Discussion
Destroyer, you is OWNED!!!
Statistics...I don't trust any of them unless i can see a detailed summary of the raw data and am allowed to draw my own conclusions...and being a BSc in Maths, surprisingly enough I can!!! Anyone got that data for me to crunch?!?
I especially don't trust people who don't even have the sense to keep their quoted "statistics" consistent from post to post...strong suggestions that they are MAKING IT UP!!!
destroyer said:
Speed Cameras were found to have a much smaller effect - the research team were of the opinion that (for a currently undisclosed reason) the result was insignificant.
Therefore, your original post is insignificant!
Read it all for yourself on the Liverpool Uni website.
You obviously didn't read my post on the other thread as I said I couldn't find it (i think another PHer said it had disappeared?)
destroyer said:
I'm not surprised that you are willing to accept view you favour with UNDISCLOSED REASONS to support them. It's pretty typical of single sighted nuggets!(sic)
This statement is irrelevant. I was not using undisclosed reasons to support my view that the findings were insignificant. The RESEARCH TEAM said that the result for speed cameras was irrelevant (not me). I said that the reason for saying that they were insignificant was unobtainable at the present moment. Therefore we have to accept the word of the team leader that the reason is valid until such a time when we can read the full report and find out the reasons for insignificance.
jewhoo said:
4. 38% fall in personal injury accidents and 30% fall in fatal and serious accidents attributable to impact on speed
destroyer in first post said:
6. 19% fall in slight and 6% fall in fatal and serious accidents due to impact speed
Care to explain why you feel the need to double one statistic and increase the other five-fold?
Well?
Seems that really awkward questions are left unaswered while the less difficult questions are answered with statistics that were collected without any sort of scientific methodology and then heavily "analysed" to provide mis-leading conclusions. I've yet to find a source of pro-speed camera data that is has well laid out, open and convincing as the case made on Paul Smith's website.
MilnerR said:
Seems that really awkward questions are left unaswered while the less difficult questions are answered with statistics that were collected without any sort of scientific methodology and then heavily "analysed" to provide mis-leading conclusions. I've yet to find a source of pro-speed camera data that is has well laid out, open and convincing as the case made on Paul Smith's website.
Indeed. For example, in a previous post, where Destroyer said that cameras were catching the correct people, I enquired as to whether he meant the safer driver. (Given that the RAC and insurance company studies have shown that it's the safer crash-free driver who is the vast majority of those caught.)
I didn't get an answer.
havoc said:I found the original study, it's utterly useless, no data at all.
Destroyer, you is OWNED!!!
Statistics...I don't trust any of them unless i can see a detailed summary of the raw data and am allowed to draw my own conclusions...and being a BSc in Maths, surprisingly enough I can!!! Anyone got that data for me to crunch?!?
Funny how Destroyer couldn't supply this info!!
host site
the study
(it's a slow site, it even timed out on me)
If you were employed in a nice cushy little 'non-job', you too might be tempted to trot out the 'corporate spin' without too much concern about its honesty. You, too, might be tempted to just keep on taking your wages, paying into your pension scheme and p*****g off your fellow man for no justified reason except cash.
The rubbish, spin, lies and obfuscation we have all experienced from Steve 'S.G.' Callaghen over the past couple of years has been almost unbelievable, until you think what take-home pay he's had from this over that time for achieving nothing except cash collection for his masters.
The rubbish, spin, lies and obfuscation we have all experienced from Steve 'S.G.' Callaghen over the past couple of years has been almost unbelievable, until you think what take-home pay he's had from this over that time for achieving nothing except cash collection for his masters.
Yep, had a look at that, and quite honestly I can glean nothing trustworthy from it. The conclusions are not supported by sufficient data to allow fair interpretation.
The formulae, to me, look like "pick a statistical tool and lets see what it does".
The most interesting thing is the graph showing accident trends over the last 30 years. That clearly shows a statisticall consistent level of accidents across ALL 30 years, despite increasing traffic volumes.
To me, that chart shows a few things:-
1) The introduction of speed cameras HAS NOT reduced accidents.
2a) Given that the rate of increase of traffic flow since 1995 has reduced, I would, all other things being equal, have expected to see a drop in the accident curves. But they have not. Indicating the "rate of improvement" in the "accidents per km driven" is falling.
2b) Point 2a is exacerbated by the widespread improvements in braking (ABS, all-round disc brakes, wider tyres, better rubber compounds) since approx 1995. The last 10 years have been a period of rapid improvement in car-safety design, both preventitive and protective, yet the rate of reduction in accidents is slacking off.
3) Following on from points 2a and 2b, I can therefore conclude that there is at least one other factor which is negating the positive effect of
(a) the long-term trend in general; and
(b) the car-design improvements specifically.
