Dear Dr Sharland

As a road safety campaigner I follow developments eagerly and in much detail. Obviously the Transport Research Laboratory has a well-deserved international reputation for excellence and in many quarters the word of the TRL is taken as gospel.

But there is something altogether rotten going on. You must act now to preserve the TRL's reputation. Certain staff members are publishing reports containing headline conclusions with little or no basis in scientific fact.

Let us examine a single critical example from TRL421 for a moment.

TRL421 includes the claim: “The percentage reduction in accident frequency achievable per 1mile/h reduction in average speed is between 2-7%. The earlier 5% figure remains a robust general rule. The reduction achievable, however, varies according to the road type and the average traffic speed.”

But TRL421 has not even attempted to establish that the claimed relationship between speed and accidents is a causal one. We believe the causality to be complex, and a very large component is simply a common cause. (The common cause being, at its simplest, bad driving leading to accidents and also leading to incautious and inappropriate speeds.) Without establishing causality it is clearly false to suggest that any “reduction in accident frequency is achievable”.

It is neither fact nor science to make that claim for accident reduction in TRL421, yet the claim from TRL421 is frequently cited in reports to government such as your own document RTS27. The claim (and others) is supporting a modern road safety policy that appears to be reversing our very long-term beneficial trend in the UK fatal accident rate.

Our view is that TRL421 and its ilk are more politics than science and should never have been published by the TRL. I could cite numerous other examples of similar critical failings, and I would be pleased to provide any details that you might require in due course.

I would appreciate your thoughts on this deadly serious matter most urgently.

Yours sincerely
 

Paul Smith

Dear Dr Sharland

I refer to my previous letter and your reply dated 3rd April. I am pleased that you are able to fully support TRL421. But I am baffled by your support and would very much like to understand better. Perhaps you could specifically answer the two emboldened questions later in this letter:

TRL421 states: “The percentage reduction in accident frequency achievable per 1mile/h reduction in average speed is between 2-7%. The earlier 5% figure remains a robust general rule. The reduction achievable, however, varies according to the road type and the average traffic speed.” 

I am extremely confident that there is no such causal link, let alone credible evidence of it, and TRL421 itself even states:

It is important to stress at this stage however that these models merely represent associations between the variables. These associations represented by the models may arise from a causal link between speed and accidents or from causal links between accidents and a number of key variables (age or mileage for example) and between speed and the same variables.

Here is an important example:

From: http://www.roads.dft.gov.uk/roadsafety/strategy/speedmanagement/03.htm
 

37. New research (Taylor et al 2000) has examined the scope for reducing collisions through speed management. Broadly each 1 mph reduction in average speed is expected to cut accident frequency by 5%. This is a robust general rule, but now we have a much fuller picture which indicates that the reduction varies according to road type as follows:

about 6% for urban main roads and residential roads with low average speeds; 
about 4% for medium speed urban roads and lower speed rural main roads; and 
about 3% for the higher speed urban roads and rural single carriageway main roads.

Out of interest, and to assist others who might read this correspondence in due course, I would like to include a story about trousers.

Suppose we observe a population of British adults and discover that taller people are more likely to be wearing trousers. Does that mean the if we give away trousers it will make people taller? Or perhaps if we travel to Africa and observe a tribe of pygmies that they are unlikely to be wearing trousers? Maybe we could restrict peoples' growth in some way and influence the patterns of trouser wearing in the future? No. The simple fact is that men are likely to be taller than women and also more likely to be wearing trousers. There's a link. We can prove it with sampling, but the link is not a causal one.

Without discovering what causes taller people to be more likely to be wearing trousers we've learned little or nothing from the observation, and we would obviously be daft to try to reduce average height by collecting up trousers.

I have suggested in previous correspondence that I consider this matter to be deadly serious. Please be sure to give it the attention it deserves. I shall not rest until I have satisfactory answers.

Yours sincerely
 
 
 
 

Paul Smith

Dear Dr Sharland

I refer to my previous letters and your reply dated 6th May.

I note that you have not attempted to address the specific questions in my letter dated 8th April, and it seems clear to me that you have no adequate answers to offer.

The authors of TRL421 clearly made a completely unjustified leap of faith in the headline conclusions, as you very well know, and in my opinion that reflects on their integrity or competence to the extent that the entire report is invalidated.

