BBC article on cycling accidents

[QUOTE=simonw7;998153Thanks I had not seen that. a lot of the discussion on the thread has been around whether or not hypothesis implied by bbc is borne out by the data given the sample size. the truth would be better represented by studies of data over a greater time period, including this paper (and presumably data has continued to be collected since this paper).
[/QUOTE]

As I keep saying, the sample size is not small, it's large. It's in the millions. It's the number of male and female cyclists in London. What is small is the number of HGV deaths observed in this sample.

As I keep saying, the sample size is not small, it's large. It's in the millions. It's the number of male and female cyclists in London. What is small is the number of HGV deaths observed in this sample.

I do agree with you on this. my point was that the data used for this bbc story is (apparently unnecessarily) limited to 2009, rather than going back 10 years, 20 years, whatever. regardless of whether the number of hgv deaths or the number of bicycle journeys is considered to be the sample. The limiting of the data to such a small period appears purely to be for the reason that this limited dataset gives the impression of supporting the story which they decided they wanted to report,i.e. it is a form of cherrypicking.

I do agree with you on this. my point was that the data used for this bbc story is (apparently unnecessarily) limited to 2009, rather than going back 10 years, 20 years, whatever. regardless of whether the number of hgv deaths or the number of bicycle journeys is considered to be the sample. The limiting of the data to such a small period appears purely to be for the reason that this limited dataset gives the impression of supporting the story which they decided they wanted to report,i.e. it is a form of cherrypicking.

I suspect it's a bit of journalistic laziness as well. Buffalo Bill has published figures for previous years on Moving Target, and they have been republished here a few times.

The millions of journeys with relatively few deaths is strong evidence that cycling is rarely fatal. The smallish number of deaths means that the actual death rate is still fairly loosly determined, relative to it's size. That makes extimates of future death rates a bit uncertain, but the conclusion that cycling is rarely fatal still holds strongly.

Split the death stats into catagories and you get even smaller numbers, smaller rates, and greater relative uncertainty. When you compare the rates for different categories, the uncertainty in each rate can easily be bigger than the difference between one rate and another.

So for the question we are discussing, it is a small sample, and the conclusions not impressively strong, but plenty strong enough in my view to justify action, given the likely costs and benefits of acting or not acting.

Firstly I’d like to not take anything away from the sad loss of life, which should be the main focus of the article and the issues its trying to highlight.

I’ve just used the calculations I normally use to look at the statistical significance of response rates to junk mail campaigns on these small sample sizes.

If you assume a 50:50 gender split in cyclists on the road. You can still, even with this small sample of 9 riders say with 99% statistical significance that the [unfortunate] result is true, that women are more likely to be killed by HGVs than men.

NB:It does assume that this sample of 9 is either random, or all HGV road deaths in a given period.

Also a higher proportion of male cyclist (Which I’m sure is the case) would only increase the significance of the results.

I don’t really want more deaths to improve the statistical significance, 9 is already 9 to many…

As I keep saying, the sample size is not small, it's large. It's in the millions. It's the number of male and female cyclists in London. What is small is the number of HGV deaths observed in this sample.

But that's the data that's relevant to this specific question. There were only 9 deaths. You can't tell anything about whether women or men are more likely to die at the wheels of HGVs from the women and men that didn't die.

Whilst I'm quite enjoying the debate on statistics between people who don't quite know as much about sample size and clustering as they are pretending, here is another way to look at the data;

100% of people who died on the left hand side of an HGV were on the left had side of an HGV. Doing everything you sensibly can to avoid being on the left had side of lorry drastically reduces the chances of dying on the left hand side of a lorry. See? No standard deviations required.

is there any sort of decent training that can be offered to cyclists who spend the majority of their riding time in the streets?

is there any sort of decent training that can be offered to cyclists who spend the majority of their riding time in the streets?

http://www.lcc.org.uk/index.asp?PageID=175

Just tell them Oliver sent you ;)

@willski

Yeah, maybe we need a statistics thread, where we can get into an nicely ill-informed bayesian vs. frequentist argument.

