Active transport and physical activity behaviours are recognised as important determinants of a number of health outcomes, including obesity. Over the last decade, there has been a significant amount of research focused on the need to quantify the ‘walkability’ of neighbourhoods or urban environments as a means of predicting physical activity behaviours. The most common methods used to create indices of walkability focus on a combination of land use mix, street connectivity and dwelling density, as developed by Frank et al., (2005). What is largely missing in this research, however, is a focus on other modes of active transport (such as cycling) and a related recognition of how different delineations (Euclidean and network) of neighbourhoods may affect results.

This thesis investigates the influence of the built environment at a number of spatial levels and different neighbourhood delineations, using both standard and novel methods. This research advances and improves our current understandings of the built environment by being the first to use a novel method based on kernel density estimation, to measure associations between the built environment, active transport, physical activity, and health outcomes in a city in New Zealand (Wellington City). This novel method is used to create an Enhanced Walk Index, improving on standard walk indices by including measures of slope, street lights and footpaths and tracks. In addition, this research is the first to test and validate indices of bikeability and neighbourhood destination accessibility (NDAI), based on the novel method.

Results of the study suggest that the novel Basic and Enhanced Walk Indices had strong significant positive associations with active transport and overweight/obesity. In comparison the standard method had weaker significant associations, potentially indicating previous research has underestimated the effect of the built environment on active behaviours and health outcomes. In addition, the novel indices of bikeability and NDAI also showed significant positive associations with active transport and overweight/obesity, however effect sizes were small. Furthermore, results varied depending on the type of neighbourhood delineation and spatial scale used. However, in general, the network buffer showed stronger associations between indices of the built environment and active transport, physical activity and overweight/obesity.

This research thus strengthens current international and national evidence on how the built environment affects active transport, physical activity behaviours and health outcomes. It expands a preoccupation with walkability to encompass other modes of transport, such as bikeability. Furthermore it provides an alternative, and potentially more nuanced novel method to assess the relationships between the built environment, active transport, physical activity and health outcomes.