The role of habit in the choice of commuting to work by bicycle

A vast literature on biking as a means of transport has emphasized the importance of psychosocial factors in shaping individual preferences. However, few studies have explored cycling habits and their impact on the decision to cycle to work. This study investigates the role of habitual cycling in the choice to commute by bicycle in an urban context. To do so, we constructed an Integrated Choice and Latent Variable (ICLV) model, where cycling habit was treated as a latent variable, revealed by the frequency of cycling for leisure, errands, and commuting. Each individual could choose among four different commuting modes: car, public transport, walking, or cycling. The kernel of the model is a heteroskedastic multinomial logit, accommodating variations in unobserved factors for commuters with different number of available travel options. Our dataset comes from a survey conducted in Cagliari and Sassari, Italy, with a sample size of 2128 individuals. Modeling results confirmed the significant influence of habitual cycling on the decision to commute by bicycle. The calculation of average treatment effects revealed that changes in factors influencing cycling habits have a greater impact than changes in Level of Service (LoS) attributes. Finally, we compute LoS elasticity effects and substitution patterns among modes following the implementation of three different project scenarios. These analyses highlighted that a simple multinomial logit model overestimates the impact of LoS attributes compared to the ICLV model including habit effect.