One size fits all?: The influence of built environment on active travel across population segments and geographical contexts

The main objective is to identify how determinants of walking/cycling differ by geographical context, population segment (defined by life stage or preference) and type of walking/cycling, thus paving the way for more diverse policies aimed at specific segments of cyclists/pedestrians. Applying different analytical approaches and using different data sets, this thesis provides evidence that determinants such as socio-demographics, built environment, attitudes, and life events have an impact on walking and cycling behavior, but also that these impacts may differ between population segments, geographical settings and types of walking/cycling. Using the Health Economic Assessment Tool (HEAT), Chapter 2 quantified the population health benefits for different demographic and socioeconomic groups in the Netherlands. The results revealed that a large number of prevented deaths in certain population groups do not necessarily reflect high walking and cycling levels but rather can also stem from large population size, a high mortality rate or both. Chapter 3 investigated the extent to which differences in the effects of built environmental characteristics on cycling behavior are related to municipality size (i.e., the spatial context) using nationwide data for the Netherlands. The results showed that the municipality size may moderate the association between environmental characteristics and cycling duration. Chapter 4 explored interactions between travel mode attitudes, urbanization level and sociodemographic characteristics in relation to bicycle commuting and cycling for other purposes (e.g., leisure, or social trips). The findings suggest the existence of synergetic mechanisms, in which a positive cycling attitude reinforces favorable environmental conditions for cycling (such as higher urbanization levels) or groups likely to cycle (elderly adults and women). Compensatory mechanisms were found in that having a positive attitude toward cycling had a weaker positive effect on bicycle commuting for those in favorable cycling conditions (such as not having access to a car). These findings may indicate that both mechanisms exist for some specific combination of cycling attitudes and individual and environmental factors. Existing knowledge about walking behavior is mainly based on cross-sectional studies, failing to address causal relations. Chapter 5 fills this gap by providing longitudinal evidence regarding the effects of life events (i.e., regarding household composition, employment, and residential relocation) on transport-related and recreational walking behavior. Childbirth and residential relocation were central determinants of transport-related walking. Specifically, relocation to less urbanized areas was significantly associated with increases in transport-related walking duration. No significant longitudinal relationship was observed between life events and recreational walking. Chapter 6 further considered three types of walking —transit-related transport walking, non-transit-related transport walking, and recreational walking—and investigated these three outcomes on both weekdays and weekends. Residential areas with a higher address density, a more connected street network, a shorter distance to the nearest train station, and more bus stop encouraged transit-related transport walking. Shorter distances to utilitarian destinations were more important for non-transit-related transport walking during weekdays. A lower address density is positively associated with increased recreational walking on weekdays, while no significant associations with the built environment were found on weekends