International Choice Modelling Conference, International Choice Modelling Conference 2017

Font Size: 
Oscar Arbeláez, Jorge Córdoba, Iván Sarmiento

Last modified: 28 March 2017


At the stage of modal partition in the classic transportation planning model, has been common to use discrete choice models, this research proposes a choice model between motorized modes and bicycle under three possible infrastructure scenarios: a) bike paths, b) shared lanes with signaling and c) none infrastructure. Data correspond to a declared preference survey with 1231 individuals, conducted in the city of Medellín-Colombia, with which hybrid discrete choice models were estimated with infrastructure as block variable and with the inclusion of latent variables. The article presents the results of applying the model to estimate potential users for bicycle in an urban context for different infrastructure scenarios and compares the results to estimate the effects of providing infrastructure for bike users.

In the study of travel demand, explanatory variables like travel time and trip cost have often been used. In the case of the bicycle, these factors do not explain adequately the its demand associated, leading to attitudes and perceptions as additional explanatory factors to the variables mentioned above. With the inclusion of latent variables that take into account the effects of attitudes and perceptions on modal choice, has been possible to include these effects in discrete choice models (Ben-Akiva et al., 2002). On the inclusion of latent variables in the case of a non-motorized mode, it is clear in the literature available that bike users have attitudes and perceptions towards subjective elements of travel experience that most non users have (Fernandez & Monzón, 2010; Heinen, Maat, & Wee, 2011; Rietveld & Daniel, 2004)

Within the framework of random utility theory, individuals seek to maximize a utility function when faced with a choice set of discrete alternatives, where in the case of transport planning, travel time and cost have been used as explanatory variables (Domencich & McFadden, 1975). The perception of safety and the risk of having an accident is associated directly with infrastructure by bike users, where having a trip sharing lanes with vehicles is perceived riskier than doing it on bike paths separated from traffic flow (Akar & Clifton, 2009; Jacobsen & Rutter, 2012). Based on the foregoing, in the experimental design, infrastructure was included as a block variable in order to test, whether the choice of an alternative can be maximized by providing the necessary infrastructure with consequent modification of risk perception on using bicycle.

Auto, bus and metro users, were surveyed using data collection with the technique of declared preferences, where the choice set included its current mode and the alternatives of bikeshare, bicycle (own), and the combined use of bikeshare and bus, or bikeshare and metro. A question to measure the desire to use bike was included which gave a percentage of 39.9% of respondents agreeing that contrast with the fact that current modal split of bike users rounds 0.5% of daily trips. Figure 1 presents the causal relationship between the explanatory variables of the model.

Figure 1. Infrastructure provision for bike users-hybrid discrete choice model.

Hybrid discrete choice models were estimated in a simultaneous approach, so that the latent variable, which measures the perception of a safe and pleasant bike journey, relates to perceptions of the possibility of an accident by interaction with other vehicles and the consequences of using a bike that affects the user like the possibility of a theft, exhaustion due to road slope, weather and manifestations of increased physical activity which in the end lead to a less pleasant experience. The indicators use for the variable "safe and pleasant journey" were measured on a Likert scale and are listed below:


y1 likelihood of a traffic accident during the trip

y2 likelihood of a theft or robbery during the trip

y3 likelihood of physical or verbal aggression during the trip

y4 physical exhaustion due to the difficulty to the slope

y5 perception of contamination during the journey

y6 perception of rain during the bicycle trip

y7 discomfort by heat and transpiring due to physical effort

After modeling bicycle choice versus auto, motorcycle, bus and metro, in each infrastructure scenario the effect on bike probability of choice was compared leading to an increase between the case of non-infrastructure to bike lanes, and an analisys of the realtive weight of the latent variable was perform in the utility funtion leading to the conclusion that in a scenario with infraesctructure bike choice would increase and the overall perception of risk would be reduced.

Keywords: bicycle, latent variables, discrete choice models, hybrid models, transport demand.


Akar, G., & Clifton, K. J. (2009). Influence of individual perceptions and bicycle infrastructure on decision to bike. Transportation Research Record: Journal of the Transportation Research Board, 2140(1), 165–172. JOUR.

Ben-Akiva, M., Walker, J., Bernardino, A. T., Gopinath, D. A., Morikawa, T., & Polydoropoulou, A. (2002). Integration of choice and latent variable models. Perpetual Motion: Travel Behaviour Research Opportunities and Application Challenges, 431–470. JOUR.

Domencich, T., & McFadden, D. (1975). Urban Travel Demand-A Behavioral Analysis.

Fernandez, A., & Monzón, A. (2010). Cyclists? Travel behaviour, from theory to reality, 17. article. Retrieved from

Heinen, E., Maat, K., & Wee, B. (2011). The role of attitudes toward characteristics of bicycle commuting on the choice to cycle to work over various distances. Transportation Research Part D: Transport and …, 16(2), 102–109.

Jacobsen, P., & Rutter, H. (2012). Cycling safety. Pucher, J, Buehler, R. Eds, 393. Retrieved from

Rietveld, P., & Daniel, V. (2004). Determinants of bicycle use: do municipal policies matter? Transportation Research Part A: Policy and Practice, 38(7), 531–550.

Conference registration is required in order to view papers.