INDICATORS OF TRANSPORT SAFETY
A CAUTIONARY NOTE
Targets, Indicators and Road Safety Strategy
Traditionally used statistics are not a good
indicator of whether roads are safe, and may even be perverse
indicators of safety. Of course, even if we get very high quality
indicators, this does not necessarily mean that local or central
government will feel the need to pursue attempts to secure change
leading to such indicators painting a better picture: the strategy
pursued, rather than concern with statistics, is what really counts.
Nevertheless, we will always be hampered by
a culture which says things are better than they actually are
by using traditional methods of analysis. RDRF came into being
largely to oppose this method of evaluating safety on the roads,
and continues to do so. We continually emphasise that retrospective
analyses of aggregated casualty statistics are a poor or even
contradictory performance indicator for assessing safety for road
users. Our other main point of emphasis is that the road safety
strategy should be an integral part of the new sustainable integrated
transport strategy, and that unless this is fully understood,
serious errorshowever unintentionalwill be made.
What's Wrong With the Current Target(s)
1. Current figures are aggregates of different
kinds of phenomenon. There is no distinction between innocent
people killed/hurt by others and people killing/hurting themselves.
Even where benign road users are legally at fault, one
can argue that being killed/hurt is different from hurting/killing
others. (The "who kills whom" question).
2. Figuresparticularly for less
serious injuriesare unreliable, because of unknown
amounts of under-reporting and mis-reporting.
3. In terms of personal experience the
figures tell us little, even if we talk about figures for a particular
group of road users. This is partly because of the morality of
issues involved (see 1 above), but also because there is no indicator
of exposure. Exposure means looking at casualty rates
(casualties of one type or another per distance, time taken, or
4. In fact, even casualty rates as depicted
fall far short of an accurate representation of our individual
problems. Rates have to be split down not just to road
user groups, but by age and more specifically by location. Sometimes
there will be different stories for different kinds of rate:
fatality rates are higher for cyclists in rural as opposed
to urban areas, but less serious casualties have lower
5. In fact, even very accurate casualty rate
measurements are hampered by being retrospective analyses
of what has happened: they do not relate directly to the danger
posed to road users. Hence the interest in indicators of
levels of danger.
A. Casualty rates, where rate is referring to
the journey made
This could be casualties per average trip,
or casualties per kilometre, for each road user group.
B. Insurance claims
There is a set of published figures from the
Association of British Insurers (ABI) in RAGB. This almost doubled
in the 1980's alone, and could be said to be an indication of
danger as it relates to the much larger number of vehicle crashes
rather than the casualties which occur in only 5 per cent-20 per
cent of them.
C. Likelihood of killing others indicator
It is possible to get a picture of the chances
of being involved in a crash where someone else dies, and translating
this into a likelihood of killing others indictor, with
a comparison between different kinds of road user, rather than
the more usual comparison in terms of being killed or hurt.
This would not, however, be a useful target type of statistic,
merely an indication of comparative danger to others.
D. Percentages of children (at different ages)
allowed to walk or cycle to school
E. Modal split or traffic reduction target
There are a number of targets which are indicative
of the aims of a sustainable transport policy, such as
proportions of journeys by more benign modes, reduction in motor
vehicle traffic overall etc.
F. Speed Targets
We should consider proportions of motorists
breaking speed limits if we have a specific speed target,
which is published already.
G. Total or average motorist "blunder numbers"
This was best described in the Autoglass survey,
where self-report (Note: motorists tend to under-report their
mistakes or bad driving) studies by motorists indicated an average
of 50 "blunders" per week. These "blunders"
are types of behaviour designated as "accident-causing"
by TRL criteria. A target would be trying to get 50-100 billion
blunders a year down by a few billion blunders a year down by
a few billion.
H. General attitudes
Questionnaires or quicker repertory grid techniques
can be used to assess residents attitudes to particular schemes,
and amplify issues like allowing children independent mobility.
Note: Severity of casualties. Apart from
the fact that these are affected by levels of medical care and
methods of data collection (i.e., under new STATS 19 rules whiplash
alone counts as "Slight Injury"), Slight Injuries have
increased by about 50 per cent since the early 50's and contradict
the picture painted by DETR figures.
Finally, RDRF has consistently emphasised two
1. The moral issue which highlights the differenceif
not conflictbetween looking after yourself and endangering
other people (community rather than just personal safety), and
2. The importance of the overall strategy,
rather than specific targets.