Do psycho-attitudinal factors vary with individuals’ cycling frequency? A hybrid ordered modeling approach

The purpose of the present study was to investigate specifically whether psycho-attitudinal factors could differ for people with different cycling frequency levels and to quantify the determinants influencing the propensity to cycle. To perform our analysis, we developed a hybrid choice modeling approach with a generalized ordered probit choice kernel, using the information collected in 2016 for 2128 individuals in two mid-size urban areas in Sardinia (Italy). Our results indicate that the latent variables Perception of cycling benefits, Perception of cycling comfort and Perceived importance of bike infrastructure positively influence the propensity to cycle, supporting the idea of a relationship between attitudes and cycling frequency. In addition, the model shows a link between different socio-demographic variables (gender, age, Body Mass Index, education level, number of cars per household, number of household members), built environment characteristics and bike usage. Computation of the pseudo-elasticity effects indicates that strategies focusing only on the physical part of the problem, such as the expansion and improvement of proper infrastructure, might not be sufficient to encourage bike use. At the same time our findings stress the importance of considering people's psychological characteristics when implementing policies aimed at promoting cycling. This can be helpful for identifying, depending on the population segment that is targeted, the most appropriate advertising/information strategy for convincing people to cycle, as well as the most effective marketing messages.