The proportion of independent to dependent trips: A new measure to understand the effects of the built environment on children independent mobility

Introduction: The research on the link between the Built Environment (BE) and Children Independent Mobility (CIM) is on the rise due to CIM's multiple health and other policy benefits. These studies have conceptualised and measured CIM using various indicators, such as mobility licence, independent time spent outdoor, and territorial range. A major problem with these indicators is that they do not indicate the extent of CIM behaviour. For example, a child is labelled as independent irrespective of whether they are independent for all trips or for a single trip or the amount of time spent independently compared to time spent with someone. This weakness of previous studies suggests applying proportionate measures to understand the relationship between the BE and CIM behaviour. Methods: This research uses the proportion of trips made independently as an indicator of CIM and investigates the BE-CIM link stratified by discretionary and non-discretionary CIM. A two-day activity-travel diary data were collected from 151 children aged between 10 and 14 years from Dhaka, Bangladesh. BE data were generated through a virtual BE audit and spatial analysis using ArcGIS and Space Syntax. Three fractional regression models were estimated, one for each of the overall, discretionary, and non-discretionary trips. One Poisson regression model was also estimated for the count of overall CIM to identify if the proportionate measure unlocked new findings. Results: Results show that: a) the proportion of CIM is higher for discretionary trips; b) the frequently visited locations do not necessarily attract more independent trips; c) proportionate measure unlocks new insights for CIM, and d) BE affects discretionary and non-discretionary CIM differently, e.g. the presence of recreational land use is associated with discretionary CIM but not with non-discretionary CIM. Conclusions: The findings suggest that the proportionate measure can effectively be applied to measure the extent of CIM, and BE needs to be designed in tandem with the type of CIM to be promoted.