The Role of Accessibility Measures in Travel Behaviour: A Focus on Conventional and Emerging Mobility Options

Abstract

Promoting active travel is essential for improving public health, reducing environmental impacts, and addressing transport inequities. Walking behaviour is associated with a wide range of personal and environmental factors. Out of these factors, the correlation between urban form and travel behaviour is well established in the literature. The extent to which built environment features are associated with active travel is often expressed through measures of accessibility, reflecting the ease of reaching destinations and activities. Given that public transport and micromobility (e.g., shared e-scooter) services may improve access to destinations and may support walking, many scholars have endeavoured to uncover how accessibility is correlated with these mobility options. Despite this growing body of research, important gaps remain in understanding how accessibility can best inform active and sustainable transport planning. This thesis therefore investigates how different accessibility measures correlate with travel behaviour across conventional and emerging transport modes in a highly car-dependent city, Auckland, New Zealand, drawing on objective mobility data and advanced modelling techniques.

The first study (Chapter 3) examined the relationship between neighbourhood accessibility and objectively measured walking behaviour using GPS and accelerometer data from 2,503 trips. An enhanced two-step floating catchment area (E2SFCA)-based accessibility index was used to quantify access from participants’ homes to destinations across eight key domains, such as education, recreation, and public transport. Findings showed positive associations between neighbourhood accessibility and walking behaviour both within and beyond home-neighbourhood boundaries, although these estimates were accompanied by substantial uncertainty. The association for walking beyond the home neighbourhood was estimated more precisely, whereas the estimate for within-neighbourhood walking was less precise. Socio-demographic factors, including age, gender, and work status, were also associated with walking patterns, with older adults and unemployed individuals recording longer walking durations. These findings suggest that neighbourhood accessibility may be more strongly related to walking beyond the immediate neighbourhood than to walking within local areas, although evidence regarding spatial differences in accessibility–walking associations remains uncertain.

The second study (Chapter 4) examined the relationship between accessibility to bus stops, measured from three spatial anchors (home, trip origin, and destination), and actual bus ridership using GPS and accelerometer data from 108 participants (835 trips). Mixed-effects logistic regression models were applied, with work status and neighbourhood residency duration tested as moderators. Results indicated that home-based accessibility was negatively associated with bus use, while origin-based accessibility showed no significant relationship. Destination-based accessibility interacted significantly with neighbourhood residency, suggesting that the association between destination accessibility and bus use differed according to residency duration. Work status marginally moderated the home-based association, suggesting that the association between home-based accessibility and bus use differed between employed and unemployed participants. These findings underscore the complexity of accessibility–ridership relationships and the importance of individual circumstances, particularly work status and neighbourhood residency duration, in understanding how associations between accessibility and bus use vary across population groups.

The final two studies (Chapters 5 and 6) focused on shared e-scooter use patterns, leveraging more than 470,000 trip records collected between April 2023 and May 2024. The third study applied a gravity-based negative binomial model to examine how domain-specific accessibility measures correlate with trip volumes. Results revealed pronounced distance-decay effects consistent with the short-range nature of e-scooter travel. Higher trip volumes were observed in areas with greater accessibility to public transport and recreational amenities, indicating a spatial alignment between micromobility use and leisure or multimodal transport nodes. Conversely, lower usage was associated with greater accessibility to shopping and education facilities, higher exposure to environmental “bads”, elevated car ownership, and increased deprivation. The fourth study employed elastic net regression to investigate how domain-specific accessibility measures were associated with e-scooter trip length. Findings indicated that higher public transport accessibility was associated with shorter trips, consistent with a possible role in first- and last-mile connectivity, whereas greater accessibility to recreation, education, and environmental “bads” corresponded with longer trips. Shopping accessibility showed directional asymmetry, with longer trips originating in retail-rich areas and shorter trips terminating in such destinations. Longer trips were also more common in areas with greater deprivation or higher car-ownership levels, raising equity concerns under distance-based pricing models.

Collectively, these studies indicated that accessibility is an important correlate of active and sustainable travel behaviour. Across walking, public transport, and shared micromobility, the associations between accessibility and travel behaviour varied across population groups according to factors such as age, employment status, neighbourhood residency duration, and local socio-economic conditions. By integrating objective mobility data with multidimensional accessibility metrics, the thesis advances understanding of how accessibility relates to travel behaviour in a highly car‑dependent setting. The findings underscore the importance of combining spatial accessibility improvements with policies that acknowledge demographic variation and address socio‑economic constraints. Altogether, this thesis provides an evidence base to support more equitable, user‑centred transport planning, illustrating that accessibility-focused strategies should be tailored to the circumstances and mobility needs of different groups.