Jeremy Broughton

Getting off your bike: cycling accidents in Great Britain in 1990-1999.

The paper extracts and tabulates selected incidence and fatality rates of cycling accidents recorded by the police in Great Britain during 1990-1999, in a database of over 30,000 standardised reports of fatal or serious injury accidents. Since usable estimates of exposure to risk in cycling are not available, the concept of exposure invariance is developed to estimate the relative risk of different sorts of bicycle/vehicle encounter. Any bias from under-reporting of serious injury accidents is shown to be boundable if the probability of under-reporting is constant.

THE THEORETICAL BASIS FOR COMPARING THE ACCIDENT RECORD OF CAR MODELS

The accident records of different models of car can be compared statistically, provided that accident data which allow the make and model of accident-involved cars to be identified are collected on a national scale: this has been done in Great Britain since 1989. This paper considers the theoretical basis for comparing safety and shows that, because of the lack of detailed exposure data, the most which can currently be achieved is to measure the level of secondary safety (also known as crashworthiness). Based on mathematical considerations, it is shown that the best measure of secondary safety of a particular model is the proportion of drivers who are injured when involved in a two-car accident where one or other driver is injured. In order to minimize bias, this proportion should be adjusted statistically to allow for the influence on the accident data of factors such as type of road and age of driver.

The actual threat posed by unrestrained rear seat car passengers

A statistical analysis of the threat posed by unrestrained rear seat car passengers to restrained drivers and front seat passengers was published in 2002. This was based on Japanese accident data from 1995 to 1999, and the conclusions have been checked using corresponding accident data from Great Britain. The law in respect of the use of seat belts and other restraint systems was less strict in Japan than in Great Britain between 1995 and 1999, so an earlier period had to be studied. The results suggest that the risk of death to a front seat car occupant is increased by about three-quarters by the presence of an unrestrained rear seat passenger behind them, a much smaller increase than was claimed by the Japanese researchers.

The British index for comparing the accident record of car models

This paper describes the second part of a study of the ways in which the accident records of different models of car can be compared on the basis of suitably detailed national accident data. An earlier paper (Broughton 1995) showed, from theoretical considerations, that the most satisfactory safety index is the one currently used by the British Department of Transport (DoT) to measure the level of secondary safety (crashworthiness). This paper presents empirical tests using British accident data from 1989-92 which confirm its value. It also describes a modelling approach which yields the same index and thus provides a theoretical justification for the DoT index. The index declines linearly with mass of model; a second safety index is developed on the basis of this relation which allows models of widely differing masses to be compared directly.

FURTHER ASPECTS OF THE BRITISH INDEX OF SECONDARY CAR SAFETY

This paper develops earlier research into statistical methods for comparing the secondary safety of car models. Two papers (Broughton 1996a,b, Accident Analysis and Prevention, Vol. 28, pp. 89-99, and pp. 101-109, respectively) had concluded that the most satisfactory index of secondary safety is the one first used in publications of the U.K. Department of Transport, referred to as the British or DoT index. This paper shows that the distribution of the risk of injury when two cars collide depends principally on the difference in mass; as this rises, the driver of the lighter car is more likely to be injured and the driver of the heavier car is less likely to be injured, while the likelihood of both being injured reduces slightly. It also shows that the level of protection in fatal and serious accidents varies between models to a significantly greater extent than the level in all injury accidents. Car models of similar mass can provide significantly different levels of protection to their occupants, so there would be fewer casualties if all models were to provide the same level of protection as the most successful current designs. It is estimated that if the safety of all models were improved to the level achieved or exceeded by the safest twentieth of models then the number of drivers injured in two-car accidents would fall by 12% and the number killed or seriously injured by 22%.

