Talk:Bicycle helmet/Archive 4

Archive 1

Archive 2

Archive 3

Archive 4

Archive 5

Archive 6

Recent string of revisions

I see many revisions have been done recently. I think they happen too fast, without the discussion that the part shows that this article needs. If nothing else to avoid "edit-wars". For instance I disagree with the assertion that a reference to the WP article Victim blaming be irrelevant -Morten7an (talk) 01:42, 27 February 2013 (UTC)

Please explain how an WP article on "victim blaming" is relevant to this article on bicycle helmets, bearing in mind that the NPOV policy extends to what links to other WP articles appear in an article. AFAIK, no bicycle helmet has ever been observed to blame a victim (of what?). I am happy to discuss the other edits I have made. Please document your concerns. Many have just been tidying up of references. Much more tidying up is required, as is obvious from a glance at the references section. There are a lot of incomplete references, a lot of references that include large verbatim quotes, when the resources they refer to are already accessible online, and there are quite a few redundant references. Tim C (talk) 03:07, 27 February 2013 (UTC)

Quotations in references

Quite a few of the references for this article include quotations from the cited source, some of them very long. This makes the whole reference section very difficult to read. The WP guidelines state that quotes should be included in references only when it make the task of editorial review easier. In the case of references which are freely available online, there is no reason to copy text into the reference as well. If there is a passage in the reference which is both germane and sufficiently important and reliable, then it should be included in the article body text (with attribution, of course). Where references are not online or are behind a paywall, then retention of existing quotes in references seems acceptable. Views? Tim C (talk) 07:23, 27 February 2013 (UTC)

Go for it! Richard Keatinge (talk) 15:34, 19 March 2013 (UTC)

I have moved one quote from a reference into a separate Quotes section, with a forward pinter from the reference to the quote, and a back pointer from the quote to the reference. The forward pointer is not elegant but is the best I could come up with - other WP mark-p solutions using magc-words don't seem to work. I will fix the other quotes in reference in the same manner, as time permits, if there are no objections.Tim C (talk) 20:59, 19 March 2013 (UTC)

This would result in a less informative article, as most people do not read references. A quick summary or a quotation of a reference is an excellent way to briefly inform the reader what the reference is about and why it is relevant. To remove those references would lead to a less informative article.

If the quotation is too long, then it should be summarised or shortened, as long as it still represents the essence of the message. Claiming the quote is too long should not be used as a way to give lower prominence to references. The approach suggested is open to abuse, as it would enable people to remove quotations from references they do not like, to give them less prominence. Given that the discussion can be quite polarised, this would open the door to unnecessary edit wars.Harvey4931 (talk) 22:11, 5 April 2013 (UTC)

In general, article text should present a complete account without the need for footnotes or extensive quotations. But there are often good reasons for using clearly-marked quotations in the text, and there are also good reasons for leaving quotations in references. One is to save the reader checking through a long text, and the other is to keep a brief record in case of link rot. The relevant bit of text in a reference can usefully do both of those things. I'm not, so far, finding similar utility in transferring long quotations to footnotes and with apologies to Tim's hard work so far, I suspect that it may not be an idea worth pursuing. Richard Keatinge (talk) 10:26, 6 April 2013 (UTC)

This article should apply to helmets for mountain bike riders too

User:MTB UK reverted a large number of changes, with the complaint that the page ought to refer to helmets for MTB riders too. I agree, but the way to address this to to add such material, not to arbitrarily and seemingly randomly revert many edits made by other contributors. Tim C (talk) 10:29, 18 March 2013 (UTC)

Roady-bias seems to be an on-going issue with this article. It's not clear that all the material you added or referenced applies to every type of helmets/helmet usage, and this is something I wanted you should demonstrate when reinserting it. MTB UK (talk) 00:57, 19 March 2013 (UTC)

Yes, several of the recent studies and scientific papers done by University of NSW include injuries to MTB riders as well as road cyclists. I'll endeavour to make that clearer as time permits. However, a better strategy is to request such enhancements on this Talk page, rather than arbitrarily deleting a lot of stuff because it fails to mention MTB riders. Tim C (talk) 01:06, 19 March 2013 (UTC)

The Corner et al. report and the Phillips helmet

User:Richard Keatinge has made an edit to the section titled "Rotational injury" which states that one of the authors of the 1987 Corner et al. study (see http://www.infrastructure.gov.au/roads/safety/publications/1987/pdf/Mcycle_Helm_1.pdf ) was also the inventor of a new type of helmet, as interviewed by New Scientist magazine in 2000 (see http://www.newscientist.com/article/dn418-soft-hat.html ). This does not appear to be the case. The name of the inventor of the helmet is Kenneth Phillips, from Middlesex, UK, whose name does not appear anywhere in the Corner et al. report, which was done at the School of Civil Engineering at the Queensland Institute of Technology in Australia some 13 years earlier.

Closer inspection of the New Scientist article reveals that the Kenneth Phillips helmet is actually intended for motorcyclists, and possibly equestrian riders and police. The article mentions motorists but I think that is a typo, and motorcyclists was the word intended. However, cyclists, as in bicycle riders, are not mentioned anywhere, and from the description of the construction of the helmet, it is hard to see how it could be used by bicyclists except in very special competition settings in which ventilation was not required (eg sprint track events). Given that the Phillips helmet is not a bicycle helmet, and that the inventor of a new patented helmet type has a clear conflict of interest, I don't think that the mention of the Phillips helmet and the quite lengthy quote from the inventor of it has a place in this article about bicycle helmets, and I have removed it from the article. If anyone wishes to make a case for its re-instatement, please do so here. Tim C (talk) 20:48, 23 March 2013 (UTC)

Is there an active scientific debate on whether helmets are useful for general cycling?

In the first paragraph of the article, there is an unreferenced statement: "There is an active scientific debate, with no consensus, on whether helmets are useful for general cycling, and on whether any benefits are outweighed by their disadvantages." Is this statement actually true? Can anyone point to evidence of an active scientific debate about whether helmets are useful for general cycling? The dominant scientific view seems to be that they protect the head and brain, and there seems to have been very little dissent from this view in the scientific literature in the last decade. There has been some debate about helmet laws in the literature, but not about the usefulness of helmets themselves. Tim C (talk) 11:10, 30 March 2013 (UTC)

That really is a silly question. If you read the article and the 161 sources, or the four other bike helmet wikipedia entries, you'll see that the statement is evidently true. Or look at the external links and compare helmets.org with cyclehelmets.org. Do you feel they are in consensus? In the last decade, Dorothy Robinson, Ian Walker and Rune Elvik have published skeptical studies. Fyhri has published two about helmets and risk compensation.

A reference might be nice, just to make it even more obvious. This one maybe, Cycle helmets – an overview, with the quote "It is at first surprising to many people that the wearing of helmets by cyclists is a controversial subject." Or maybe this one Bradley Wiggins knows a lot about cycling. But he might be wrong about the safety benefits of wearing a helmet. The Daily Mail "The whole issue of cycle helmets is so strange that David Spiegelhalter, the very excellent Professor of the Public Understanding of Risk at Cambridge University, a man who knows more about accident stats than almost anyone else, half-jokingly refuses to talk about it. 'Cycle helmets' he once told me. 'I try not to go there.'"

