, known in the environment and politics literature as 'Cheap Talk'. It is therefore possible to look at the interaction of the responses regarding perceptions with the effectiveness of these two methods. This paper attempts to link individuals' perception and their choice making process to better understand this process. Astandard MNL and Heteroskedastic Multinomial Logit model (HMNL) estimation were used to analyse the data. It was found that adding a Cheap Talk script amends the valuation of improved rolling stock in the case study in the expected direction; however, bias may well remain. Adding more attributes to the SP experiment, i.e. increasing task complexity, did not show a significant impact on the estimation bias, but contributed to higher error variance in responses. Significant effects of respondents' perception on their valuation of new rolling stock have been found.
Abstract