Safe Speed Forums

On the CSCP forum, the "2004 fatalities" thread (http://www.cumbriasafetycameras.org/forum/index.php?showtopic=857) the anti-camera movement is making quite a bit from the increased K, while the Scameraship itself refers to the combined figure and shines their light on the overall reduced figure. I've been thinking how the ratio may be changing, and I believe it has a lot to do with secondary safety in vehicle engineering. I've just placed the following post on the referred CSCP thread in response to a post proclaiming that the ratio of K to SI cannot change in this way. I think the "cannot" needed challenging. comments invited:

Can't be Roger, although it might be some part of it. The same sort of ratio changes apply to pedestrians and cyclists too. See:

http://www.safespeed.org.uk/serious.html
http://www.safespeed.org.uk/serious2.html

I also had a go at modelling the idea you have put forward (based in part on the concept "you can't be worse than dead") and I didn't produce anything even remotely convincing. PLEASE show me how it works if you can make out a convincing case.

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

I am putting together a simple (!) excel sheet. Back later with it.

The spreadsheet may not be ready 'til tomorrow - I am being summoned by someone close to me..

HOWEVER - a little more thought and possiblyt a simpler explanation ?

Pre-cameras, lots of concentration, most accidents scrub off speed in the run-up thanks to driver concentration and natural behaviour - more slower accidents, fewer faster accidents, more injuries less deaths.

Post cameras - lots of concentration on speedo - more driving to the needle, less slowing on route to a collision.....

I was reading a comment made by someone else on another website and they were saying the definition of a serious injury has changed, hence the fall in SI figures but the rise in fatalities. If a serious injury has been tightened up so there are fewer of them then they were then that explains why fatalities as a proportion of SIs are increasing. If you could find out what previous years SI figures using the new definition then you could see exactly what is happening rather than the spun version.

Personally, I think it's more to do with the magician's 'multiple-out' technique.

I.e., if serious injury's go down but deaths go up, they publicize the serious injuries and claim that that was what they were trying to reduce.

If the opposite happens, they publish the deaths and claim that that was what they were aiming for.

If both go up, the publish some other result like 'personal injury collisons' (easy to define however you want), or just say 'we are working harder than ever to reduce accidents, and will be following up our campaign with a wide range of new measures' (even though what they've tried so far has failed).

AIUI, the definition of "serious injury" has changed to incorporate an overnight stay or longer in hospital. If this is so, the decline of the Health Service could be going a long way to "improving" the KSI figures. If only because of lack of beds, hospitals now routinely discharge patients that until recently would have been kept in overnight under observation. So, anyone coming in from an RTA who would have been kept in under observation but who is discharged represents a reduction of one in the KSI statistics.

IOW, the decline of the Health Service may be responsible for a reduction in KSI even in the face of an increase in RTAs.

_________________
Will

The definition hasn't changed since well before 1960. But there's some evidence of "external influences".

Study this: http://www.safespeed.org.uk/serious.html

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

A possible explanation lies in the following passage copied from the DfT document: Road Casualties in Great Britain Main Results: 2003

Thanks for that. My basic premise was that external influences (i.e. other than road safety) were "massaging" the figures. While the definition hasn't changed, the way that casualties are treated has, which implicitly alters the SI numbers. Taken with the other factors you cite (and perhaps some you've missed:) the effect is to reduce the SI numbers even without reduction in the number or severity of incidents. That reduction masks the number of fatalities when rolled up into KSI. which is what the scameraphiles use to hide their failure.

_________________
Will

Add to that the fact that there must be many cases where "casualties" used to be taken to A&E but now aren't because the ambulance will be trapped for hours waiting for A&E to cut it's queues before it is allowed to add it's "casualty" to their list and respond to the backlog of calls it's failed to answer.