These citations could be put in but I'm not sure it would help people understand an article that is already quite long and a little too detailed. The article should give the reader an understanding of the debate without going into every detail and argument. Erik Sandblom (talk) 21:10, 4 April 2013 (UTC)

My question is a serious one - where exactly is recent (i.e. active) the scientific debate about whether helmets are useful for general cycling, and on whether any benefits are outweighed by their disadvantages. The Elvik meta-analysis is about helmet effectiveness, and it confirms previous meta-analyses that helmets are effective in protecting against head, brain and facial injuries, but aren't effective in protecting against neck injuries. Elvik's paper doesn't address benefits versus disadvantages at all. So that paper doesn't seem to be part of a debate about helmets being useful for general cycling. The Fyhri et al. paper is about risk compensation - again, it doesn't address the issues of whether helmets are useful for general cycling, and on whether any benefits are outweighed by their disadvantages. The de Jong and Newbold mdels are about whether any benefits are outweighed by disadvantages of mandatory helmet laws, not of helmets per se. So that leaves Robinson's articles. Does one author make for an active scientific debate. But there may be more articles that are part of a debate - if so, which articles? Or should the statement be re-worded to indicate that the active scientific debate is about helmet laws (and/or promotion), but not about bicycle helmets themselves? Tim C (talk) 22:22, 4 April 2013 (UTC)

It's right there on the page: "According to the new studies, no overall effect of bicycle helmets could be found when injuries to head, face or neck are considered as a whole." (Elvik) The whole concept of risk compensation means that behaviour will compensate for helmet wearing, thus a zero-sum game.

The fact that much of the literature is about helmet laws is precisely because it's highly questionable whether any benefits are outweighed by their disadvantages. Disadvantages include personal taste like helmet hair or simply not wanting to wear one. Erik Sandblom (talk) 22:38, 4 April 2013 (UTC)

Here's another source, on the British Medical Association which has changed it's position on helmet laws at least twice since 1992. British Medical Association. Cyclehelmets.org And here's a general comment, "Several studies have shown that bicycle helmets have the potential of reducing injuries from accidents. Yet, no studies have found good evidence of an injury reducing effect in countries that have introduced bicycle helmet legislation." Bicycle helmets – A case of risk compensation?

Again, I'm not sure a citation war is the best way to help the general public understand the helmet debate. It's better to explain the different viewpoints, while leaving out the fact that every statement will have its detractors. Erik Sandblom (talk) 22:13, 4 April 2013 (UTC)

I don't think anyone has ever suggested that any bit of material on the head would have precisely zero effect on head injuries, but there are certainly plenty of people who don't wear helmets even under legal threat. The absolute risk of serious injury from riding your bike is very small after all, and of serious head injury still smaller. Robinson's work may be seen as a definitive answer to the issue of whether mass helmet use noticeably reduces the proportion of head injuries, but scientific debate seems to continue, publications do continue to come out, and indeed Tim's talk page says he's working on a further analysis. de Jong's and Newbold's analyses are of compulsory helmet use, but to the extent that perceived pressure to wear a helmet may also discourage cycling they may apply to helmet promotion also. Elvik has recently re-analysed the hospital-based studies in new and illuminating ways. And so on. I'd say that was an active scientific debate, personally. The statement as written seems quite good. Richard Keatinge (talk) 08:53, 5 April 2013 (UTC)

Tim C has commented that the dominant scientific view seems to be that helmets protect the head and brain; that the only dissenting articles in the scientific literature in the last decade seem to be by one author (Robinson); and has questioned whether one author constitutes an active scientific debate.

In a 2006 letter to the BMJ, under the Heading 'Arguments against helmet legislation are flawed', Hagel claimed that 'Robinson’s opposition to helmet laws is contrary to published evidence on the effectiveness of bicycle helmets . . .six studies have examined the relation between helmet laws and head injuries, and all found a reduction in head injuries after legislation was enacted'. In the same article, which started on the same page that Robinson's article finished on, Hagel also noted that the correlation coefficient corresponding to one of the graphs in Robinson's article suggested that 'much of the variation in the percentage of head injuries is explained by helmet use. In a 2009 Transport Research Laboratory report ('The Potential of Bicycle Helmets to Prevent Injury - A Review of the Evidence'), Hynd was dismissive of Robinson's 2006 paper, stating that the definition of head injuries was 'inadequate', 'very vague and not useful for this type of study'. I am not aware of any responses (in the scientific literature) refuting the criticisms expressed in either of these articles, it seems that there is no active SCIENTIFIC debate.

Tim C's initial post included the word 'scientific' at least 3 times (plus once in the heading):'Can anyone point to evidence of an active scientific debate about whether helmets are useful for general cycling? The dominant scientific view seems to be that they protect the head and brain, and there seems to have been very little dissent from this view in the scientific literature in the last decade'. In commenting 'That really is a silly question', then citing helmets.org and cyclehelmets.org as evidence of lack of consensus, Erik Sandblom seems to have not noticed the use of the word SCIENTIFIC in Tim's question 'Is there an active scientific debate on whether helmets are useful for general cycling?'. Neither helmets.org nor cyclehelmets.org are scientific publications.

Erik Sandblom went on to comment that 'Dorothy Robinson, Ian Walker and Rune Elvik have published skeptical studies'. Richard Keatinge has commented that 'Elvik has recently re-analysed the hospital-based studies in new and illuminating ways'.

In the 2011 paper, Elvik wrote: 'Do bicycle helmets reduce the risk of injury to the head, face or neck? With respect to head injury, the answer is clearly yes, and the re-analysis of the meta-analysis reported by Attewell et al.(2001) in this paper has not changed this answer.' The differences between the results of Attewell's original study and Elvik's 2011/2012 study are indeed tiny:

• Attewell's study found that unhelmeted cyclists were 2.4 times more likely than helmeted cyclists to sustain a brain injury, in Elvik's 2012 re-analysis, this number was 2.5
• Attewell's study found that unhelmeted cyclists were 2.5 times more likely than helmeted cyclists to sustain a head injury, in Elvik's 2012 re-analysis, this number was 2.4
• Attewell's study found that unhelmeted cyclists were 3.7 times more likely than helmeted cyclists to sustain a fatal injury, in Elvik'2 2012 re-analysis, this number was 4.3

In a later post, Erik Sandblom states that evidence of an active scientific debate is 'right there on the page: "According to the new studies, no overall effect of bicycle helmets could be found when injuries to head, face or neck are considered as a whole"'. I have previously noted (at http://en.wikipedia.org/wiki/Wikipedia:Reliable_sources/Noticeboard#Bicycle_Helmet_Research_Foundation ) that what is posted on the cyclehelmets.org with respect to the Elvik study is grossly misleading, it seems that Eric has not actually read the Elvik paper/s.

Erik Sandblom does not give a citation for the study by Ian Walker, he may be referring to a 2005 article listed on the same cyclehelmets.org page as the Elvik study. The Walker article listed on the page was not in a scientific publication, but a cycling magazine.

The arguments that helmets are not protective are restricted to sources such as cyclehelmets.org. Nick-D has commented that 'The front page of the website alone makes it pretty clear that this is an advocacy website: all the stories are about how bike helmets are a bad thing. As such, this doesn't seem to be a reliable source given that it's only presenting one side of the debate (a genuinely disinterested academic or research-type organisation wouldn't structure their website in such a way)' Stuart.Jamieson has commented that the BHRF, which operates the cyclehelmets.org website, presents itself 'as an expert organisation when that is by no means proven.'

