Abstract
Active transport plays an essential role in young people's daily physical activity, the lack of which is likely to lead to health issues such as obesity and heart disease. Active transport to school (ATS) is an easy and sustainable way for adolescents to reach their recommended daily physical activity level and start to live a healthy lifestyle. Understanding adolescents' ATS behaviour could help local authorities, including city councils and schools, with decisions made in encouraging young people to be more active and therefore foster long-term health benefits. A key aspect of that understanding is the many different factors that influence ATS. These include built environment factors (e.g., population density, intersection density and land use), socio-environmental factors (e.g., distance to school), socioeconomic factors (e.g., access to vehicles, home income level), demographic factors (e.g., age, gender and co-ed school attendance), and factors at the individual level (e.g. parental, sibling and peer influence).
Arguably most ATS factors, such as land use, distance to school, income level and age, have a spatial component. Furthermore, the association between ATS and these spatial factors is not stationary over space. However, previous ATS studies exclusively examined the association between ATS and ATS factors using a non-spatial statistical approach. Therefore, the full potential of a spatial approach, underpinned by properties such as distribution, association and autocorrelation, has not yet been realised. This indicates a deficit in understanding a key analytical aspect of ATS and ATS factors and the follow-on lack of spatial evidence to support decision-makers in promoting ATS usage. Geographical Information Science (GIS) has the spatial analytical and visual spatial communication theory and practice content required to fill this knowledge gap. This research thesis applied qualitative and quantitative GIS approaches for exploring, transforming and displaying spatial data from the real world to examine the spatial aspects of ATS and ATS factors in Otago, New Zealand.
To identify any significant spatial patterns of ATS use in association with ATS factors, exploratory spatial analysis methods, including kernel density estimation and local indicators of spatial association, provided qualitative and descriptive evidence that the association between ATS and all ATS factors are significant and spatially varying within and between different settlement types. Significant spatial association difference was found with the home neighbourhood deprivation index (a surrogate for the household income factor), varying within a large urban area and medium urban area pairing. Differences in spatial patterns of gender were reported within the large urban area, half of the medium urban areas, and most of the small urban areas.
Moving from exploration to explanation of the described spatial associations, a spatial modelling approach through geographically weighted logistic regression was applied to quantify the spatial relationships found. A stepwise model calibration approach using goodness-of-fit indicators determined one “winner” model from a total of 127 spatial model candidates. This spatial modelling confirmed and quantified the spatial association between the adolescents' probability of using ATS and two ATS factors: residential density and mixed land use entropy in the home neighbourhood.
The results of the qualitative and quantitative spatial approaches were included as mapped and graphed evidence in a geovisual analytics-based decision-making support (DMS) tool to help decision-makers in promoting ATS usage among adolescents. A series of tests were conducted to evaluate the acceptance and usability of the DMS tool. As a result, the tool was considered useable and acceptable in day-to-day decision-making. The visual exploration of ATS usage using the tool replicated the reported spatial association findings through the qualitative and quantitative spatial approaches.
In conclusion, this research thesis was a pioneering and robust attempt at a spatially-oriented understanding of adolescents' ATS and proved the necessity and applicability of using GIS in understanding adolescents' ATS usage. The quantified spatial associations reported and the geovisual analytics-based decision-making support tool designed in the research thesis could be used by decision-makers to initiate actions to promote ATS usage among adolescents. In addition, this study recommended that a more human-oriented and data-driven place-based approach could be appropriate for future studies and decision-makers to further explore adolescents' ATS and understand the human experiences (ATS usage) in the place-making processes and the effect of places on human behaviours.