But anyway, some people have died on the right hand side lorries and as far fewer people pass lorries on the right we have much less statistical evidence about how safe that is or isn't. So passing on the right instead of left may not be so great.

Not overtaking lorries at all doesn't work either, because they can come from behind and overtake you, and often there is little you can reasonably do about that.

Then there is the problem that if avoiding the left of lorries is the generally accepted wisdom it becomes more excusable for lorries to kill people on their left, and there will always be plenty of people on the left of lorries because the dynamics of the road often put them there and cycle lanes & ASLs encourage them to be there.

hey moth,

you are good at this.

you should post more often.

nice one.

They were random stops:

City of London [Police] spot checks on HGVs [were] carried out on 30
September 2008 as part of the Europe-wide Operation Mermaid2, which is
intended to step up levels of enforcement of road safety laws in
relation to lorries.
On this one day, 12 lorries were stopped randomly by City Police. Five
of those lorries were involved in the construction work for the 2012
Olympics. All of the twelve lorries were breaking the law in at least
one way

Repeat: a 100 per cent criminality rate among small random sample of
HGVs on the streets of central London. The offences range included
overweight loads (2 cases), mechanical breaches (5 cases), driver
hours breaches (5 cases), mobile phone use while driving (2 cases),
driving without insurance (2 cases) and no operator license (1 case).

just to add the that, the lorry that hit me had no operator license and 3 tacho offences.

still hasn't gone to court a year later. the defence keep stalling....

just to add the that, the lorry that hit me had no operator license and 3 tacho offences.

still hasn't gone to court a year later. the defence keep stalling....

It's time to acknowledge how completely nuts it is to have huge machinery lumbering round narrow streets and posing a disproportionate danger to vulnerable road users.

It is time to point out that the beeb article, complete with the drawings of the "kill a cyclist and get away with it if you say they were in these shaded areas", is misdirection.

People ought to be allowed to make mistakes and not die or be crippled, that's what the aim should be- a town where the most dangerous vehicles are perpetually aware of their lethal potentiality and the drivers behave accordingly.

Just thought I'd point out that there's a mailing list set up at seemesaveme.com for the sort of statistical debate people have been having:

http://www.seemesaveme.com/wiki/index.php/Mailing_Lists

Join the Enforcement/Legal/Data list.

Of course, if you'd rather discuss these issues here, I'm not one to stop you, but it would be good to focus it in the work of the Action Group.

It's time to acknowledge how completely nuts it is to have huge machinery lumbering round narrow streets and posing a disproportionate danger to vulnerable road users.

It is time to point out that the beeb article, complete with the drawings of the "kill a cyclist and get away with it if you say they were in these shaded areas", is misdirection.

People ought to be allowed to make mistakes and not die or be crippled, that's what the aim should be- a town where the most dangerous vehicles are perpetually aware of their lethal potentiality and the drivers behave accordingly.

+1

But that's the data that's relevant to this specific question. There were only 9 deaths. You can't tell anything about whether women or men are more likely to die at the wheels of HGVs from the women and men that didn't die.

You can't tell anything about whether women or men are more likely to die under HGVs unless you know roughly how many women and men didn't die.

You can't tell anything about whether women or men are more likely to die under HGVs unless you know roughly how many women and men didn't die.
Put another way you need to know how many cyclists out there are men and how many are women. No one has really good data on this, guesses range between 25%-35% women cycling in London, less in the rest of the country.

Different studies have shown women account for between 25% and 33% of cyclist fatalities in London - roughly in line with the numbers out there. However when looking at types of fatal crashes women seem to be involved in around a half of HGV related deaths and a much smaller proportion of non-HGV related deaths. Studies from Germany show a similar disproportionality as does anecdotal evidence from other places.