I am not aware of any active SCIENTIFIC debate about whether helmets are useful for general cycling.Linda.m.ward (talk) 07:21, 7 April 2013 (UTC)

Linda Ward, Elvik writes "New studies, summarised by a random-effects model of analysis, indicate no net protective effect." Perhaps that could be used as a source for the contested statement?

Hagel et al's reply to Robinson 2006 is to simply ignore the whole point that motorists and pedestrians seem to gain the same benefits as cyclists, even though they were not wearing helmets.

The helmet-skeptic evidence is not restricted to cyclehelmets.org and Ian Walker published in Accident Analysis and Prevention. Also I wonder why cyclehelmets.org is not considered scientific, when they have an editorial board which approves the material and whose members have written in several peer-reviewed journals. Erik Sandblom (talk) 19:32, 9 April 2013 (UTC)

The Bicycle Helmets in Australia page has a quote from Ron Shepherd: "no-one disputes that helmets work. But why aren't motorists required to wear them? .... on average you'd have to be wearing your (bicycle) helmet since 509 BC for it to protect you."

Although Curnow argues that helmets prevent against wounds and skull fractures, but not brain damage from rotational injuries, the main scientific debate is surely about costs and benefits, at it's no different to the arguments for wearing helmets in cars. Racing drivers wear helmets, but most people think the the risk for ordinary driving are not high enough to justify the inconvenience. Racing cyclists and mountain-bikers wear helmets, but many people (including scientists who argue that cycling without a helmet increases life expectancy compared with not cycling) don't think the risks for ordinary cycling justify the inconvenience. That, surely is what is meant by "useful for general cycling, and on whether any benefits are outweighed by their disadvantages." Things are not usually considered useful if the disadvantages outweigh the benefits. Dorre (talk) 09:07, 7 April 2013 (UTC)

Dorre -- You're referenced comment about motorists wearing helmets is a Straw Man argument. This page is about cycling helmets only. However, if you want to make that point, why don't we make cyclists wear seat belts, only sell bikes with air bags, introduce stability control on bikes and only sell bikes with roll cages? Surely all these things, that are standard in cars, would make cycling safer? Right? Well... that's just another Straw Man.

The fact of the matter is the helmet (along with maybe good cycling infrastructure and the cyclist's brain protected by the helmet) is the only item comparable to motor vehicle safety for bikes.JakeOlivier (talk) 09:25, 7 April 2013 (UTC)

Jake, Dorre is simply using car helmets to illustrate the point that "Things are not usually considered useful if the disadvantages outweigh the benefits." It's not enough that there may be benefits. The benefits have to outweigh the disadvantages. Erik Sandblom (talk) 19:15, 9 April 2013 (UTC)

Erik -- How is that relevant here? The auto industry spends millions upon millions trying to make driving vehicles safer. Research has led to the many improvements I listed above. Each of them were weighed against advantages and disadvantages. In terms of safety equipment engineered to make cycling safer, there is only the helmet (clearly, infrastructure is important too, but it is not safety equipment). To compare to modes of transportation with vastly different engineering concerns is not supported by those types of arguments. JakeOlivier (talk) 07:28, 10 April 2013 (UTC)

Right, and the auto industry decided not to pursue helmets, perhaps because they are customer driven and thus sensitive to the tradeoff. Not just benefits, but disadvantages too. Car industry talks bike helmets, silent on car helmets

It's inappropriate for you to use wikipedia as a discussion forum to promote your views and ask spurious questions like "Is there a scientific helmet debate". I think you should consider starting a forum where you and like-minded people can talk about the details. Once you have a group-effort essay explaining how helmets are not controversial, you can link to it. Don't use wikipedia as a discussion forum. Erik Sandblom (talk) 08:43, 10 April 2013 (UTC)

Erik -- You're comment that I'm using WP, in your words, "to promote your views" is not founded in anything. This discussion is about whether it is appropriate for the WP article on bicycle helmets to say that there is an active debate about their efficacy (which I did not start, by the way). I only responded to a fallacious argument by pointing to the error in logic. How exactly is that promoting my view? I am an academic who has published several peer-reviewed papers in this area and, as an expert in this area, I am merely pointing to the evidence/facts.JakeOlivier (talk) 02:06, 11 April 2013 (UTC)

Jake. I think it's fair to say we are all pointing to facts. However, that wearing a helmet while cycling should not be compared to wearing a helmet for other activities is an opinion and not a fact.

Regarding promoting your views, I was primarily referring to "Dorre's misinterpretation of statistical results". That discussion does not belong on wikipedia, though I can say in your favour that it's not a monologue. Erik Sandblom (talk) 07:29, 11 April 2013 (UTC)

The 2nd sentence on the project page states that 'There is an active scientific debate, with no consensus, on whether helmets are useful for general cycling, and on whether any benefits are outweighed by their disadvantages'.

Nobody has provided any evidence of an active scientific debate on whether helmets are useful for general cycling.

There is no active scientific debate 'on whether ANY benefits' of HELMETS 'are outweighed by their disadvantages', there seems to be consensus in the scientific community that helmets are clearly efficacious. I am not aware of any active scientific discussion of any disadvantages of helmets.

The ('scientific') debate is not about whether helmets work, it is about whether helmet LAWS 'work'.Linda.m.ward (talk) 04:26, 11 April 2013 (UTC)

Thanks User:Linda.m.ward. I'll edit the sentence in line with your useful suggestion. Tim C (talk) 06:39, 11 April 2013 (UTC)

Linda Ward, what about when Elvik writes "New studies, summarised by a random-effects model of analysis, indicate no net protective effect" (2012) or when Walker writes "wearing a bicycle helmet led to traffic getting significantly closer when overtaking" (2007). How is that not "an active scientific debate on whether helmets are useful for general cycling"? Erik Sandblom (talk) 07:12, 11 April 2013 (UTC)

Colin at cycling Corner et al http://www.infrastructure.gov.au/roads/safety/publications/1987/pdf/Mcycle_Helm_1.pdf

Corner discusses the density of foam and how good or bad a specific density may be. Corner provides details of the very high rotational accelerations, 69k rad/sec/sec, page 121. UK research found value up to about 20k at impact speeds of 8.5m/sec. How safe are helmets is one debate.

Some research has indicated that helmets lead to higher accident rates, eg NZ 2012 http://www.cycle-helmets.com/nz-clarke-2012.pdf note 20% higher rate

and

Robinson DL; Head injuries and bicycle helmet laws; Accid Anal Prev, 28, 4: p 463-475, 1996 http://www.cycle-helmets.com/robinson-head-injuries.pdf Tables 2 and 5

and

Clarke CF, The Case against bicycle helmets and legislation, VeloCity Munich, 2007 http://www.ta.org.br/site/Banco/7manuais/colin_clarke_cycle_helmet.pdf details a number of reports indicating a higher rate. A list of advantages v disadvantages for helmets is provided, so a debate could consider each in turn.

The helmet debate tends to focus on reports like Attewell, it did not measure risk per hour of cycling, it users reports to compare outcomes from wearers and non-wearers, typically for cyclists who had a choice. It was mainly comparing two groups. These groups tend to have different approaches. This is probably a separate issue comparing injuries for different ages and types of cyclists, typically teenagers may not in inclined to wear helmets and quite often they are at fault when it comes to accidents, whereas adults may be slightly more inclined to wear helmets. The accident rate can vary by around a factor of 10 to 1 per km of travel for different types of cyclists.

