TRL fudge
"Speed and accidents - let's put the record straight"

... or maybe not


In the TRL newsletter for September 2002 an article appeared entitled "Speed and accidents - let's put the record straight". The article then goes on to deliberately attempt to distort and misrepresent the clear and accurate record made by TRL323. We find it quite sickening.

The article (which is freely available from the TRL by downloading their newsletter) appears in full below with our analysis and comments below that.

SafeSpeed's analysis:
There is a vast amount of evidence demonstrating the strong link between vehicle speed and road accidents.
There is a vast amount of evidence that driving at inappropriate or excessive speeds for the conditions is highly dangerous. As every responsible driver knows. There's no evidence that exceeding a speed limit in itself is dangerous, and in many cases of modern speed limits the safest group of drivers at 85th percentile speeds will be exceeding the limit. There is plenty of scientific research which verifies that the lowest accident risks belong to drivers at 85th to 90th percentile speeds. If the speed limit is set below this threshold, then the safest drivers are criminalised. 
So why does material keep appearing in the media suggesting the effect is small?
Because every single study of real world accident causation verifies that the effect is small.
The issue is so important we feel it is time to reiterate the true position. In the 1990s a number of police forces conducted a limited trial of an experimental accident reporting system. The results were reported clearly in TRL Report 323 but they have frequently been misquoted.
Misquoted by the Government, the DfT and now, even the TRL itself. Yes. 
Speed increases the impact of many of the factors which contribute to accidents. For example, “aggressive driving” or “driving too closely” are both much worse at speed. Such factors were recorded in the system separately from speed; but speed plays a big part in their effect on accidents. 
Aggressive driving is aggressive driving. Driving too close is driving too close. Driving too fast is driving too fast. These serious errors may be combined or may be separate. That's why TRL323 properly codes them separately. To suggest that certain behaviours are "worse at speed" is to suggest that motorways are more dangerous than roads in built up areas. It's simple not true and the intention is to mislead.
The system also allowed speed to be recorded in its own right. The total effect of speed on accidents is obviously the sum of both types of factor.
Where excessive speed was recorded as an accident causation factor we should believe that it was an accident causation factor. Where it wasn't recorded at all we should be very confident that it wasn't an accident causation factor. 
Misunderstandings in the press appear to have resulted in two ways. First, speed identified as a separate factor in its own right was present in 15% of accidents, not the 7.3%, or lower figures, that are often wrongly
TRL323 (quite brilliantly) allowed for contributory factors to be coded as "definite", "probable" and "possible". "Excessive Speed" was coded as a definite contributory factor 126 times across 2,795 completed report forms. That proportion is just 4.5%. The 15% figure (which does not appear in TRL323) even includes 68 cases where excessive speed was coded as a "possible" factor! It follows that 15% over represents the proportion of accidents where speed was a cause or a contributory factor.
Secondly, the 15% is only one part of the total effect of speed on accidents. When allowance is made for all of the other speed-dependent factors, the contribution is, we believe, much greater.
This is pure obfuscation. Are we talking about speed as a contributory factor? What sort of "speed" is this? What other "speed dependent factors"? And where they say "we believe" they are asking us to share in their blind faith. i.e. "despite the fact that we have no evidence (or we would have quoted it) we believe that the contribution is much greater.
This means that speed is far more important in causing accidents and increasing their severity than the misquoted figures suggest. Importantly, other TRL studies have directly examined the relationship between speed and accidents. These are summarised in TRL reports 421 and 511. They avoid the inevitably subjective judgements associated with studies of contributory factors which, for example, involve estimating what an appropriate speed is in each situation.
TRL 421 and TRL 511 are works of fiction as evidenced by built in logical fallacies. We wrote to the TRL about this problem and have had totally inadequate replies. (click here)
• Studies of individual drivers have examined how drivers’ speed choice affects their likelihood of accident involvement. Accident records of more than 10,000 drivers were related statistically to their observed speeding behaviour. These showed clearly that accident risk rises the faster a driver travels: at 25% above the average speed, a driver is about 6 times as likely to have an accident than a driver travelling at the average speed. (see graph)
We broadly agree. But it is necessary to look at the behaviours in greater detail. It is entirely true that some incautious and reckless drivers exceed safe speeds and have additional accident risks. It's also true that some under confident drivers fail to keep up with the traffic and have increased accident risks. These behaviours affect both speed and accident risk. It simply isn't true that asking a reckless driver to travel a few mph slower will reduce his accident risk. His accident risk and his high speed both arise because he is reckless and a different solution is required.
• Road-based studies looked at how speeds on a given road affect accidents occurring there. Several hundred thousand observations of vehicle speed on almost 300 roads of different types were related statistically to the numbers of accidents on those roads. These showed clearly that the faster the average speed of traffic on a given type of road, the more accidents there are. Injury accidents rise rapidly as average speed increases, if all else remains constant.
We have not commented on this claim before, but we have long known how absurd it really is. We can completely destroy it with simple logic. Firstly, it is well known that faster roads are safer (TRL511 confirms this). Secondly a road has no speed at all until drivers choose to use a certain speed on that road. It follows that for a road to be driven faster, drivers must perceive the road to be suitable for a higher speed. If drivers perceive a road as suitable for a higher speed, then it follows that the road should not be classified in the same group as another road where drivers perceive that a lower speed is required. Therefore it is impossible to claim that similar roads are driven at different speeds - if they are driven at different speeds, then those roads are not similar to the drivers. The TRL claims absolutely depend on identifying "similar roads" which are driven at "different speeds". But truly similar roads are not driven at different speeds - you would need a different population of drivers for that to happen. On this basis we know for sure that the TRL's system of roads classification was pure illusion, and therefore no real data was available to check against accident frequency.
Many ‘before and after’ studies of measures which slow traffic and result in substantially fewer accidents have also been reported. These measures include, for example, traffic calming schemes in 20mph zones – where injury accidents were more than halved (TRL Report 215).
We're getting used to "before and after" studies depending on a regression to the mean benefit illusion. We also accept that if a specific location is being driven at inappropriate speeds then measures to slow traffic will very likely be effective in reducing accidents. Having not read TRL215, we can't comment further.
These studies together provide extremely robust evidence of how speed affects accidents. They are large-scale studies, of real traffic on real roads, involving rigorous statistical analyses. The conclusions are unambiguous. Remember, 10 people die and 100 are seriously injured on our roads per day. Improvements in driver behaviour have the potential to cut these statistics dramatically; reducing drivers’ speeds will play a vital part in this.
Well. TRL323 says no such thing. TRL421 and TRL511 are flawed beyond redemption. This article seeks to twist and misrepresent the TRL's own research.
Conclusion and further reading