The way it adds up women are much less likely than men to be killed by a car and slightly more likely to be killed by HGVs. To be honest I have no idea why this is so. One suggestion is that women are smaller than men and so are more vulnerable to inattentive lorry drivers; that doesn't explain why they seem to be much less at risk from car collisions.

Meanwhile, is anyone up for a quick demo at City Hall tomorrow morning (wednesday 14th) before the Mayor's Question Time discusses him cutting all the funding for the only police unit in the country with the skills to assess safety procedures of HGV operating companies. ?

Different studies have shown women account for between 25% and 33% of cyclist fatalities in London - roughly in line with the numbers out there. However when looking at types of fatal crashes women seem to be involved in around a half of HGV related deaths and a much smaller proportion of non-HGV related deaths. Studies from Germany show a similar disproportionality as does anecdotal evidence from other places.

The way it adds up women are much less likely than men to be killed by a car and slightly more likely to be killed by HGVs. To be honest I have no idea why this is so. One suggestion is that women are smaller than men and so are more vulnerable to inattentive lorry drivers; that doesn't explain why they seem to be much less at risk from car collisions.

Yes this is the pattern that seems to be present in the Inner London data in that BMJ paper that Sharkstar provided the link to.

Much as I I am reluctant to publicly speculate in a relatively uninformed manner on such a sensitive topic, I can't help thinking that this pattern could show that men and women cyclists are to a large extent equally likely to be killed by HGVs because it appears that they may not be seen by the driver, hence no gender determination is made, which means that potential differences in driver behaviour towards women and men cyclists do not occur, hence equal probability of collision/fatality for men and women.

However, women cyclists are less likely to be killed by other vehicles [which do not share the blind spot issue of HGVs] purely because they can be seen to be women, which may in itself be sufficient to invoke some extra caution on the part of the driver. This low level of fatalities of women cyclists (when compared to men cyclists) caused by non-HGVs is very apparent in the Inner London data in the BMJ paper.

Obviously this assumes that cycling behaviour between genders is roughly equal which may not be the case.

Just thought I'd point out that there's a mailing list set up at seemesaveme.com for the sort of statistical debate people have been having:

http://www.seemesaveme.com/wiki/index.php/Mailing_Lists

Join the Enforcement/Legal/Data list.

Of course, if you'd rather discuss these issues here, I'm not one to stop you, but it would be good to focus it in the work of the Action Group.

Please do join the SeeMeSaveMe group. This is a great discussion with loads of useful info. We need to capture it and use it.

Well said on this point below, totally agree. Kulveer Ranger and his special cycles for the poor wee girls cos they arent very confident... Eildih was hardcore, but careful, not at all reckless, and in front on the right!! Repeat - you are not safe on the right, nor in front.:

"But, more than anything else, the focus has to shift to i) the sheer volume of HGVs on urban thoroughfares, and ii) the lawless and reckless manner in which they're operated by drivers and their companies. These endless variants of the same article - what are the cyclists doing wrong? - only serve to reinforce the way (enforcement of) the law views operating motor vehicles as an inherently blameless activity."

Meanwhile, is anyone up for a quick demo at City Hall tomorrow morning (wednesday 14th) before the Mayor's Question Time discusses him cutting all the funding for the only police unit in the country with the skills to assess safety procedures of HGV operating companies. ?

Yes, a demo. Except Im in northumberland and its a bit late in the day.
Here is a question that will be asked tmrw:

HGVs and safety
Question No: 2652 / 2009
Jenny Jones
Will the cuts to the Commercial Vehicle Enforcement Unit make London’s roads less safe for cyclists?

Just a thank you for bringing this kind of thing to people's attention.

Came up behind a lorry indicating right at a junction yesterday. No room to pass on the right, thought about passing on the left, remembered all the discussions bout it on here recently, stopped behind it. A second later and the dick goes and turns left.

Cheers all.

Shit Hannah, lucky/well thought through.

I avoid the fucking things like the plague, even buses don't bother me like HGV/Arctics do.

Except those bendy buses. They're a different story all together...