Candidates for this are:-
- Reduction in average driver ability (be it observation, car control, judgement);
- Deterioration in road surfaces and general conditions (e.g. sight lines, drainage,...);
- Introduction of "distraction factors" (excessive signage, speed cameras.
Personally, I believe it to be a combination of all three.
Driver ability is falling, as people a) get reliant on their safety systems, ABS, crumple zones, and airbags. They are also getting sucked in by the "speed kills" message and are failing to use judgement anymore - exemplified by the rising number of "failed to see" or "failed to judge" accidents (e.g. rear-enders)
Road surfaces are clearly deteriorating...and this will have a minor negative impact on road safety.
Road signage HAS got more confusing recently, and speed cameras are a clear distraction - just look at the number of people who brake when they see them!
The formulae, to me, look like "pick a statistical tool and lets see what it does".
The most interesting thing is the graph showing accident trends over the last 30 years. That clearly shows a statisticall consistent level of accidents across ALL 30 years, despite increasing traffic volumes.
To me, that chart shows a few things:-
1) The introduction of speed cameras HAS NOT reduced accidents.
2a) Given that the rate of increase of traffic flow since 1995 has reduced, I would, all other things being equal, have expected to see a drop in the accident curves. But they have not. Indicating the "rate of improvement" in the "accidents per km driven" is falling.
2b) Point 2a is exacerbated by the widespread improvements in braking (ABS, all-round disc brakes, wider tyres, better rubber compounds) since approx 1995. The last 10 years have been a period of rapid improvement in car-safety design, both preventitive and protective, yet the rate of reduction in accidents is slacking off.
3) Following on from points 2a and 2b, I can therefore conclude that there is at least one other factor which is negating the positive effect of
(a) the long-term trend in general; and
(b) the car-design improvements specifically.
Candidates for this are:-
- Reduction in average driver ability (be it observation, car control, judgement);
- Deterioration in road surfaces and general conditions (e.g. sight lines, drainage,...);
- Introduction of "distraction factors" (excessive signage, speed cameras.
Personally, I believe it to be a combination of all three.
Driver ability is falling, as people a) get reliant on their safety systems, ABS, crumple zones, and airbags. They are also getting sucked in by the "speed kills" message and are failing to use judgement anymore - exemplified by the rising number of "failed to see" or "failed to judge" accidents (e.g. rear-enders)
Road surfaces are clearly deteriorating...and this will have a minor negative impact on road safety.
Road signage HAS got more confusing recently, and speed cameras are a clear distraction - just look at the number of people who brake when they see them!
destroyer said:
Read it all for yourself on the Liverpool Uni website.
Where? I can't find it. However I found a remarkably similar, London-based study, that miraculously comes up with some extraordinarily similar statistics - hosted on the DfT website. Collusion, coincidence or statistical clustering?
destroyer said:
Here you go James:
It would seem that the only part of the information from the report that the comments here refer to are the newspaper headlines. Some of the conclusions of the report went like this:
1. Cameras are most effective on roads with higher speeds
2. Typically a 4% fall in accidents per 1 mph reduction in speed
3. 8.4 mph fall in mean speed
4. 38% fall in personal injury accidents and 30% fall in fatal and serious accidents attributable to impact on speed
5. Cameras reduce accidents – less than observed change but worthwhile benefits
6. The methodology used allows for regression, trend, route changes and speed changes
Now how this reporter managed to make out that the cameras had no effect from those conclusions beats me but maybe he/she has some advanced analytical experience that can see through these conclusions and turn them on their head. I somehow doubt even the members here would believe that!
So there you have the response you requested, now can you get the headline and the comments to match the conclusions because I can’t. Perhaps you should look at the report rather than the c*#p the tabloid rags feed you next time and you will spend less time with your foot in your mouth.
Regards
The man behind a few cameras.
It's quite amazing how somebody who quotes so many statistics knows so little about 'Regression to the mean'. Or is it a case of 'ner ner I'm not listening' [/Fingers in ears mode off].
He's cornered.
Destroyer duz statistics Jewhoo said:
It would be a start if the stats quoted were correct, or at the very least, consistent
Consistency is no good if the statistics are consistently wrong. They need to be correct.
We do need data to guide us in our search for sensible policies, whatever the subject in question, but unless the data is reliable we are probably better off without it. That is why I base my views on my own experience and observations, and I'm sticking with that approach until something better comes along.
Best wishes all,
Dave.
TripleS said:
It would be a start if the stats quoted were correct, or at the very least, consistent
Consistency is no good if the statistics are consistently wrong. They need to be correct.