Considering a serious head injury, say about once in 2000+ years of cycling – see Ron Shepherd’s article, average person may cycle for say 50 years. So helmets may have a chance of reducing a serious head injury for an individual, say once in 2000 years of cycling. But if their accident rate increases by 20%, they have a higher risk of being involved in an accident and sustaining injuries in general in their 50 years of cycling.

One debate is whether helmets improve safety. Also the debate is whether they improve health and safety and additionally if legislation is justified. If safety is measured by all injuries then helmets fail, if only head injuries are considered, that would still be debateable because wearers will incur more impacts. Comparing the 2 groups, wearers v non-wearers is not assessing safety per km of travel. The debate cannot be easily resolved because of the variables involved and difficulty in obtained precise data for helmets effects alone. When helmets laws come in the debate takes another stage because non-wearers ask for good proof that helmets are safe, want to see why a legal requirement has been imposed with possible excessive fines. Data shows cyclists are at a similar risk of serious head injury to that of pedestrians and the risk is low. Imposing the legal requirements sets the stage for a health and safety debate, with a wide range of issues that can be considered and with the end result that helmet laws are questioned. The evidence for helmet use is not conclusive. The health implications affect many more people than the remote risk of head injury. Discrimination in accident compensation runs hand in glove with helmet promotion and legislation. It seems the debate can only continue. ~~Colin at cycling~~ — Preceding unsigned comment added by Colin at cycling (talk • contribs) 21:32, 11 April 2013 (UTC)

Does anyone have any major objections to the form of words in the opening sentences to the article as they now stand? If not, then I think it is time to close and archive this debate. Tim C (talk) 21:59, 11 April 2013 (UTC)

Yes, I object and think it should be reverted to the way it was. I have twice pointed out that Elvik and Walker question the usefulness of helmets for general cycling. So far nobody has replied but you changed it anyway Tim, why? Elvik writes "New studies, summarised by a random-effects model of analysis, indicate no net protective effect" (2012) and Walker writes "wearing a bicycle helmet led to traffic getting significantly closer when overtaking" (2007). How is that not "an active scientific debate on whether helmets are useful for general cycling"? Erik Sandblom (talk) 05:13, 12 April 2013 (UTC)

It may not be a good idea to cite the Elvik meta-analysis at the very beginning of this article at this stage because it is quite likely to be retracted in the near future - see https://github.com/timchurches/meta-analyses/blob/master/benefits-of-reproducible-research/benefits-of-reproducible-research.md The Walker study is about motor vehicle driver behaviour, not about helmet effectiveness. Tim C (talk) 05:39, 12 April 2013 (UTC)

Don't you think it might be a good idea to wait for Elvik to retract the study before determining that there is no debate? The Walker study is about how cycle helmets influence motorist behaviour. Fyhri found that cyclists go slower when they take off their helmets. And do you have a reply to Colin? Erik Sandblom (talk) 06:14, 12 April 2013 (UTC)

Yes, that's what I said, the Walker study is about motor vehicle behaviour when they observe a helmeted versus an unhelmeted cyclist. OK, I'll edit the opening sentences again. No I am not going to reply to User:Colin at cycling who seems to be writing about studies from teh 1980s and 1990s - that's not current scientific debate. Tim C (talk) 06:31, 12 April 2013 (UTC)

Thank you, a reversion would be appreciated. Richard Keatinge (talk) 07:57, 12 April 2013 (UTC)

~~Colin at cycling~~

2012 Some research has indicated that helmets lead to higher accident rates, eg NZ 2012 http://www.cycle-helmets.com/nz-clarke-2012.pdf note 20% higher rate, note the "Cyclist’s injury risk per hour increased by 20–32%."

2007 Clarke CF, The Case against bicycle helmets and legislation, VeloCity Munich, 2007 http://www.ta.org.br/site/Banco/7manuais/colin_clarke_cycle_helmet.pdf details a number of reports indicating a higher rate. A list of advantages v disadvantages for helmets is provided, so a debate could consider each in turn.

2006 UK research found value up to about 20k at impact speeds of 8.5m/sec. How safe are helmets is one debate.

The debate is also about reports trying to mislead the public.

With these results I have modified the opening to reflect the above.

~~Colin at cycling~~ — Preceding unsigned comment added by Colin at cycling (talk • contribs) 08:33, 12 April 2013 (UTC)

~~Colin at cycling~~

The statement below may not be very reliable and changes to reflect actual reports could be worthwhile.

“The evidence on bicycle helmets is mixed - some studies indicating that helmets provide a clear benefit and others finding either no clear benefit or that they have failed to improve safety.[3][4][5][6]”

"some studies indicating that helmets provide a clear benefit" Bambach el al is provided. This report compares outcomes for helmeted as a group to non-helmeted as a group. It does not provide a safety level per hour of cycling for each group, eg non-wearers may cycle more hours per week than wearers on average, exposure data is not available. A noticeable feature in the details was for the ages, 55% of non-wearers were in the 0-19 age group compared with 19% for wearers. A higher proportion of non-wearers had BAC over 0.05. Other differences can also be seen from the report. The fatality data provided, 18 from 42, out of 106 recorded in the accident statistics leaves a serious question about the data considered. In that the report is comparing quite diffeeent groups who could have quite different head injury rates, means it cannot really be considered to show that helmets provide a clear benefit, the overall accident rate per hour cycled would have to be included. Nearly all reports finding helmets provide a benefit are based on comparing 2 groups,

In addition some reports, NZ 2012, Summary - Cyclist’s injury risk per hour increased by 20–32%”, The ECF ( European Cycling Federation) stated "the evidence from Australia and New Zealand suggests that the wearing of helmets might even make cycling more dangerous", indicating safety was actually reduced.

Slight changes to the sentance to reflect the above mentioned information. ~~Colin at cycling ~~ — Preceding unsigned comment added by Colin at cycling (talk • contribs) 19:29, 14 April 2013 (UTC)

Dorre's misinterpretation of statistical results

A recent edit by Dorre states

"For pedestrians (as a comparison group), there was sudden and drop in both head and arm injuries also coinciding exactly with the onset of the helmet law, which was less likely to have happened by chance (P=0.02) than the change in the head to limb injury rates for cyclists (P=0.03). In both cyclist and pedestrian groups there was a less pronounced statistically-significant downwards trend in the head-to-arm injury ratio, over and above the sudden drop in in the ratio observed in cyclists only, and the sudden drop in both head and arm injuries observed for pedestrians. [35][36]"

This is an incorrect interpretation of the results from a paper I co-authored. In our paper, we primarily considered a comparative interrupted time series model to assess changes in cycling hospitalisations around the helmet law in NSW Australia with head injuries as the primary outcome and arm injuries as the comparator. Additionally, we performed sensitivity analyses comparing cycling head/leg injuries, pedestrian head/arm injuries and pedestrian head/leg injuries.

