GIS-based spatial analysis of child pedestrian accidents near primary schools in Montréal, Canada

In Montréal, Canada, accidents affecting child pedestrians (5 to 14 years old) remained almost constant from 1994 to 1999 despite the great amount of prevention measures. Moreover, the elementary public school environment has been barely taken into account by past and present research on factors affecting the risk of accident even though children attend school most weekdays. We argue here, therefore, that the integration of the local environment into the spatial analysis of child pedestrian accidents could help to reduce them. Accordingly, we have integrated socio-economic and environmental data into a geographic information system in order to perform a geographically weighted regression and results demonstrate that the average network distance separating accident and closest school is less than 500 meters, thereby confirming a relationship of proximity between these two locations. Results also demonstrate the relevance of adding a spatial dimension to the regression model by suggesting that prevention initiatives should take into account the particular nature of each neighbourhood so that more relevant risk factors can be targeted.