Effectiveness of speed indicator devices: An observational study in South London

Speed indicator devices (SIDs) are vehicle activated signs displaying the real-time speeds of passing vehicles. They are part of the array of road safety measures for managing speeds on local authority roads in the United Kingdom. This paper documents an evaluation of the effectiveness of SIDs that was carried out in South London in 2008. SIDs were installed at 10 sites in South London for periods of between one and three weeks. The overall effect of the SIDs being installed was a reduction in vehicle speeds of 1.4mph and a significant reduction in the proportion of vehicles exceeding the speed limit. The research also investigated the distance over which the SIDs are effective and the effect on vehicle speeds once the SID is removed. The results showed that a speed reduction occurred over a distance of up to 400m and that once the SID is removed then, in general, the mean vehicle speeds returned to pre-SID levels by the end of the first week.

A collection of evidence for the impact of the economic recession on road fatalities in Great Britain

There was a considerable reduction in the number of fatalities on British roads between 2007 and 2010. This substantial change led to debate as to the cause of the reduction. Multiple sources of information and evidence have been collated including STATS19 road accident data, population data, socio-demographic data, economic patterns, weather trends and traffic and vehicle data. Summary analyses of these data sources show a reduction in overall traffic, a large reduction in HGV traffic, a reduction in young male drivers, a reduction in speeding, and a reduction in drink driving during the recession period. All of these reductions can be associated with a reduction in fatal accidents and have led to the conclusion that the economic recession changed behaviours in such a way that fewer people were killed on the roads in Britain during this period.

A Comparison of ILP and Propositional Systems on Propositional Traffic Data

This paper presents an experimental comparison of two Inductive Logic Programming algorithms, PROGOL and TILDE, with C4.5, a propositional learning algorithm, on a propositional dataset of road traffic accidents. Rebalancing methods are described for handling the skewed distribution of positive and negative examples in this dataset, and the relative cost of errors of commission and omission in this domain. It is noted that before the use of these methods all algorithms perform worse than majority class. On rebalancing, all did significantly better. The conclusion drawn from the experimental results is that on such a propositional dataset ILP algorithms perform competitively in terms of predictive accuracy with propositional systems, but are significantly outperformed in terms of time taken for learning.

The influence of car registration year on driver casualty rates in Great Britain

A previous paper analysed data from the British national road accident reporting system to investigate the influence upon car driver casualty rates of the general type of car being driven and its year of first registration. A statistical model was fitted to accident data from 2001 to 2005, and this paper updates the principal results using accident data from 2003 to 2007. Attention focuses upon the role of year of first registration since this allows the influence of developments in car design upon occupant casualty numbers to be evaluated. Three additional topics are also examined with these accident data. Changes over time in frontal and side impacts are compared. Changes in the combined risk for the two drivers involved in a car-car collision are investigated, being the net result of changes in secondary safety and aggressivity. Finally, the results of the new model relating to occupant protection are related to an index that had been developed previously to analyse changes over time in the secondary safety of the car fleet.

Estimating the potential impact of vehicle secondary safety regulations and consumer testing programs on road casualties in emerging markets

ABSTRACT Many of the 1.24 million deaths on the roads each year occur in emerging markets such as Malaysia and Brazil. Vehicle safety is not yet fully established in these countries and vehicle use is growing quickly, leading to increasing casualties. If vehicle safety regulations and consumer testing programs, similar to those in Europe, were introduced then casualty savings could be accelerated. Malaysia and Brazil have progressed this with the introduction of Asean and Latin New Car Assessment Programme (NCAP), respectively; this study aims to quantify the benefits of also applying secondary safety regulations, such as those applied in Europe, to emerging markets. Statistical modelling is used to estimate the secondary safety benefits observed in Great Britain due to the introduction of regulations and Euro NCAP. These casualty savings are then applied to the predicted number of casualties in the emerging markets, based on the current vehicle safety situation and casualty rate trends. The results show that, if car secondary safety improvements are transferred to the Malaysian vehicle fleet, between 1,200 and 4,300 car-user fatalities could be prevented between 2015 and 2030. In Brazil it is estimated that between 12,500 and 34,200 fatalities could be saved over the same time period.