Road Safety News

Deaths and serious injuries down a quarter near speed cameras: RAC Foundation

Friday 7th June 2013

A new report published today by the RAC Foundation includes analysis which shows that on average the number of fatal and serious collisions in the vicinity of fixed speed cameras fell by more than a quarter (27%) after their installation.

The report, compiled by professor Richard Allsop of University College London, includes analysis of data for 551 fixed speed cameras in nine locations. In addition to the fall in fatal and serious injuries, the analysis shows an average reduction of 15% in personal injury collisions in the vicinity of the cameras.

However the research also highlights 21 camera sites (in these areas) at which, or near which, the number of collisions appears to have risen enough to make the cameras worthy of investigation in case they have contributed to the increases.

The data analysed in the report, Guidance on the use of Speed Camera Transparancy Data, was released in 2011 as part of a government move to make speed camera operations more transparent to the public.

The estimates for collision reduction were made allowing for the more general downward trend in the number of collisions in the nine areas in recent years, and for the effect of regression to the mean at sites where collision numbers were unusually high in the period before the cameras were installed.

The study comes in the wake of  the 2011 instruction from government that speed camera data going back to 1990, detailing accident statistics before and after fixed speed cameras were installed, be made publically available.

Since 2011, only 12 out of 36 of the organisations (a mixture of councils, police forces and safer roads partnerships) responsible for the figures have published the information in a format which complies with official DfT guidance.

Click here to read the full RAC Foundation news release.


Comment on this story
Report a reader comment

What's your view - comment on this story:

I confirm that I have read and accept the moderation policy and house rules relating to comments posted on this website.
Your comment:
Your name and location:
Your email:

With respect to Mr Owen's comment I would submit that, since the vast majority of studies on speed limits and enforcement are done by the same government organizations which imposed such limits, what do you think would happen to a report that the limit accomplished nothing positive? Answer: It would be quietly shelved never to be seen again. However, now and then a long-time 'silly-servant' (like myself) gets pensioned off (like me) and takes an interest in traffic safety (like me). Having thus been freed from the compelling necessity of job preservation and thus able to follow wherever science leads her/him the truth will out! I was able to investigate the first European imposition of speed limits in late fall of 1973. Although the politicians who had imposed those limits claimed that the large reduction in the motorway fatality rate was due to those limits I was able to show that what happened in only some countries, i.e. the speed limits, could not possibly be responsible for the large reductions in motorway fatality rates which happened in every country in 1974.
see details in'97_detroit_paper).htm
Al Gullon, Ottawa, ON, Canada

Agree (6) | Disagree (2)

I am very interested in whether people understand Allsop's report. The conclusions of every report depend on the data presented and the methods of analysis employed therefore these are the areas I tend to look at. I would like to ask:

Do readers feel they
1) understand regression to the mean (RTM)?
2) understand the method Allsop used to account for RTM?
3) understand the potential of the method and its limitations?

Questions such as these are, I feel, essential when considering all speed camera reports and many road safety reports generally.
Dave Finney, Slough

Agree (9) | Disagree (2)

Mr Jones. To clarify the matter, I was looking at a continuous video covering a half hour time frame. The PC was targeting every vehicle without exception and the equipment tagged every reading over 40mph for easy retrieval when he reviewed the tape at the end of the shift.
Nick Elmslie, New Milton

Agree (7) | Disagree (0)

Just one last (respectful) comment from me to wrap things up. Eric is right - I was for many years involved in speed management/enforcement/traffic and road safety etc. And whilst that background would indeed cause some raised eyebrows if I was a regular contributor to a forum on gynaecology or space flight, this is after all, a road safety forum for - one presumes – people with an appropriate background (it doesn’t have to be exclusively professional) to contribute to, so it’s hardly a bombshell.

