International Choice Modelling Conference, International Choice Modelling Conference 2015

Font Size: 
Back to the source: modelling information sources acquisition and use for travel choice with a revealed-preference study
Severine Marechal

Last modified: 11 May 2015


A wide array of travel information sources are available, which differ significantly in their format and in content. Any given traveller uses a subset of the available sources, in varying ways and in different contexts. Travel information is extensive, multifaceted and constantly updated but nonetheless potentially erroneous. As technology and societal norms evolve, the usefulness and relevance of different sources evolves accordingly. Some sources become obsolete as they do not respond anymore to the travellers’ expectations. For example, Transport Direct has been as multi-modal journey planner website for the last 10 years but has now closed down its services ( Therefore travel information providers have an interest in understanding attractiveness of information sources and aim to offer useful information. Modelling how the features of these sources have different impacts on individuals’ perception and travellers’ behaviour helps to address this question. For example, the credibility of waiting times given by a phone app or a station board varies across individuals and trip contexts, and its value is useful for operations of information providers.

The use of information in travel choice has long been studied, although the acquisition and use of information has recently been considered as a choice itself. In existing travel information modelling literature, several basic approaches have been developed for acquisition of information: a change from initial to final plan via information (Peirce and Lappin 2003), a choice between adding a piece of information or stop searching and travel (Chorus et al. 2013). However these concepts are usually simplified to be understandable, for the need to fit in specific travel behaviour model or the ease of simulation modelling. More complex models considering detailed features of travel information have been developed using stated preference (SP) data (Wang et al. 2009; Ben-Elia et al. 2013; Chorus et al. 2013) while only a few elaborated an RP-SP survey (Pathan et al. 2011). However realistically simulating the processes of information acquisition, selection and use is difficult in a SP study.

This research contributes to literature by presenting a modelling approach to the choice of information for travel behaviour and a data collection protocol, different to ones found in previous work. The modelling framework is three-fold and represents the joint choice of travel information and its subsequent impact on travel choice. Considering a one-time instance of actively seeking information before travelling, the first choice is an assessment of travellers’ need for travel information depending on their preference and trip context and is modelled using latent classes to capture similarities in information-need behaviour for groups. The second is a combination of any of the travel information sources available to the individual for which multiple discrete choice models or cross-nested models are used. The third is the decision to travel based on the information provided and is modelled using a multinomial choice model, which alternative specific parameters also depend on the information source. The choice in all stages also depends on characteristics of the individual, the context of the trip instance, and satisfaction towards alternatives.

This research aims at collecting quantitative data using a RP survey in order to model both the information acquisition and travel choice behaviours without any experiment-related bias. The data collection protocol is elaborate and involves several steps including focus groups, pilot survey and full survey. The survey sample focuses on adults who commute by using, at least as part of their journey, public transport provided by the Transport for London authority.

Two focus groups assisted in designing the survey on information acquisition and travel choice. The qualitative findings were used to create the hypotheses that will be tested using the quantitative data for the full scale RP data collection.

The pilot survey was developed using a web-based platform and was designed to include questions on demographics, trip preferences, usually used travel information sources and their attributes. In this study, a source of travel information is a combination of its media and provider.

Early findings from preliminary data collection show that the multiplicity of sources of travel information is important, with more than 75% of commuters actively seeking more than one source of information. Amongst the sources listed, smartphone and websites are most used media, while Transport for London, Google maps and phone apps are most popular providers. The characteristics of sources (type of information consulted, frequency of use, perceived difference between information-reported and actual times) are analysed for each source and therefore impact on their satisfaction scores. For example, phone apps, social media, Google Maps and BBC travel news get highest scores regarding their format (ease of access and understanding), while staff at stations, social media and Transport for London get highest scores regarding their content (accuracy, consistency).

The paper will present estimation, results and interpretation of the information need, acquisition and use in travel choice models. It will also present details about the data collection protocol for this revealed preference study.

Ben-Elia, E., Di Pace, R., Bifulco, G.N. and Shiftan, Y. (2013). "The impact of travel information’s accuracy on route-choice." Transportation Research Part C: Emerging Technologies 26(0): 146-159.

Chorus, C.G., Walker, J.L. and Ben-Akiva, M. (2013). "A joint model of travel information acquisition and response to received messages." Transportation Research Part C: Emerging Technologies 26(0): 61-77.

Pathan, A.F.H., Bonsall, P. and de Jong, G. (2011). How the Choice of Mode is Conditioned by Information Source. Choice Modelling Conference 2011, Oulton Hall, Leeds, 4 July 2011 -6 July 2011.

Peirce, S. and Lappin, J. (2003). "Acquisition of traveler information and its effects on travel choices: evidence from a Seattle-area travel diary survey." Volpe National Transportation.

Wang, X., Khattak, A. and Fan, Y. (2009). "Role of Dynamic Information in Supporting Changes in Travel Behavior." Transportation Research Record: Journal of the Transportation Research Board 2138(-1): 85-93.

Conference registration is required in order to view papers.