Pedal cycle accidents at T-junctions

Analysis of the relationships between accident factors using coded STATS 19 recorded injury accident data can frequently reveal patterns and correlations which aid the understanding of accident causation. However, there are usually very many different types of possible conflict which may be grouped under the broader heading of accidents of a particular kind. Unless these different conflict types are recognised and taken into account, the significance of underlying correlations between factors may not become evident. Cycle accidents at priority controlled T-junctions represent about 40% of the total reported number of cycle accidents in the U.K. Furthermore, such accidents generally fall into one of a small number of types. This paper describes a preliminary analysis which makes use of the injury accident data for the City of Manchester from 1986/87 for those accidents involving cyclists at T-junctions. The accidents are classified as conforming to one of a small number of types and the predominant types and the implications of these classifications are described. Examples of the use of contingency tables and log-linear models are given but it appears that the latter technique requires a larger database to allow conclusive results to be obtained.