This forum seems to be open to anyone however, so my comments about two new contributors were to hint at the that fact that not every contributor does has an appropriate background and their motivation to comment may not be the same as those who do and perhaps their statements and claims should be read with this in mind.
Hugh Jones, Cheshire

Agree (3) | Disagree (4)

Just to clear up any misunderstandings - it's not that I have not accepted the report in question, but as I said in one of my earlier comments, I have always known that accidents at any particular camera site could go up, or down, or stay the same,i.e the cameras would not make any difference which is why the report was not of great interest to me. The cameras' benefits extends far beyond their site of installation but the report didn't mention that so I didn't think it painted as full a picture as it could have done.
Hugh Jones, Cheshire

Agree (2) | Disagree (7)

ALL CONTRIBUTORS: this thread is in danger of becoming too personal and descending into an unseemly spat. Please consider this if you intend to make further contributions. Thanks for your cooperation.
Nick Rawlings, editor, Road Safety GB newsfeed

Agree (5) | Disagree (1)

It is unfortunate that Hugh Jones chooses to vilify new or infrequent contributors, who show some fresh and enlightened thinking.

I believe Mr Jones once earned his living operating or managing speed enforcement. That is why he cannot accept that no report, including Prof Allsop's latest, provides any evidence that speed enforcement has a net road safety (casualty reduction) benefit.

I also note Rod King believes that there are "huge benefits" from zero-tolerance speed enforcement. He cannot be aware of the 40 negative effects of speed enforcement, nor the deaths attributable to speed enforcement.

Elmslie and Walker, I salute you.
Eric Bridgstock, Independent Road Safety Research, St Albans

Agree (6) | Disagree (5)

Mr Elmslie: Are you an expert on these matters or someone who has fallen foul of speed enforcement possibly? Your comments have the familiar ring of one who is in denial and looks desperately for reasons why an offence really wasn’t that bad at all. You nearly had me fooled for a moment. i.e. “Not one car passed at less than 33mph”. You were looking at a video only of those vehicles whose speeds were at or above the level that the camera operator might record the speed of and would not show the vast majority who were driving below the limit.
Hugh Jones, Cheshire

Agree (3) | Disagree (8)

Thanks for putting us right on all that Mr Walker. Are these your own findings, or are they what you’ve read on an anti-speed camera website somewhere? When putting pen to paper it’s always best to check what one is saying is actually correct, as in some cases it can make the writer look uninformed and biased - welcome to Road Safety GB.
Hugh Jones, Cheshire

Agree (4) | Disagree (8)

The thing that many folk choose to ignore or forget is that speed limits were initially set at least partly based upon the technical capabilities of motor vehicles many decades ago. They were also set below the level that would have been reasonably "safe" on the understanding that large numbers of drivers would overstep the mark. The effects of this under setting of limits has always been visible just by observing the speed that cars are travelling in any uncongested 30mph limit. I was once able to view 30 minutes of a camera van video. Not one car passed at less than 33mph. The device was set to enforce at 41mph. Anyone doing 40 or less was not set a NIP.

Even the Police were unwilling to prosecute every driver that passed. When I first started driving a police car would follow you and if your driving was above the limit but not unsafe you would be left to carry on without being stopped. If you were stopped where only the excess speed was at issue you would often be given a short lecture and allowed to proceed. You were far more likely to be prosecuted for other, in my opinion, far more important offences.
Nick Elmslie, New Milton

Agree (8) | Disagree (3)

Whilst there are many views on static speed cameras, the flaw seems to be in the idea that we set a speed limit that we wish to be complied with everywhere and then we rely on static and highly visible fixed cameras for the majority of the enforcement.

This is the same approach as with physically calmed and isolated 20mph zones. There are highly effective locally, but communicate the idea that these are the most important places to comply and it is less important to comply elsewhere.

If we are to be serious about setting maximum speeds for not only safety reasons, but for convenience, emissions, noise, liveability, etc, then we must get serious about enforcement.

I believe that there would be huge benefits from a zero-tolerance of speeding based upon random and covert manned and unmanned cameras. We also have emerging technology which will make them even cheaper and more discrete. Maybe unpopular at first, but hugely effective.
Rod King, Cheshire 20's Plenty for Us

Agree (8) | Disagree (12)

About 20 years ago the Department of Transportation decided to abandon the well proven principle of setting speed limits at the 85th percentile speed of free flowing traffic under good conditions that had served the UK so well with continuous and substantial year over year reductions in the fatality and crash rates.

The "new" plan from the early 1990s and up to today is to set artificially low posted limits plus speed cameras, and to urge local authorities to set speed limits at the mean speed of traffic to define about 50% of all drivers as violators.