Dorre's comments are incorrect given the parameterisation we used for our models. For those that are interested, the paper can be found here http://handle.unsw.edu.au/1959.4/50858. In our analyses, we used indicator variables for the type of injury (head was always 1 and other types 0) and the pre- and post-law periods (post-law was always 1 and pre-law 0). So, the p-value given by Dorre for changes in pedestrian head and arm injuries with the law of p=0.02 is for arm injuries only. Using the model estimates, the estimated change in pedestrian arm injuries with the law is exp(-0.250)-1=-0.22. So, the model estimates a significant drop in pedestrian arm injuries of 22%. The term INJURY*LAW is a comparison between changes in pedestrian head and arm injuries with the law. We can use that to estimate the change in pedestrian head injuries with the law as exp(-0.250+0.106)-1=-0.13. So, head injuries for pedestrians dropped an estimated 13% with the law. The p-value of 0.41 indicates pedestrian arm and head injuries do not significantly differ; however, it does not indicate the 13% in head injuries is significantly (or not significantly) different from 0% (or no effect). I could re-parameterise the model to the get that p-value if you like, but what's given there is clearly incorrect.

With regards to cycling injuries, the estimated drop in arm injuries is exp(-0.112)-1=-0.11 and for head injuries exp(-0.112-0.322)-1=-0.35. So, head injuries for cyclists dropped by 35% with the law while cycling arm injuries dropped at a similar rate to pedestrian head injuries at 11%. The results for the cycling head/leg comparison was similar, and we reported a 29% adjusted decline in cycling head injuries with the law. Our recent 2013 paper modified this analysis a bit by reporting the most parsimonious model and a model in which inter-month correlations were accounted for. Each of these gave similar estimates of the decline in head injuries, but with much more significant p-values (p<0.001 in each case).

If Dorre would like to make a direct comparison between head injuries for cyclists and pedestrians, that would need to be done as it's own analysis. However, we can get a pretty good idea of the result using the model estimates which would be a comparison of 35% vs 13% (not that unlike the original comparison with arm injuries).

The reason for using pedestrians as a control is normally to adjust for time series trends. Because you assumed the trend for child cyclists was exactly the same as that for adults 6 months earlier (you have never explained why!) it's difficult to know what to make of your statistical adjustments. However, doing the same adjustment for pedestrians can provide an insight in the methodology, in particular if the assumption of different trends before and after the law is valid, or if exaggerates any effect of the law. The fact that you obtained an even more significant reduction in pedestrian arm injuries implies that there may be a problem with the methodology.

Dorre made another edit claiming Robinson used "similar data" in her 1996 paper. Her analyses heavily relied on survey data taken one month every year for four years. This is a missing data problem with 44/48 months not observed (about 92% missing data). I'm confused how census data of all hospitalisations aggregated by month is similar here?JakeOlivier (talk) 10:11, 7 April 2013 (UTC)

The only statistic of real-life importance is the number of head injuries relative to the amount of cycling. Your analysis cannot tell whether the number of injuries per cyclist increased or decreased, because you ignore the survey data showing big reductions in cycling compared to the reductions in head injuries. As you know, children often have a higher rate of head injury for the same accident severity, and children represent a very high proportion of cyclist hospital admissions. Children are also quite impressionable and in other situations (running an obstacle course) took significantly more more risks when wearing a helmet and other safety gear. A similar change in behaviour for kids on bikes would lead to an increase in falling off when doing tricks or off-road biking. Such falls are likely to have a higher risk of arm injury relative to the risk of head injury, irrespective of whether or not helmets are effective. High impact collisions with motor vehicles have a much higher risk of severe head injury than off-road falls, even if the cyclist is wearing a helmet. Reported on-road cycling injuries in NSW (all severity, many not requiring hospital admission) declined from 1,860 in 1990, to 1,088 in 2010, but NSW hospital admissions for arm injuries skyrocketed from 660 in 1991 to 1620 in 2010. The changing pattern - reduced on-road injuries but skyrocketing arm injuries suggests a change in the type of cycling from on-road transport cycling to off-road sports cycling that carries a much higher risk of arm injury compared to the risk of head injury. This change appears to have started with the helmet law. The limited survey data support this idea - ABS survey on cycling to work or education as percentage of total trips (Aust-wide, because of high sampling variation) were: 1996 - 1.9%; 2000- 1.1%; 2003 - 1.3; 2006 - 1.6; 2009 - 1.5; 2012 - 1.6. The same for census data on cycling to work - 1.09% in 1986 in NSW, 0.93% in 2011.

A recent paper in Risk Analysis suggests that helmet laws increase accident rates per cyclist by discouraging the safest cyclists. So we might expect the helmet law to change the head:arm injury ratio by increasing arm injuries from off-road falls, rather than reducing the real risk of head injury. Dividing the number of head injuries to children by estimates of the amount of children's cycling supports this conclusion, leading to the considerable debate about helmet laws. Your analysis does not compare numbers of head injuries with the amount of cycling, so it doesn't tell people what they need to know in order to determine whether helmet laws are beneficial.

I agree with Richard Keatinge that this isn't the ideal place to discuss this, but editors need to understand that the only important statistic is the risk of head injury per cyclist. If head injuries fall by 29% and cycling falls by 40%, the overall risk of head injury has increased - the average cyclist would have been safer without the law. If, at the same time, pedestrian head injuries fell by 10-20%, the comparison is even more marked - pedestrians are enjoying safer conditions, compared with reduced safety for cyclists. Dorre (talk) 22:40, 9 April 2013 (UTC)

Dorre -- You're only "proof" cycling has declined comes from surveys that were never meant to be analysed that way. As I've mentioned many times here, this is a "garbage in, garbage out" type of analysis. Head injuries are clearly the primary outcome here. But, the question in any analysis involving an interrupted time series, is what to compare it to? Cycling rates were never estimated in NSW around the time of the helmet law -- no matter what DL Robinson has stated. And if you are going to insist they be used, then you need to come to terms with 44/48 months of missing data. Although Walker's counts declined, helmet wearing proportions did not vary much post-law indicating that decline was not related to helmet wearing. That all adds up to a not very convincing argument. After fitting our models, the residuals (or what's left) correlated highly between head and arm injuries. This wasn't as evident with other models.

The rewording of the comment "and even stronger statistical evidence (P<0.02) for an equally sudden drop in the number of pedestrians (a comparison group) with arm injuries at exactly the same time" is clearly wrong. This statement ignores the fact that the most parsimonious and GEE models give similar point estimates but p-values less than 0.001. It also ignores the fact comparing p-values in that way (i.e., 0.02 < 0.03, so "stronger evidence") misses the point that the point estimate for the decline in cycling head injury is 35% and for pedestrian arm injury 22%. Further, p-values from interaction terms are known to have less power than main effects.

