Factors Affecting Crashes at Signalised Intersections

Monash University Accident Research Centre - Report #62 - 1994

Authors: K. Ogden, S. Newstead, P. Ryan & S. Gantzer

Abstract:

This report presents the results of a study of signalised intersections in Melbourne, focussing on the differences in site and operational characteristics between sites with a "high", "normal" and "low" accident frequency over the 5 years 1987-91, based upon an analysis of accident data and entering traffic volumes. The study indicated that the majority of the variation in accidents was not explained by traffic volumes, but by other factors. While no single factor was identified which would lead to a dramatic improvement in safety at signalised intersections, a range of measures were identified which would likely contribute to improved safety, if applied at specific sites where relevant.

Executive Summary

In 1992-93, MUARC undertook a comprehensive study of crash patterns at signalised intersections in Victoria (Ogden and Newstead, 1994). One of the outcomes of that study was a recommendation that further work be done, focussing on a detailed examination of sites with a poor crash record. This recommendation was accepted, and this report presents the results of that study.

It is reasonable to suppose that, all else being equal, the more traffic that an intersection carries, the more accidents will occur there. Therefore, any analysis of the factors contributing to crashes at a signalised intersection must take account of the traffic flow. This project takes account of this factor explicitly, by considering those sites which had both higher and lower accident frequencies than might be expected on the basis of variations in accident frequency due to traffic flows alone.

Data were collected for a range of sites with a poor accident record, for which VicRoads had traffic data for the number of entering vehicles. For comparison, a range of nearby sites with a better accident record was also included in the database.

Analysis of the data indicated that about 21 per cent of the variation in accidents was statistically explained by variations in traffic flow. Thus, the majority of the variation in accidents is explained by other site or operational factors.

Details of a range of site characteristics were obtained, and analysis was performed on the basis of three intersection groups: those which, in relation to the volume of entering traffic, had a "high", "normal" or "low" accident frequency over the 5 years 1987-91.

The results of this analysis are that factors other than just traffic volumes are contributing to accidents at signalised intersections. However, the study was not able to identify any single prime factor which would lead in itself to a dramatic improvement in safety at signalised intersections. Nevertheless, based on this and previous studies, a range of strategies or measures was identified which, if applied at specific sites where appropriate, will contribute to safety. These include:

particular attention to sites with an upgrade approach
potential for the use of coloured or textured pavements on approaches
potential for use of skid resistant pavements
continued use of fully controlled right turns where the relevant guidelines are met
potential for active detection of the approach of a heavy vehicle
selective use of active advance warning saying "prepare to stop" or similar
continued use of mast arms where the relevant guidelines are met
use of exclusive right turn lanes where space exists or can be cost-effectively provided
avoidance of use of narrow lanes at intersections where possible
provision of a median where possible
recognition of a reduction in conspicuity created by visual "noise" (e.g. advertising)
recognition of the effect of complex decision environments on driver performance
conversion of signals to roundabouts where relevant guidelines are met

The study also suggests that:

research be conducted on driver perception on approach to a signalised intersection
an educational campaign highlighting signalised intersections as "the most dangerous site on the road" be considered
Traffic flow data be collected for those signalised intersection sites with very high accident frequencies, so that a further study of these sites could be undertaken, at present this information is not available.

The key is that those responsible for the design and operation of signalised intersections carefully appraise each site on its merits, have regard to known accident characteristics of signalised intersections, and apply sound and judicious engineering judgement to determine what particular features will likely contribute to safety at that site. The above recommendations will be useful in this regard.

Sponsoring Organisation: Baseline Research Program - Department of Justice, Transport Accident Commission, Royal Automobile Club of Victoria (RACV) Ltd, VicRoads