These totally improper methods are supported by the insurance industry who can surcharge the insurance premiums of safe drivers who get tickets issued for the "crime" of driving safely.
James C. Walker, Michigan, USA

Agree (10) | Disagree (6)

I have searched widely for "proof" that a high level emphasis on speed cameras and speed limit enforcement has a general effect on serious road trauma using my home state, UK, Netherlands & now France as references. There is absolutely no evidence of a measurable general effect. My approach is to look at the change in fatalities per 100 m KM over time, and especially over the time each jurisdiction has made dramatic changes in the levels of speed camera deployment or speed limit enforcement. There is no evidence of an additional decrease in fatality rate.

Strangely, the one set of site data in the report as Table 1 shows an increase in total casualties post the speed camera installation. Would have thought Allsop would have chosen a site where there was a reduction.

A non-random sample of cameras cannot be relied on in research. In respect of casualties at camera sites the report used data from 9 partnerships - 551 out of around 3,450 cameras or 16%. And in respect of speed the numbers are 132 out of 3,450 or 3.8%.

The confidence limits he describes are for the sites analysed - - they are not for all camera sites in UK. So his figures of 25% reduction with confidence levels of 13%-35% reduction mean nothing in the overall UK context.
John Lambert, Victoria Australia

Agree (8) | Disagree (6)

What is ground-breaking about Allsop's report is the method he used to exclude RTM effects. When the authorities said there was no method that could fully account for RTM at speed camera sites, I was surprised so I developed such a method. Reports are now making use of this method:

2009 The method was published on my website and I asked Professor Linda Mountain for comments
2010 (Jan) Professor Linda Mountain co-authors a report proposing using part of the method
2010 (Nov) Allsop uses part of the method to indicate levels of RTM
2012 (Feb) I use part of the method to report speeds and collisions at speed camera sites
2012 (Nov) I publish the 1st report to use the full method (mobile speed cameras)
2013 (June) Allsop's report uses the full method (fixed speed cameras)

I am very pleased to see Professors Mountain and Allsop proposing and using this method as it should bring a genuinely new understanding to a major problem in road safety analysis. If used correctly, the method could enable scarce resources to be best deployed thereby really saving lives.
Dave Finney, Slough

Agree (7) | Disagree (2)

Nick (Elmslie): I think the reason some disagreed with your comment is because they're rational and not easily swayed by rhetoric. How do you think Messrs Finney, Bridgstock and Francis feel now, having spent countless hours of their free time delving in to this subject over many years, producing stats and reports only for you to come along and clear it all up in one long paragraph?
Hugh Jones, Cheshire

Agree (7) | Disagree (5)

The amount of people breaking speed limits, it is obvious that people think they are set too low. But then as any road can have many 'safe' speeds depending on prevailing conditions, the whole thing is a road safety dead end.
Terry Hudson

Agree (8) | Disagree (5)

Looking at it from another more positive angle, I'm sure you spotted that twice as many people 'agreed ' with your post as 'disagreed'!
Nick Rawlings, editor, Road Safety GB newsfeed

Agree (5) | Disagree (2)

The fact that 5 people disagreed (so far) with my previous post tells me a lot about those people. I only made a comment on my observations of drivers reaction to cameras and upon the accidents that I witnessed. It's not possible to disagree as they were my first hand observations and yet 5 people bothered to hit the button. A blinkered view and an unwillingness to accept anything that doesn't "fit" is the only thing that would lead anyone to bother to disagree with my post. The same kind of thing is evident in many pro speed camera reports.
Nick Elmslie, New Milton

Agree (4) | Disagree (9)

I thank Richard Owen for his contribution and assure him that Allsop's methodology for excluding the RTM effect is the same as mine.

The method is simple: Identify the SSP and then exclude data from within that period.. Allsop gives an example on page13, Table 1. The final column assumes the SSP was 1996-1998 for that speed camera (the “1”s in the last column). Allsop then uses the years with 0,0 in the last 2 columns as his “before” data and the years with 1,0 as his “after” data.

As we can see, Allsop has separated the data into 4 periods which I call (in chronological order) PreSSP, SSP, ASBiC and after installation.

This method does rely on accurate identification of the SSP (that's where inaccuracies in Allsop's report may lie) but may I join with Richard Owen in looking forward to clarification by Professor Allsop of the method he used to exclude RTM effects?
Dave Finney, Slough

Agree (7) | Disagree (5)

It appears that the debate has raged for some time and with some familiar views expressed by the usual suspects. For the record, I don’t receive payment from camera manufacturers or course providers but I am employed to carry out analysis of road safety statistics and report on the findings.

I congratulate Dave Finney on his attempts to carry out his own analysis which must take up a lot of his time and although I disagree with his methodologies, it is much better to see a debate informed by analysis rather than opinions. I did address Mr Finney’s most recent report on the RSGB website and pointed out what I felt were the most significant flaws. Upon inspection of the RACF report and its methodology it appears that they are not the same (as Mr. Finney’s) but I think it would be best if the report’s authors could clarify that point.

Whatever your opinion on speed cameras they are only a small part of the road safety toolkit, albeit one loved by the popular press, and it would be a shame to see the same circular debates raging for another 20 years.
Richard Owen, Banbury

Agree (11) | Disagree (5)

Professor Allsop's report may be very significant. Analysing speed camera effectiveness is all about excluding regression to the mean (RTM) and Allsop explains his method on page3 from “For each camera...”. I urge people to read the 3 bullet points and the following paragraph.

Now read my method and you'll see they are the same:

What Allsop and myself are saying is that data that influenced site selection cannot be used to evaluate speed camera effectiveness. Linda Mountain and Mike Maher also proposed this. The accuracy of the method relies on identifying the SSP but the data Allsop used was not sufficiently detailed to enable him to do this. I would like to thank Richard Owen and his team for supplying the detailed database that did allow me to do this for mobile speed cameras.

We are, after 20 years of speed cameras, finally finding some agreement!
Dave Finney, Slough

Agree (10) | Disagree (6)

Comparing accident figures – before and after - at any camera site, will usually show, randomly, increases, decreases or no change. That’s not to criticise the use of cameras, as you would probably find the same thing after the introduction of some other types of remedial measures to address an accident cluster site. The reason is that, relative to the total number of vehicle movements, the number of actual collisions at a site over say, a three-year period, is actually too small (single figures sometimes) to be influenced.

The effectiveness of cameras - which reports never acknowledge – is more global than local and is as a result of the prosecution of the reckless and careless drivers, which leads either to a ban and therefore off the road and out of harm’s way, or a change for the better in the individual’s behaviour. That’s almost impossible to measure but shouldn’t be overlooked.
Hugh Jones, Cheshire

Agree (12) | Disagree (7)

In my previous employment I drove over one million miles in a ten year period. Fixed cameras and then mobiles were introduced roughly in the middle of that period. I noticed no reduction in general speeds even when drivers were unaware of the location or likely locations of the cameras and vans. Drivers only slowed momentarily as they saw the camera and then speeded back up, often to a slightly higher speed than before. During the covering of that million plus miles I saw a number of crashes as they happened and could fairly accurately judge the cause. Only on the motorway did I ever conclude that speed was a large factor. By speed I mean too fast for whatever was occurring on that stretch of road, not by any means speed over the 70mph limit. On all other roads the accidents I saw had speed as only a minor factor with poor observation/judgement by far the main cause.
Nick Elmslie, New Milton

Agree (14) | Disagree (10)

Anyone who thinks that vested interests do not play a big part in all areas of human activity, including speed cameras policy, should read Parkinson's Law on how humans and organisations behave. One such law is that from the moment any organisation is formed its primary objective becomes survival and expansion, not what it was set up to achieve.
Idris Francis Fight Back With Facts Petersfield

Agree (12) | Disagree (24)

To be clear, my use of "vested interests" does not refer to Road Safety Professionals but to organisations who work (typically) in the speed industry (camera manufacturers and providers of speed awareness courses) and the individuals and bodies who receive funding from those organisations while promoting the (questionable) benefits of their products and services.
Eric Bridgstock, Independent Road Safety Research, St Albans

Agree (12) | Disagree (23)

As an "amateur", I believe that the fact that I am not paid for my work and others are, does not add one jot of extra legitimacy to my views or arguments.

Road safety in this country would be far inferior if we did not have professionals using their expertise and experience to make our roads better places to be for all.

Of course that does not mean that I agree with every Road Safety Professional and their views. I have always felt that a candid, sometimes robust and positive debate always helps to develop better initiatives and outcomes.

