Understanding route preferences is essential to designing accessible and attractive cycling networks. Energy expenditure is recognized as a critical aspect of cycling but its influence on utilitarian cycling behaviour is still poorly understood. We address that gap in knowledge by exploring how individual cyclist preferences regarding energy expenditure are reflected in their route decisions. Using choice modeling, we seek to determine whether a measure of an individual’s disutility of energy expenditure relative to travel time \(\left(MR{S}_{et}\right)\) moderates their preferences for route attributes. Mixed path-size logit models with latent segmentation are estimated with and without \(MR{S}_{et}\) using a set of 1590 observed utilitarian cycling trips in Vancouver, Canada. Results show that incorporating normalized \(MR{S}_{et}\) improves model sensitivity and reveals that cyclists with more energy-conservative preferences are more sensitive to route distance and relatively less sensitive to most other route attributes (including hills). Speed choices likely mediate these preferences, as energy-conservative cyclists ride slower, disproportionately increasing the time cost of a given detour distance while allowing them to mitigate the energy cost of ascending hills. This study contributes to more nuanced understanding of heterogeneity in cycling behaviour, which is critical for designing equitable cycling networks and policy. For example, energy-related interventions (such as e-bike provision) could lead to safer route selection for energy-conservative cyclists who tolerate lower safety for shorter distance routes. Future work on cycling route choice models should aim to incorporate speed choice explicitly, to distinguish the effects of time versus energy costs.