It's highly notable that they do not reiterate the "one third lie" (click here). In fact they don't even attempt to put a percentage to the number of accidents where excessive speed is a cause or contributory factor, except to misuse the "15%" calculation from TRL323 data.

We don't know why the TRL seek to mislead in this way. Perhaps certain staff members have deeply held beliefs about the relationship between speed and accidents. Perhaps they are contracted by the DfT to tell lies.

Whatever way it is spun the evidence for the speed accident relationship claimed is unlikely to be applicable to "normal drivers" and "speed limit enforcement". Our letter of 21st May 2003 to the TRL Chief Executive asks a number of absolutely critical questions that the research does not answer. (click here) We have had no reply at all to the letter.

It has become quite normal for road safety to be based on lies, flawed research and misrepresentations of research. See our "big lies" page. Quite how or why this has come about we have no idea.

It's highly illuminating to note that TRL421 makes unjustified leaps in its headline conclusions (click here), and perhaps even more illuminating to note that the authors of TRL421 are also the authors of this nasty little newsletter article. In fact as far as we can tell all the modern rubbish which comes out of TRL has Marie Taylor's name on it.

new Comments

Safe Speed encourages comments, further information and participation from our visitors. See our (forums).

Read about our comments policy (here).

Many pages (including this one) have a specific associated forum topic. You can (view) or (add comment) to the forum topic for this page. Posting in the forum requires simple registration.

Calling for real road safety, based on truth

We have a strict editorial policy regarding factual content. If any fact anywhere on this web site can be shown to be incorrect we promise to remove it or correct it as soon as possible.
Copyright © SafeSpeed 2003, 2004
Created 13/09/2003. Last update 8/03/2004