You keep mentioning the census data, but fail to mention these "trends" aren't distinguishable from horizontal lines. These are really, really small proportions. You keep ignoring my points, so I'll try again. Suppose I randomly generate 10,000 points on the unit square. I'll make 109 red and 9891 blue. I'll do that again with 93 red and 9907 blue. Can you honestly tell me you would notice a difference?JakeOlivier (talk) 07:46, 10 April 2013 (UTC)

Like so many recent edits, this discussion is interesting to me though I guess not to most readers of an encyclopaedia. It would be illuminating if it took place in a less opaque fashion (i.e. with all the raw data public and with open access to every detail of the analyses, with public comments and, we might hope, a consensus statement at the end). In the meantime it is far too detailed for an encyclopaedia. I would be fascinated to see such a debate and there may be ways of organizing one on the Internet. Jake, Dorre, and others involved, would you be interested in participating? Somewhere more appropriate than Wikipedia? Richard Keatinge (talk) 11:32, 7 April 2013 (UTC)

In her 1996 paper, Robinson only looked at NSW hospitalisations for injuries in child cyclists, not adults, and aggregated the injury data by year, not by month. With respect to the 2011 Walter et al. paper, there has been open discourse about it in the scientific literature - a critique of the paper by Rissel, and a response by its authors, both published in AAP. In the response, we mention that the data have been made publicly available. Worth a read. Tim C (talk) 20:25, 7 April 2013 (UTC)

Can you confirm that your model assumes the trends change with the law? In other words, you assume at the trend for children from July 1991 to December 1992 is identical to the trend for adults from January 91 to June 92, and the trend for adults from July 89 to December 90 is the same as that for children from January 90 to June 91. This seems a very strange assumption. Normally a trend would be fitted to the entire data series, to account for common temporal effects, e.g. admission policies and (for injuries caused by motor vehicles) common road safety factors. The normal way of treating your data would be to include monthly counts for children and adults in the same model, fitting seasonal factors to monthly counts of each age group, common trends, plus 0,1 factor for the law in a single model.

It's good to hear that you have obtained permission to make the data publicly available. Can you let us know how to find it? The only thing I could see were the aggregated counts for adults and children of head and arm injuries that had already been adjusted separately for (by implication potentially different) seasonal effects, to which, as far as I understand it, the model fitted by your group assumed the same trend for children as adults, except delayed by 6 months, with trends that were magically assumed to change with the introduction of the helmet law. There appeared to be no comparable data for pedestrians or for leg injuries. I assume that's not what you mean by making the data publicly available, so perhaps you could let us know how interested people can obtain the real dataset of monthly counts for children and adults by motor vehicle involvement (and ideally injury severity)? Dorre (talk) 21:30, 7 April 2013 (UTC)

Dorre -- I'm confused by your statement

"Can you confirm that your model assumes the trends change with the law? In other words, you assume at the trend for children from July 1991 to December 1992 is identical to the trend for adults from January 91 to June 92, and the trend for adults from July 89 to December 90 is the same as that for children from January 90 to June 91. This seems a very strange assumption.".

We originally performed analyses for kids and adults together because of "contamination" of helmet wearing due to different law dates (there's a clear increase in kids helmet wearing after the adult law but before their's. This is a conservative approach. This statement ignores the fact that we did analyse kids and adults separately in our 2013 paper (http://handle.unsw.edu.au/1959.4/52350). Please read it. We found the helmet law effect for kids and adults individually was similar to what we originally reported using their combined data. JakeOlivier (talk) 02:32, 11 April 2013 (UTC)

Researchers need to contact the NSW Ministry of Health to obtain NSW hospitalisation data, as has always been the case. Tim C (talk) 21:42, 7 April 2013 (UTC)

That wasn't quite what I had in mind by "publicly available", in fact a better description would be "impracticable to get the data and impossible to be sure, having got it, that we have exactly the same data for exactly the same codes that you used". I had in mind good scientific practice, making the original counts used for your analysis available on line. A meaningful and ethical scientific process absolutely requires those original counts; what you describe as "misinterpretation" of data is an attempt to guess at what the original counts showed, and without the original data the discussion cannot progress. Would you be prepared to make them available? I don't mean here, as I say this is at best tangential to Wikipedia, but Web space elsewhere could be made available for the data and indeed for a more appropriate off-Wikipedia discussion if you need it. Richard Keatinge (talk) 08:41, 8 April 2013 (UTC)

We are very sorry that we have failed to comply with what you had in mind. However, good, and ethical, scientific practice dictates that researchers comply with the restrictions put on the use and distribution of the data provided to them by health authorities. As was made clear in our response paper, we are not at liberty to provide copies of the data we used to other researchers, beyond the counts that have already been released. However, that does not mean that it is, as you hyperbolically put it, "impracticable to get the data and impossible to be sure, having got it, that we have exactly the same data for exactly the same codes that you used". To get the data, a researcher needs to write to the NSW Ministry of Health, as I indicated above. Writing a letter or an email is not impractical for most researchers, as far as I am aware. The specifications of the data we used are provided in our papers, but if other researchers need further details, then they need only contact the corresponding author by email, using their real name and their institutional email address, of course, as is the usual practice. Tim C (talk) 10:01, 8 April 2013 (UTC)

This is now way off topic for Wikipedia, but I note that the data are not, in fact, publicly available in any usual sense of the phrase. It may yet be possible for me to get them (and it will take a lot more than a simple written request); I'll let you know if that turns out to be possible. Richard Keatinge (talk) 10:43, 8 April 2013 (UTC)

The injury count data for cyclists, to which the models were fitted, as well as all the SAS program code which provides the exact specifications of those statistical models, are all freely available at http://www.unsworks.unsw.edu.au/primo_library/libweb/action/dlDisplay.do?vid=UNSWORKS&docId=unsworks_9752 That is all that we have been given permission to make available by the provider of the data. If you want more than that, as I have said, you will need to contact the NSW Ministry of Health. You seem to be more familiar with the requirements of that organisation than me, judging from your assertion that more than a written request will be required. There is nothing more to be said here in response to your inquiries about access to these hospitalisation data - please direct further inquiries about data access to the NSW Ministry of Health, and inquiries about our papers to the corresponding author. Thank you. Tim C (talk) 11:30, 8 April 2013 (UTC)

I agree that this discussion is going nowhere useful and is best terminated. I merely point out that the injury counts needed to make optimal comments are not at the link you supply; I take your point that access to the injury counts therefore requires fresh (and onerous) application to the NSW Ministry of Health. If I get a swift and satisfactory answer I'll leave a brief comment here. Richard Keatinge (talk) 12:43, 8 April 2013 (UTC)

Hey TimC, JakeOlivier et al. Why don't you team up with helmets.org and publish your thoughts there? I agree with Richard that Wikipedia is not the place for this very detailed discussion of bike helmets and BHRF. The wikipedia entry is already too complicated for a general audience, and I'm sure BHSI would appreciate your expertise. Wiki's role is to summarize sources such as BHSI and BHRF (and obviously the peer-reviewed publications), not to disect every detail. Erik Sandblom (talk) 19:45, 9 April 2013 (UTC)

Looking at Dorre's most recent change to the interpretation of the model by Walter et al., it seems that Dorre has confused the interpretation of a log-linear model with the interpretation of a linear model. The main effect terms in a log-linear model do not 'mean the same thing' as the main effect terms in a linear model. The first-order interaction terms in a log-linear model correspond to the main effects in a linear model. The main effect terms in a log-linear model refer to the marginal totals, the 22% drop associated with the LAW variable does NOT represent 'a sudden and sustained drop' in pedestrian arm injuries at the time of the helmet law'. As previously noted by JakeOlivier, the INJURY*LAW term compares (pedestrian) head to arm injuries before/after the law, for pedestrians the p-value was 0.41. The statement that the model showed 'even stronger statistical evidence (P<0.02) for an equally sudden drop in the number of pedestrians (a comparison group) with arm injuries at exactly the same time (but no change in the head:arm injury ratio in pedestrians)' is egregiously incorrect, and should be removed as a matter of priority.Linda.m.ward (talk) 22:56, 10 April 2013 (UTC)

