Stated Preference Survey Design in Air Travel Choice Behaviour Modelling
A Collins, J Rose, University of Sydney, AU; S Hess, Imperial College London, UK
This paper discussed the development of an air travel behaviour SP survey that presents respondents with real-world look alike booking process.
Like in most other areas of choice analysis, the vast majority of studies of air travel choice behaviour now make use of Stated Preference (SP) rather than Revealed Preference (RP) data. However, experience as to the optimal design of such surveys in the context of aviation is still very limited, especially when compared to topics such as mode choice or departure time choice analysis. Additionally, there are significant differences between the choice situations presented to respondents in air travel surveys and those actually faced in real life.
Traditionally, SP tasks involving airport or airline choice present respondents with the choice between two or more alternatives described along multiple attribute dimensions. Commonly used attributes in such tasks include ticket cost, travel time to the airport, flight time, departure and or arrival time, number of connections, and on-time performance of the carrier. Often, one of the alternatives presented corresponds to a historic real-life trip made by the respondent.
Typically, these SP tasks resemble grids with one attribute per row and one alternative per column. The realism of these tasks may be considered somewhat dubious, particularly as the tasks present respondents with full information which they may not possess in real life. For example, many travellers may seek to purchase tickets from a travel agent or on-line website in which case they will have only have access to information that the agent or website provides. Further, such sources are unlikely to provided detail with regards to on-time performance or other attributes that are commonly included in SP surveys. Finally, while the ordering of alternatives in the SP surveys is generally random (albeit with any RP alternative presented first), the ordering in a real world booking process can follow certain criteria, such as lowest air fare first, or preferred departure time first. These issues with the realism of the design have a potentially significant impact on model results and implied choice behaviour.
The purpose of this paper is to explore more realistic methods for capturing SP data using airline choice as the context of interest. In this paper, we compare and contrast results from SP data collected using more traditional SP survey methods with data collected from a survey where respondents use websites that are made to look and react exactly like real world on-line booking sites. In this study, respondents complete both the traditional and more realistic tasks, allowing for a within and between subject analysis of the two methods. The results from this analysis should not only provide further insights into what drives air travel choice behaviour, but will also give important guidance for further SP studies in this area.
Association for European Transport