International Choice Modelling Conference, International Choice Modelling Conference 2017

Font Size: 
How different user segments perceive multimodal alternatives: A survey of 3 EU cities
Amalia Polydoropoulou, Ioannis Tsouros

Last modified: 28 March 2017

Abstract


Transportation, as a sector, accounts for 23% of greenhouse gas emissions in the EU as of 2014 as opposed to 15% in 1990. The use of multimodal alternatives, where a part of the trip is conducted via a sustainable mode of transport (public transport, active transport), is widely regarded as a sustainable measure towards the goal of emission reduction and a more sustainable future.

This paper presents an investigation of the multimodal preferences of various user groups, classified by their statements towards the environment, the use of private motorized transportation and their level of tech-savviness/usage of social media. The simultaneous user classification and choice modeling is done via a hybrid choice model.  The study aims to investigate further the relationship between the various user types constructed by attitudinal data and the choice of using a multimodal option for everyday transportation needs.

The survey is designed to run on the web, for a more flexible and faster distribution and data collection. The initial data was collected through the promotion of the web survey from the research project’s (OPTIMUM – H2020 funded) partners in the three pilot cities: Birmingham, Vienna, and Ljubljana.

The first section of the questionnaire presents questions relevant to the use of multimodal travel information services. More specifically, the respondents are asked a series of questions to measure the effect that travel information has on their choices, or to measure how useful is travel information for them. The second section collects the socio-demographic data needed for the model. The third part contains the psychometric questions that harvest the user’s attitudes and perceptions to use for the measurement of the latent variables. The users are asked a series of questions (the total of which is attached in the appendix) which range from the perception of the environment to their attitudes towards car usage, their opinion on transport policy, other values and opinions and also the perception of one’s self. The fourth part of the questionnaire contains the persuasive techniques experiment, where respondents are presented with various persuasive techniques in the form of short storyboards and classify and evaluate each presented technique. The final, fifth part of the web survey is the Stated Preference (SP) experiment. In this section, users are presented with various scenarios of mode combinations, along with travel times, travel costs, different weather situations and a list of incentives to choose from.

 

The behavioral model is built on data collected through the EU funded project OPTIMUM. The survey which is web-based and is distributed in two languages (English and Slovenian) is the model input. The model, through an integrated choice and latent variable framework, simultaneously classifies the users based on latent traits and models their choices concerning probabilities for choosing a certain multimodal alternative. The model results are presented and the effect of different latent variables on the final choice is revealed. Also, the probabilities of the user’s and the user classes to choose an alternative are presented. Finally, an application of the model is described, acting as a sensitivity analysis for travel time and travel cost.

To our knowledge, this is the first study that focuses on passenger multimodal choices and spans across three different EU countries. The results should enable policy makers to focus on certain user types to promote multimodal options and to push forward certain types of incentives.

Initial user segmentation results reveal four distinct user groups. User group 1includes the Aspiring Environmentalists, people that have pro-environment attitudes and perceptions and are aware of the adverse effects of motorized transport has on the environment. We measure this factor by using the Env latent variable. Note that this methodology does not cluster the sample into one group of “Aspiring Environmentalists” but rather uses the pro-environmental latent variable as a function that actively affects the final mode choice. So the latent variable is used to classify users depending on their environmental stances but not in just one cluster of pro-environmentalists, but on a continuous scale depending on their answers. User group 2 includes the Car Addicts, people who are used to and enjoy driving their private vehicle. For this group natural environment should be exploited for human advantage. They also regard themselves as outgoing and not reserved. This user group will be measured in the final version of the model using a latent variable CarAddiction. User group 3  contains people that have strong pro-social network perceptions and also are not willing to make sacrifices for the environment or have any concerns about the negative effects of car use. We call them tech-friendly/eco-indifferent. User group 4 contains people that are accustomed to using their car but also have some environmental worries about the negative effects of motorized transport and have positive views about cycling. We call this group “Disappointed Motorists.”

Initial model results include the positive correlation between the “Environmental Friendliness” latent variable and the choice of a public transport alternative. Also, the model results reveal the distributed values of time for the sample and for each alternative based on the calculation of travel time, travel cost and latent variable distribution.  The effect of the other latent variables will be presented in the final version of the paper, as well as, the effect that the incentives have on the final choice.

 


Conference registration is required in order to view papers.