Quoting Jake (above): "So, the p-value given by Dorre for changes in pedestrian head and arm injuries with the law of p=0.02 is for arm injuries only." The fact that the INJURY*LAW term is not significant is irrelevant - it just shows that there was a very similar drop in pedestrian head + arm injuries, as shown in Fig 3B. I originally thought the P-value for LAW applied to both head + arm injuries (Table 2, which reports it as an effect of LAW, really needs explanatory footnote), but was happy to revise the text in the light of the information provided by Jake. Dorre (talk) 23:44, 10 April 2013 (UTC)

Dorre - I don't understand why a footnote is needed when our model is explained in length in our paper (page 2066-67) and in the subsequent response (and both are freely available). As mentioned above, we've also made available the data used and SAS code to run our models. Simply put, you can decompose the model used into four parts in terms of pre/post-law and head/arm that are solely dependent on time. They are:

Arm, pre-law: b0 + b1(TIME)

Head, pre-law: (b0 + b2) + (b1 + b4)(TIME)

Arm, post-law: (b0 + b3) + (b1 + b5)(TIME)

Head, post-law: (b0 + b2 + b3 + b6) + (b1 + b4 + b5 + b7)(TIME)

So, the estimated head injury rate is (b0 + b2) right before the law and (b0 + b2 + b3 + b6) and, thus the change with the law is (b3 + b6). Similarly, the change in arm injuries with the law is (b0 + b3 - b0 = b3). Therefore, INJURY*LAW term (given by b6), is the change in head injury with the law while "adjusting" for the change in arm injury. The actual proportional change is then exp(b6)-1 since we used a log link function. It's just simple algebra. JakeOlivier (talk) 03:03, 11 April 2013 (UTC)

Most reviewers of journal papers expect tables to make sense independently of the text. If I sent round a paper like that to my work colleagues, it would not pass the internal review and I would not be allowed to submit it for publication. Dorre (talk) 03:14, 11 April 2013 (UTC)

Dorre -- What is the relevance of your clearly personal comment? Well... I did send this paper, and others, to my academic colleagues (experts in injury, statistics, engineering, psychology, public health, epidemiology...) who were all impressed. So were the reviews we got back from the journal. So what? Perhaps you should spend some time actually reading our papers and not openly criticizing our work when you clearly don't understand something. JakeOlivier (talk) 03:47, 11 April 2013 (UTC)

I agree with Jake - we have gone out of our way here to explain the statistical model in our paper to another WP editor who added commentary about our paper to the article which just wasn't correct. We have pointed to follow-up papers which further explain the analysis, and to freely-available copies of the statistical program code which was used to fit the models. What more could we have reasonably done? Despite this, the response from User:Dorre is a snide ad hominem remark about the quality of our work. My feeling is that the edits by user:Dorre regarding the study in question should be reverted, and this discussed closed. Does anyone other than user:Dorre disagree with this course of action? Tim C (talk) 04:19, 11 April 2013 (UTC)

Jake asked why I thought a footnote was needed. I answered that the convention was that people should be able to understand the information presented in the tables without having to go through the text for details of the parameterisation. Reporting a P-value for the effect of LAW that is in fact just the P-value for the effect of LAW on arm injuries is a tad confusing. Jake clearly disagreed with my comment, so I simply explained that other people have different conventions. I had already corrected the text, but Jake still thought it necessary to go into great detail about the models, and continues to claim that the data are freely available, even though we were discussing pedestrians, and I haven't yet been able to see any data whatsoever for pedestrians.

While Tim and Jake don't think a significant reduction in pedestrian arm injuries coinciding exactly with the law is of interest, I hope other editors agree that this provides important insight into the methods used in this paper. This information is relevant and should be retained, or reference to the entire paper deleted. Dorre (talk) 05:05, 11 April 2013 (UTC)

Dorre, I think the information is important but I'm not sure it belongs on wikipedia. Criticism or supplemental information not included in the study belongs in a journal, on cyclehelmets.org or maybe on a blog. Wikipedia is not a scientific forum in this sense, it is a tool for the general public to understand the scientific debates. It's the big picture and not the details that are important for wikipedia. Erik Sandblom (talk) 07:43, 11 April 2013 (UTC)

I agree, Erik. It would be better if complicated details could be omitted, but the second best option is to report the effect on both cyclist and pedestrian injuries so that people at least balanced evidence. Isn't that why other aspects such as risk compensation and the research showing that helmet laws are likely to discourage the safest cyclists (potentially affecting the type of collisions and injuries)? Dorre (talk) 08:18, 11 April 2013 (UTC)

I would probably choose to omit the entire thing but I doubt if we will reach consensus on doing so. If we are to keep it in, we do clearly need to mention the pedestrian control-group analysis as well. Omitting it would give a severely biased picture. Richard Keatinge (talk) 09:03, 11 April 2013 (UTC)

Large scale destructive edits by User:Chris_Capoccia

User:Chris_Capoccia has made large scale destructive edits to many references in this article, removing URLs to abstracts and/or full-copies of papers, and in some cases removing all details of the reference except for a PubMed ID or a DOI. These destructive edits then triggered the attention of an automated citation bot, which then made a whole raft of further changes, mostly inappropriate (due to the preceding destructive edits). I have reverted all these changes (one had to be reverted manually), and placed a warning template on the user talk page of User_talk:Chris_Capoccia. If the user makes further desctructive edits, then s/he will need to be reported to WP authorities. Tim C (talk) 22:57, 17 April 2013 (UTC)

The addition to "Cycling risk and head injury" section by User:Colin at cycling

User:Colin at cycling added the following text and references to the section "Cycling risk and head injury":

GB accident data shows the injury rate per billion km for cyclists has increased in recent years and indications are that the helmet wearing rates may also have increased.

If the juxtaposition of these facts:

1. cyclist injury rates per kilometre cycled in the UK going up slightly 2005-2011
2. helmet wearing rates in adult road cyclists in the UK increased from 30.7% in 2006 to 34.3% in 2008, up from 16% in 1994.

is intended to imply some sort of causal association (or a lack of causal association i..e lack of effectiveness of helmets), then the edit violates NPOV.

If the juxtaposition of the two facts is just co-incidence, then the edit is off-topic for this article (although arguably the helmet wearing rate information could be used in a different section). Finally, just providing a URL does not constitute an adequate reference. The authors, title, journal or publisher of the work or information cited must be given. It is unreasonable to expect other editors to fill in such details on your behalf subsequently.

For all these reasons, I have deleted this material. Tim C (talk) 01:29, 20 April 2013 (UTC)

Correct use of named references.

I have just spent about 30 minutes fixing errors in the references caused by incorrect use of named references. Please note that the syntax for referring to an existing named reference is:

<ref name=Whatever2011 />

and is not

<ref name=Whatever2011></ref>

The latter form replaces the details in the named reference with nothing when the page is rendered. The bast way to insert named references is to click on the Named References clipboard icon in the toolbar above the editing area. Tracking down and fixing this misuse of named references is very time-consuming. Please edit with care. Tim C (talk) 22:27, 21 April 2013 (UTC)

Conflicts of interest

I, like most editors of this page (presumably including someone who calls himself Colin at cycling) have a major conflict of interest - I cycle for transport. Repealing helmet laws won't prevent anyone who wants to from wearing a helmet, but it will encourage people who currently consider cycling too dangerous to see it in a different light. An increase in cycling will increase my personal safety. Currently, if I wanted to commit suicide, all I'd have to do is continue cycling round a roundabout when a truck is approaching.