We do need data to guide us in our search for sensible policies, whatever the subject in question, but unless the data is reliable we are probably better off without it. That is why I base my views on my own experience and observations, and I'm sticking with that approach until something better comes along.
Best wishes all,
Dave.
Same here! I know a lot of drivers who drive very high mileages and none of us have experienced what Destroyer tells us will happen if we drive 'too fast' on roads like motorways.
What Destroyer doesn't seem to grasp is things like the dangers of overtaking with too low a speed differential and not being able to maintain separation from other vehicles all around you. This is far more crucial to safe driving on all but the busiest motorways when it's hard to avoid situations like these (but this is down to woeful underfunding of the transport infrastructure anyway).
What are Destroyer's measures doing to prevent people from falling asleep behind the wheel, driving too close to the vehicle in front or pedestrians trespassing on the motorway?
If you go to the DfT website, and look at the 112 pages (!)of the UCL study - A 3 year evaluation of safety cameras - LINK TO PAGE - it gives lots of data, and the formula used to validate the stats. This is the study that comes up with near identical conclusions to Linda Mountain's more recent work. I suspect there is a degree of collusion or plagiarism, especially since the potted conclusions are the same as Destroyer quoted yesterday. Or am i missing something here?
Well I just proved that Destroyer makes 'em up as he goes along so it doesn't really matter where he gets 'em from.
As for consistency, I was trying to point out that it would be easier to show them that they are wrong, but as the stats keep changing when challenged, they will never get it and keep on bringing out "new" stats.
As for consistency, I was trying to point out that it would be easier to show them that they are wrong, but as the stats keep changing when challenged, they will never get it and keep on bringing out "new" stats.
nubbin said:I disagree, that report has no useful data at all, only summaries.
If you go to the DfT website, and look at the 112 pages (!)of the UCL study - A 3 year evaluation of safety cameras - LINK TO PAGE - it gives lots of data, and the formula used to validate the stats.
nubbin said:
This is the study that comes up with near identical conclusions to Linda Mountain's more recent work. I suspect there is a degree of collusion or plagiarism, especially since the potted conclusions are the same as Destroyer quoted yesterday. Or am i missing something here?
UCL study said:This is the only mention of this effect in the report
Because of this requirement to identify the sites as suitable for this particular safety measure, and in particular that the collision record was not the sole criterion for selection, the established statistical phenomenon of regression to the mean (also known as bias by selection) will not apply in full measure.
So the report clearly does not fully account for RTM (IMO it didn’t account for it in the slightest), while Linda Mountain's claims to, yet they yield the same results? 
I have the press releases; these may clear up some confusion. Don’t ask me how I got them 
Here they are, in their unedited entirety (my longest post ever):
I believe in the 80/20 rule, though the 90/10 rule may be more applicable here – in reverse.
It is clear that Destroyer and his chums are utilising 90% of the total resource to solve 10% of the problem and has at best been lying about the effectiveness of his partnerships.
The summaries also fail to exclude benefits from other potential improvements at camera sites (new features other than the cameras themselves) e.g.
new speed camera plus new zebra crossing = reduction of KSI at camera site
Here they are, in their unedited entirety (my longest post ever):
said:
Press briefing on Speed Cameras:
Recent research carried out by Dr Linda Mountain and Dr Mark Hirst at Liverpool University and Prof. Mike Maher at Napier University has investigated the impact of various types of speed management scheme on speeds and accidents on 30mph roads. The study was based on data for some 149 speed management schemes on 30mph roads – 79 cameras, 39 engineering schemes with some form of speed hump and 31 other types of engineering scheme (chicanes, mini-roundabouts etc.)
The study found that schemes with vertical deflections were on average most effective: they resulted in a 44% fall in all accidents, a 35% fall in fatal and serious accidents (FSAs) and reduced mean speeds by 8mph. With cameras the effects were a fall of 22% in all accidents, a fall of 11% in FSAs, with an average fall of 4mph in mean speeds.
While our study suggests that humps give the greatest reductions in both speeds and accidents, cameras still give worthwhile safety improvements and can be used more widely than humps. Cameras are normally used on busier routes – particularly those used by buses and emergency vehicles – where speed humps would not be appropriate. In our study we did find that cameras were used on busier roads at locations which had more accidents in the first place. This means that although the percentage accident reductions look very different, in fact, in terms of the fall in the number of accidents both humps and cameras result in savings of about 1 accident per kilometre per year.
It would be wrong to suggest that these results 'prove that speed cameras do not save lives'. They indicate that, for the sample of sites examined, a statistically significant reduction in fatal and serious accidents could not be positively established. In other words a larger sample of sites may show a larger or smaller reduction in fatal and serious accidents but our best estimate from this data set is a fall of 11% not a zero change.
said:
What does our research show?