_________________
http://www.itsyourduty.org.uk

I too have studied this in some depth, and the only plausible explanation I can come up with is that "something" is causing a percentage of "SI" accidents to be downgraded as "Slight" injury. As Paul has demonstrated with his excellent analysis this only needs to be a relatively small percentage in order for the whole thing to suddenly make sense.

The notion that fatalities AND slight injuries can increase whilst serious injuries fall is little short of ludicrous.

_________________
CSCP Latin for beginners...
Ticketo ergo sum : I scam therefore I am!

Your logic implies that an accident gives injuries on a kind of sliding scale between none, slight, serious and death. Do you accept that the speed of the accident partly accounts for that? After all, very common collisions at low speeds between cars generally result in no injuries which (until recently) often go unreported, except to insurance companies who pay for the new bumpers and lights etc. A heavy collision is generally much more serious, especially if a pedestrian is involved.

If you do accept that speed partly accounts for it, then it follows that speed reductions would be prioritised in areas where serious or deadly accidents are most likely. Areas where slight accidents (despite their high frequency) happen would be un-prioritised, and therefore likely not to fall (or even to rise). It then follows that such measures affect the KSI figures alone and your argument boils down to saying that it is ludicrous for fatalities to increase whilst serious injuries fall. But fatalities have not increased, they have stayed steady or dropped despite more traffic. If we see the K figure relatively unchanged (despite massive increase in distance travelled) and the SI figure lower, that is far from ludicrous – it is a good result for cameras.

If you do not accept that speed partly accounts for it, then a heavy accident would have an equal chance as a light accident of resulting in no injury. And a light accident would have an equal chance as a heavy accident of resulting in serious injury or death. Yet experience shows otherwise, doesn’t it?

_________________
I stole this .sig

Actually, fatalities have increased - 2003 was up on 2002 remember?

I fully accept your model that a correlation exists between impact speed and injury severity, it is that that makes the current stats look so suspicious.

Accepting this as a premise, then for "SI"s to genuinely fall one of two things needs to have happened: either a general drop in the number of collisions, or a reduction in impact speeds (or of course both!).

Taking the first notion, we know this hasn't happened - there hasn't been a significant drop in the total number of accidents which would generate a pro rata reduction in SIs.

Now consider the alternate theory - that impact speeds have reduced. Again it doesn't fit, even if we also accept the notion that the effects of cameras have been specifically targeted at a particular type of accident. If they had somehow been successfully targeted at "severe" accidents then how come they haven't had any effect on the severest accidents of all - the fatal ones? If they were reducing impact speeds such that a significant number of SIs were mitigated to "slight", then a similar proportion of fatals would surely have been similarly mitigated to "SI"s.

Even if the opposite were true, that camera activity had somehow been targetted away from fatal accidents, then the opposite effect would surely have happened - for each SI mitigated to a "slight" then there would be several "slights" that ceased to be injury accidents at all.

Sorry, but the only model that truly fits the current statistics is that the definition of "SI" has been varied downwards by a few percent. I fully accept that this effect may in itself be unintentional, though I don't accept that the way the DfT / SCPs lean on it as a misleading barometer of success is unintentional.

_________________
CSCP Latin for beginners...
Ticketo ergo sum : I scam therefore I am!

and 2002 was down on 2001. Look at SafeSpeed's graph pages - the figures are up and down like a fiddler's elbow over a period of a year or three. I'm talking about a significant change in the fundamentals, not statistical jitter. The figures now are better than the 90's, despite more traffic. Simple as that.



Over a period, I'd expect the number of Ks and SIs to relate to the frequency (total number) and severity (speed) of collisions, and therefore for Ks and SIs to relate closely to each other. Again, look at the graphs and see the spread of the yearly deviations. Ks and SIs are linked, but the run of data you are using is too short. I'm not saying that the disparity is random - there are underlying reasons, such as the weather, the state of the nation and what's on telly! But they can't be modelled easily.