UK road safety policy is being based in part on wild conclusions in TRL research which have no basis in science or observable fact. The situation is outrageous, dangerous and unacceptable.

The TRL has always been a highly respected body with a well-deserved international reputation for excellence. Indeed, I have long respected it myself, and have absolutely no desire to damage the TRL's reputation. But if you are unprepared to assist in setting the record straight about TRL421, I shall have no alternative but to take my case to the press and the public.

Unfortunately TRL421 is not the only problem. I have similar concerns about TRL511 and subsidiary documents including RTS27 and the TRL Newsletter article from September 2002 entitled “Speed and accidents – let's put the record straight”. 

I believe a telephone conversation would be beneficial. I give you my word that I shall not record the call or use it against you in any way. Please have Sue Surridge call me to arrange a mutually convenient time for the call. Alternatively a substantial reply in writing would be perfectly acceptable.

I look forward to hearing from you in due course.

Yours sincerely
 
 
 
 

Paul Smith

Dear Dr Sharland

I refer to my previous letters and your reply dated 16th May.

We are finding it difficult to make any progress and this will be my last letter to you on the subject. Please afford me the courtesy of answering the following points individually and specifically.

1. In your last letter, you refer to “the evidence relating speed to accident frequency”, Can you give me any reference to any work that establishes speed in excess of a speed limit as being causally related to accident frequency?

2. Can you give me any reference to research into speed and accidents where causes in common have been eliminated in order to determine the relationship between speed and accidents for normal drivers exceeding a speed limit? (In particular it is important to eliminate the obvious potential distortions caused by: speeding drunks, speeding escaping criminals, speeding joyriders in stolen cars etc, where the cause of the speed and the cause of the accidents may be the same, but are not causal linked, and cannot be usefully applied to “normal motorists”.)

3. Can you give me any reference to research where the link between speed and driver concentration has been explored? In particular I am very concerned that increased speed enforcement on motorways would reduce concentration and lead to increases in sleep related accidents. It is also quite probable that there would be an increase in inattention accidents.

4. Can you give me any reference to research where the proportion of inappropriate speed accidents above and below the speed limit has been investigated? There is some Canadian research in this area which appears to indicate that two thirds of inappropriate speed accidents take place within the prevailing speed limit. If that research can be properly applied to the UK, then there are some figures being widely quoted which need to be reduced by 67%.

5. Can you give me any reference to research where the negative effects of increased speed enforcement have been explored? I am particularly concerned about impacts on “drivers’ priorities” where at a critical instant they might be more concerned about compliance with a legal speed limit instead of an unseen dangerous situation developing ahead. Other important negative factors include: Risk compensation: Where a driver at a lower speed may unknowingly preserve risk values by driving closer or more aggressively. Attitude effects: Where increasing the restrictions on a driver may make him care less about his driving in general. Traffic displacement: where an increase in speed enforcement on one route causes drivers to chose an alternative route which is more dangerous.

6. Can you give any explanation for the loss of trend in the UK fatal accident rate? We have enjoyed a long-term beneficial history showing an average 5% per annum reduction, but in the last decade it has stalled and is likely to show a real increase for the first time in modern history when the 2002 figures are announced in a few weeks’ time.

7. Have you got any idea why the UK roads are just about the safest in the World? I suggest that it is largely a “Hendon halo” where the techniques and attitudes from the Police Driver Training Establishment at Hendon have pervaded all aspects of UK driving culture via official recommendations, legislation and the driving test. Whatever the cause(s), we would clearly be extremely unwise to turn our backs on the principles which have made our roads the safest in the World, but that is exactly what we are doing.

8. Will you make any comment at all about the claim in TRL421 which started this correspondence? The silence has been deafening. I pointed out that it was neither fact nor science to claim a potential reduction in accidents on the basis of the reported speed/accident relationship unless causality was also established.

9. And finally, have you ever involved highly trained drivers when setting research objectives? You might be surprised and enlightened by the perspectives that they have to offer.

I have published our previous correspondence to the Internet at: 

http://www.safespeed.org.uk/trl.html [this page]

I look forward to hearing from you in due course.

Yours sincerely
 

Paul Smith