Accident records for about 9000 streets in local authority residential estates, free from substantial through traffic, in 20 towns, were analysed in relation to population statistics, traffic flow, observations of children and parked vehicles, design parameters etc, using multiple regression and other techniques. The emphasis was on pedestrian accidents, which were predominantly to children. Accident rates (per house, or per person, per year) were found to be strongly correlated with traffic flow and with design factors such as street length, highway width, presence of shops, schools and buses etc which were themselves strongly correlated with traffic flow. Rates in culs-de-sac were very low. The multiple regression analysis suggested that the direct effects of factors such as carriageway width and curvature were probably not large, and might be either positive or negative. Similar conclusions were reached for parking densities and car ownership rates. There appears to be a serious problem of child pedestrian accidents in traditionally designed estates. It is suggested that the low accident rates in culs-de-sac (and also in 'segregated' estates) indicate that better design might achieve substantial improvements. (Author/publisher)
Samenvatting