We studied 149 speed management schemes on 30mph roads where speeding problems are most acute (around 60% of drivers exceed the 30mph limit) – 79 cameras, 39 engineering schemes with some form of speed hump and 31 other types of engineering scheme (chicanes, mini-roundabouts etc).
In terms of percentage accident and speed reductions we found that schemes with vertical deflections were on average most effective: they resulted in a 44% fall in all accidents, a 35% fall in fatal and serious accidents (FSAs) and reduced mean speeds by 8mph. With cameras the effects were a fall of 22% in all accidents, a fall of 11% in FSAs, with an average fall of 4mph in mean speeds. (The effects of other engineering schemes are somewhere between the two: falls of 29% and 14% in all accidents and FSAs respectively and a reduction of 3mph in mean speed.)
Speed humps are thus more effective than cameras?
Yes it is fair to say that our study suggests that humps give the greatest reductions in both speeds and accidents but cameras still give worthwhile safety improvements and can be used more widely than humps.
Cameras are normally used on busier routes – particularly those used by buses and emergency vehicles – where speed humps simply wouldn’t be appropriate. For through traffic routes cameras would normally be considered a more appropriate choice.
In our study we did find that cameras were used on busier roads which had more accidents in the first place. This means that although the percentage accident reductions look very different, in fact, in terms of the fall in the number of accidents both humps and cameras result in savings of about 1 accident per kilometre per year.
Your estimates of the effects of cameras appear rather lower than some other studies. Why is this?
In our study we took account of both trends in accidents and regression-to-mean (RTM) effects. If these effects are not accounted for schemes such as cameras and humps will appear more effective than they really are. Although most studies take account of the fact that the general downward trend in accidents will account for some of any observed fall in accidents most do not properly allow for RTM effects.
RTM arises because sites chosen for cameras, humps and so on are normally selected because they have had a particularly large number of accidents over the last 3 years. Selecting sites in this way means that the sites chosen will tend to be those with more than expected accidents in the 3-year period used to select sites and as a result they will have fewer accidents in a subsequent period of time simply because the numbers recorded in the first period were abnormally high. If accidents are observed before and after the implementation of a road safety scheme part of any observed accident reduction will be due to this RTM effect. In this study we found that RTM accounted for falls in accidents of between 3 and 16% while trend accounted for falls of around 2 to 6%. Clearly if these effects aren’t accounted for the effectiveness of any road safety measure will be over-estimated.
So what should the government do now?
In my view the government should continue to try to encourage drivers to reduce their speeds. There is much evidence to suggest that speeding increases the risk of accidents and it is important to encourage drivers to drive more slowly. Cameras, speed humps, chicanes and so on can all help to achieve this. The goal really should be to ensure that speeding becomes as socially unacceptable as smoking and drink-driving.
But your work finds no significant reduction in the most serious accidents. Doesn’t this mean that cameras should be scrapped?
No. The fact that the reduction in FSAs wasn’t statistically significant doesn’t “prove that cameras don’t save lives” as has been quoted in the press. Our best estimate is that they save 11% of fatal and serious accidents over the section of road up to 1km either side of the camera. The lack of statistical significance simply means that with the data we had available we couldn’t positively establish that there is a reduction. The average reduction for all cameras could be bigger or smaller than the one we observed but a fall of 11% (not zero saving) is our best estimate.
But surely the government should be using more humps and fewer cameras – cameras are just used because they generate a profit.
Humps are simply not appropriate on most busy roads – for example they can cause problems for buses, ambulances and fire engines. Although speed cushions may be used to overcome this problem it’s not at all clear that widespread use of such devices would be any more popular than cameras.
The revenue from cameras isn’t a profit – it is used to fund the operation of the cameras and of course no one has to pay the fines generated by cameras they just have to drive within the speed limit.
I believe in the 80/20 rule, though the 90/10 rule may be more applicable here – in reverse.
It is clear that Destroyer and his chums are utilising 90% of the total resource to solve 10% of the problem and has at best been lying about the effectiveness of his partnerships.
The summaries also fail to exclude benefits from other potential improvements at camera sites (new features other than the cameras themselves) e.g.
new speed camera plus new zebra crossing = reduction of KSI at camera site
supraman2954 said:
nubbin said:
If you go to the DfT website, and look at the 112 pages (!)of the UCL study - A 3 year evaluation of safety cameras - LINK TO PAGE - it gives lots of data, and the formula used to validate the stats.
I disagree, that report has no useful data at all, only summaries.
Sorry, I should have made it clearer - go to the box that says "Downloadable files" and click on the link to the .pdf file. That's where all the data is.
Gassing Station | Speed, Plod & the Law | Top of Page | What's New | My Stuff