_________________
I stole this .sig

I have looked at the graph pages - the underlying trend is that the fatality rate (ie taking traffic levels into account) was falling, levelled off over the last ten years, and has now for the first time started to rise (which I seem to recall was predicted by Paul).


If you can come up with a model whereby cameras cause some SIs to worsen into fatalities yet others to improve into "slights" then you might be onto something. If not then I'll continue to believe what seems so far to be the most plausible explanation - that cameras are having no benefit, that the drop in SIs is due to changes in their recording or classification, and that the "real" picture is that all the accident classifications are following approximately the same trends.

_________________
CSCP Latin for beginners...
Ticketo ergo sum : I scam therefore I am!

This is the graph that I find most revealing:



Definitions, references etc on page:
http://www.safespeed.org.uk/smeed.html

The drop in 2002-2003 is the first real (i.e. non-blip) fall in the history of motoring.

The rise in 1997-1998, I believe is due to excluding deaths from hospital contracted infections following a road crash from the stats.

The divergency from trend started in 1993/4 and has been progressively worse since then. We built up to the trend reversal in 2003.

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

Here is my chart from the figures quoted in another thread of the relationship between Killed, Serious Injury and Minor Injury per VKM using 1991 as the 100% start point.



The SIs still show a drop in comparison with Ks but not as dramatic as other representations.

I believe that the rise in proportion of fatalities caused by the non compliant road users could be in part responsible for the strange movements in the figures. The fatalities among compliant groups is IMO coming down inexorably, almost as always, whereas the other road users, the high risk groups have increased, and despite their small numbers they are having a dramatic effect on the fatality stats.

In many of the non compliant groups, eg boy racers, drink drug drivers, weekend bikers, the ratio of K to SI is (although I have not analysed it) probably significantly higher than the ratio of collisions in other road user groups. Reasons could be

• older cars,
• impaired ability to avoid collisions,
• impairment preventing natural survival instincts,
• a higher percentage of dramatic collisions causing a higher percentage of fatality,
• more chance of drink drivers and new drivers losing control for no reason causing vulnerable road user fatalities.
• Probably numerous other reasons which I can't think of at the moment..

These non compliant groups are increasing rather than diminishing, and I believe this battle between the two groups is possibly responsible for the strange figures.
I put this forward as a theoretical explanation with not a lot of fact to back it up other that we do seem to be dealing with more and more fatalities caused by non compliant groups these days.

Police officers at RTCs provide the injury classification, so I do not accept that they are being pressurised to change the classification as has been suggested. I think the BiB on here would be letting the cat out of the bag if that was the case. I know I would.

_________________
Fixed ideas are like cramp, for instance in the foot, yet the best remedy is to step on them.

Ian

This sounds a reasonably plausible explanation for the change in the proportions.

It's certainly the case that pretty much every time you hear of some terrible fatal crash, the news report goes on to say "a man has been charged with driving without insurance/while disqualified" etc.

Do we have any figures for the proportion of fatal crashes over time that have involved at least one uninsured, unlicensed or disqualified driver?

_________________
"Show me someone who says that they have never exceeded a speed limit, and I'll show you a liar, or a menace." (Austin Williams - Director, Transport Research Group)

Any views expressed in this post are personal opinions and may not represent the views of Safe Speed

Interesting to compare Ian's graph with this one:



I believe the trends from 1990 to 1993/4 were far from typical due to the recession at the time. It's also important for us to look at changes that have been taking place in the "speed camera decade".

I agree with Ian's suggestion that high severity crashes by "non compliant" road users are on the increase. Unlike Ian, I blame speed cameras for this increase by the following mechanisms:

• Policy switched resources from traffic policing to camera policing.
• Camera policing tends to create lawless behaviours (like avoiding vehicle registration requirements).
• Traffic police deter lawless behaviours - fewer trafpol means more lawless behaviour.
• Speed cameras cause frustration. Frustration causes rebellious behaviour. We have a new generation of rebel bikers to show for it.

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety