A large scale study has been carried out to determine levels of pedestrian activity in representative housing areas and to examine their influence, along with other factors, on annual pedestrian casualty rates. Data were collected on land use and layout, population and socioeconomic characteristics, pedestrian numbers (by age and sex), traffic, and casualties to pedestrians in 474 one kilometre squares distributed in the regions of England and Wales. Analysis of this data resulted in a group of models in which annual casualty rates per sq km of housing area were explained in terms of (a) pedestrian and traffic data, (b) population and census data, and (c) land use/layout data. The best of the models (which were tested against an independent data set) explained up to 77 per cent of the variation about the mean casualty rate (r=0.88). but standard errors of estimate were disappointingly high. Factors influencing the size of these errors are examined, and possible practical applications of the models are discussed. (Author/publisher)
Samenvatting