Perhaps those objecting so strongly to professionalism and commercialisation would do better honing their arguments and presenting them more constructively and with less acrimony than perpetually harping on about "vested interests".
Rod King, Cheshire 20's Plenty for Us

Agree (24) | Disagree (14)

Professor Allsop has written a very good report that moves the debate on. He explains just how difficult it is to analyse speed camera data although doesn't mention that the difficulties would not have arisen had speed cameras been subjected to RCTs (scientific trials):

To exclude RTM effects, Allsop uses a short version of probably the best method available. The full method has been on my website for many years:

I used the full method to analyse mobile speed cameras:

So why such differences in the results? Allsop has assumed the SSP is the previous 3 years but he provides no evidence of this. Many sites in his report were pre-hypothication therefore, with variable ASBiCs and free choice of SSP, the SSP for many sites may have gone back 5 years before installation. In short, Allsop may still have substantial RTM in his results whereas my report uses the full method and therefore has no RTM in the final results.

Allsop states (Px) "no relationship is evident between changes in numbers of PIC and changes in speed" and my research also finds this:

We now have much to agree on: data sources, analysis methods and some conclusions.
Dave Finney, Slough

Agree (13) | Disagree (8)

Yet again marked preference for fantasy over fact! Not that I mind - I am not here to be popular but to be right.

John Lambert was, in the 1990's, in charge of the State of Victoria's camera policy and delighted with results - until he saw the same elsewhere without cameras.

He notes these words in the RAC Paper: "...little or no correlation between the estimated changes in number of collisions and in average speed. The correlation coefficient is 0.0037..... locations of the points from these cameras show no particular pattern within the scatter of 132 points".

Would anyone here who "dislikes" our objections now step up to the plate and explain how (in Scotland even more so) cameras are credited not only with bringing about reductions that do not relate in any way to speed, but also are 3 times greater than involve speed in the first place? (Please declare financial or employment interests)
Idris Francis Fight Back With Facts Petersfield

Agree (13) | Disagree (21)

Nick contrasts the views/approaches of Prof Allsop and Richard Owen with those of Idris Francis and me. Idris Francis and I have not received a penny in connection with our road safety campaigning, and do not expect to.
Eric Bridgstock, Independent Road Safety Research, St Albans

Agree (12) | Disagree (24)

Before I took up this role I assumed data analysis and interpretation was a fairly straightforward exercise - how wrong I was!

We have diametrically opposed views expressed in this news item and discussion thread. On the one hand professor Allsop and Richard Owen; on the other Eric Bridgstock and Idris Francis.

The beauty of this website is that people with very different views can express them - and our readers can indicate with whom they concur, via the 'agree/disagree' facility, or through additional posts.

As always, it will be fascinating to see what others think, and who our readers believe to be correctly interpreting the data.
Nick Rawlings, editor, Road Safety GB newsfeed

Agree (28) | Disagree (1)

Cameras reducing speeds by 2-3 mph and speeding by <50% cannot reduce KSI by 2.8 times the 9% that involve speeding, even as secondary or "possible" causes. Minimal Partnership data, innumerable errors and omissions, as my own research confirms (e.g. 90% falls in traffic volume after/before at some Scottish sites, 17mph and even 30mph increases at others!). At 100 out of 429 sites speeds either did not fall or rose - though that did not prevent accident reductions there being claimed as camera benefit.

Mr. Owen has "yet to see a report of similar quality" demonstrating other than success. Has he forgotten already the excellent Finney Report (RSGB Knowledge Centre) that all reductions in Mr Owen's own area, Thames Valley, happened before cameras started operation? As indeed they almost invariably do across the country, due to regression to the mean?

This RAC report is not remotely credible.
Idris Francis Fight Back With Facts Petersfield

Agree (14) | Disagree (34)

A report of interest only to the vested commercial interests.
Eric Bridgstock, Independent Road Safety Research, St Albans

Agree (12) | Disagree (39)

Another excellent analysis of speed camera performance by the RACF. The level of statistical analysis, while daunting to some, really lends weight to the findings. It will take me another few hours to digest the results in their entirety but they certainly match my own previous analysis (24% reduction vs long term trend in Berkshire). I have yet to see a report of similar quality from any source that demonstrates anything other than the repeated success of speed enforcement in reducing road casualties.
Richard Owen, Banbury

Agree (28) | Disagree (15)