Background: Cycling for transportation has the potential to contribute to an increase in people's physical activity levels. A growing body of evidence links the natural and the built environment to cycling. Whereas previous studies were mostly done within one city or one region, the present study covers the whole of the Netherlands, allowing an investigation of whether associations between environmental characteristics and cycling are context-specific. The study examines the extent to which objectively measured natural and built environment characteristics contribute to cycling duration in the Netherlands, as well as the differential effect of environmental characteristics on cycling duration by municipality size.
Methods: Our sample from the Dutch National Travel Survey 2010-2014 comprised 110,027 people aged 20-89 years, residing in 3163 four-digit postal code areas, nested within 387 municipalities across the whole of the Netherlands. Multilevel Tobit regression models were fitted to assess the associations between the natural and the built environment with average daily cycling duration (in minutes), while adjusting for individual and household characteristics. Interaction effects of natural and built environment characteristics and municipality size on cycling duration were also investigated.
Results: Higher address density, more bus stops, and shorter distance from home to the nearest train station were positively related to cycling duration. Respondents were more likely to cycle on days with higher temperatures, less wind, and less precipitation. Interaction tests showed that increased street density and address density were less cycling-promotive in small urban areas compared to medium or large cities. On the other hand, the positive association between number of bus stops and cycling duration was weaker in the largest and medium-sized cities compared to small urban and rural areas.
Conclusions: Interactions suggest that relations between environmental characteristics and cycling duration are context-specific (i.e., dependent on circumstances that differ between highly urbanized and less urbanized areas). Our findings need to be replicated in other countries to gain more insight into the interplay between environmental factors and municipality size.
Keywords: Cycling; Multilevel regression model; Municipality; Natural and built environment; The Netherlands.