In 1998/1999, measurements of a national sample of more than 500 intersections of urban through-roads with mopeds, bicycles, and pedestrians were carried out (See C 14622 (ITRD E201705)). Motorized vehicles, mopeds, and bicycles were counted. Features of the intersections and the branches were linked to the 1994-1998 crash database. For that period, this report presents a quantitative judgement of the safety of a number of intersection types that are distinguished by (traffic) functional features. Here, the intersection safety is expressed in the crash rates, i.e. the number of registered injury crashes per million motor vehicles counted. Intersections with a 50 km/h speed limit were distinguished by the following features: a) crossroads or roundabout, b) the number of branches (3 or 4), c) the type of bicycle facilities, and d) the types of priority regulation: traffic lights, a special priority regulation (on priority roads or a priority intersection), and priority for traffic from the right. The features of the intersection types distinguished appear to correlate with the motor vehicle traffic volume on the intersection. On both the 3- and 4-branch intersections, the average 24-hour volume increases in the following sequence: priority from the right, special priority regulation, and traffic lights. The 4-branch intersections always have a greater volume than the 3-branch intersections. The 4-branch roundabouts have a volume in between that of crossroads with a special priority regulation and those with traffic lights. Besides the features of the intersection types, their crash rates also correlate with the motor vehicle volume. It is, therefore, only possible to compare intersection types within the same volume class. Of all the intersection types, the 3-branch crossroads with priority from the right and an average volume of less than 5,000 motor vehicles a day have, on average, the lowest rate (0.03 injury crashes per million motor vehicles). 3-branch crossroads with traffic lights and an average volume of more than 20,000 motor vehicles a day have the second lowest rate (0.06). 3-branch crossroads with other priority regulations or traffic lights and an average volume of 5,000 - 20,000 motor vehicle a day have higher rates (0.08 - 0.13). Of all 4-branch intersections with an average volume of 5,000 – 20,000 motor vehicles a day, roundabouts have the lowest rate (0.07) and 4-branch crossroads with traffic lights the highest rate (0.18). 4-branch crossroads with an average volume of more than 20,000 motor vehicles a day nearly always have traffic lights; their rate is 0,13. The safety differences between intersection types were examined by means of (discrepancies with) the design principles of Sustainable Safety. Possible explanations can be found especially in the functionality of the intersections and the homogeneity of the traffic flows. As far as the predictability is concerned, a lack of behaviour observations makes it impossible to make any statements. It can, however, be assumed that relative high crash rates are partly due to a not always correct interpretation by the road user of the layout and regulations. A definite judgement of a particular intersection type can only be made after specific behaviour studies have been carried out. The found differences in rates may certainly not be used to change unsafe (scoring) intersection types into 'safe' types by changing their distinguishing features. To do this, it is always necessary that the function of the network in which the intersections are situated, plays its part in the judging. After all, this is a comparative study and not one in which the effects of measures are quantified using before-and-after measurements.
Abstract