It's funny that a helmet can change driver attitude, but it does. A friend of mine was cycling in town and came over to say hello to a retired couple he knew from a dance club. Apart from the helmet and prescription glasses that go dark in the sun, he was wearing normal clothes. But the attire was enough to make the couple think they might get mugged, until they recognized him. Even if only 1% of drivers think in this way, cycling changes from a pleasant, safe, enjoyable experience to something quite terrifying. I learned to cope - asserting my right to be on the road, being prepared to stop on roundabouts to give way to the entering drivers who would normally have given way to me. But the whole string of near misses (compared to not a single one before the helmet law was introduced) was almost enough to make me give up cycling.

Jake and Tim will tell you that cycling is booming, because you can buy an alloy-frame mountain bike with suspension and shimano gears for under $100. Great for off-road stuff, but not much use for transport without a rack, mudguards and lights. These would cost very little extra if included in the specs to the manufacturer, but double or triple the price when they have to be bought separately. Bikes are sold this way because the intended market isn't transport cycling. I see a few people struggling on these bikes with backpacks, but the discomfort and inconvenience must put a lot of people off. Extremely cheap bike imports have led to booming sales, but very transport cyclists. Backpacks also raise the centre of gravity, leading to increased risk of falls and arm injuries. People who see me cycling ask how on earth I can get up the hills. The answer is simple - once out of the police zone I remove the helmet as soon as I start feeling hot, to avoid the danger of sweat getting into my eyes.

So yes, if conflicts of interest need to be declared, then the fact that I cycle for transport ranks much higher than any other possible conflict of interest, as I assume it would for Colin at cycling. The second major conflict of interest would be that, as a citizen of the world, I'd like to encourage other people to use bicycles for transport for health benefits and environmental reasons. Dorre (talk) 01:14, 25 April 2013 (UTC)

Thank you for those anecdotes and thoughts, but please note that wikipedia Talk pages like this are not a general discussion forum - see Wikipedia:Talk_page_guidelines - the first paragraph states: "The purpose of a Wikipedia talk page (accessible via the talk or discussion tab) is to provide space for editors to discuss changes to its associated article or project page. Article talk pages should not be used by editors as platforms for their personal views on a subject." The reasonable request was for editors of this (and related) WP articles who are also authors of papers or other material referenced in these WP articles to declare such a conflict of interest if it is not obvious from their WP user name, particularly when editing content relating to papers they may have authored. Are we to assume from your response above that User:Dorre has no such conflict of interest? Tim C (talk) 01:32, 25 April 2013 (UTC)

User:Tim.churches, if I interpret User:Dorre correctly they were being subtle and you should maybe pause and read between the lines before diving in with legalese - it does you no service. Just a thought. Kiwikiped (talk) 05:15, 26 April 2013 (UTC)

Toddler saved from fall by helmet

I've put in a brief notice of this incident; it's a single anecdote but in the context of helmets having consequences not intended by their designers it may be appropriate for inclusion. What do others think? Richard Keatinge (talk) 08:28, 25 April 2013 (UTC)

It made me think of the other extreme: "It is unlikely I'll walk one day, said the 65-year-old Mr Kerec, who takes half a cup of drugs a day. He was nearly garrotted by the strap from his own helmet after he believes it became tangled in the other bike and yanked his head back. The nasty gash across his neck is a chilling reminder of his brutal accident..." http://www.canberratimes.com.au/act-news/insurance-for-cyclists-would-limit-pain-and-hurt-20130413-2ht1w.html#ixzz2RW6vcMEx

Perhaps other editors would like to comment on whether either or both should be included. Dorre (talk) 22:31, 25 April 2013 (UTC)

I think that none of these anecdotes should be included, and that the text should just note that there is a small risk of accidental strangulation risk in children wearing helmets when not riding, and that some helmet standards have been modified to take this into account, retaining the technical references and maybe one or two of the strangulation report references as examples, and deleting the rest. Otherwise, we'll need to include a section with lots of "The helmet saved my life" anecdotes in order to maintain a NPOV. My feeling is that anecdotes have no place at all in an encyclopaedic article, not even in footnotes. Tim C (talk) 22:41, 25 April 2013 (UTC)

OK. I'd like to suggest that this article should take brief notice of un-designed consequences both good and bad, and that a sentence, two at most, on what those consequences are should then be supported by full references including quotations. Or in a footnote, but personally I think the present system works well. Tim makes a good point in that we should also take notice of the many anecdotes of "a helmet saved my life". Since the existence of many of these is not controversial, perhaps we should just have a sentence describing them. If we want a reference we could use http://cyclehelmets.org/1209.html, though Tim and Linda might reasonably have further suggestions. Richard Keatinge (talk) 08:03, 26 April 2013 (UTC)

I think that including freak accidents of this nature, which are so unlikely to occour to the vast majority of helmet users, may come across as an attempt to fill the article with as much trivial anti-helmet bias as is physically possible. There's already a sentence about cases of young children playing suffering death/brain damage as a result of accidental hanging - which has 17 different citations(!), wont that surfice? Obscurasky (talk) 10:05, 26 April 2013 (UTC)

Opps - may have misread the above thread. I was responding to the '65-year-old Mr Kerec' story. Obscurasky (talk) 10:50, 26 April 2013 (UTC)

Reporting anecdotal evidence

In this specific case, I suggest again that the unintended consequences of helmets are well and appropriately described by perhaps one sentence each, supported by all the relevant references. Causality is fairly obvious in these cases and doesn't need any further remark. Separately, Google Alerts brought me over 300 anecdotes of "a helmet saved my life" before I stopped counting. A sentence mentioning these anecdotes would be an entirely reasonable addition, and since the fact of their existence is uncontroversial it doesn't necessarily need a reference at all though one or more would be fine. I agree with Dorre that causality is in grave doubt in these cases, and we need to use a form of words that does not endorse causality in Wikipedia's voice, but we don't, I suggest, need to indulge in any further argumentation on the subject here. Richard Keatinge (talk) 10:08, 27 April 2013 (UTC)

"My Helmet saved my life" stories would more logically fit into a section on public attitudes to helmets, in conjunction with other relevant information on how the number of such stories compares with mortality rates before helmets were common. This attitude was demonstrated by the claims made comparing the injuries of Chris Boardman and Fabio Casarteli, and that a helmet saved Boardman's life. Boardman, however, cycled without a helmet in his video arguing for better facilities to improve the health and wealth of the nation.

These are worthwhile topics of sufficient interest to be covered, but they really don't belong in a section on anecdotal evidence. More importantly, I don't think it is fitting to describe the numerous cases of hanging by helmet straps as "anecdotal evidence". Dorre (talk) 22:14, 27 April 2013 (UTC)

Case Control studies

Undid post by Tim.churches Reverted to previous script, important to provide information for readers to able to see concerns have been published, without having to search for information.

Case control studies only provide a comparison, not a measure of safety per hour cycling, they are only a guide. Riders wearing helmets may spend less time on average cycling, exposure data is required before making strong claims.

~~Colin at cycling~~

No, it is not acceptable for User:Colin at cycling on the one hand to delete mention of and reference to an online critique by Jake Olivier of a study by Colin Clarke, but on the other hand for User:Colin at cycling to insist that online criticism by Colin Clarke of another study be included in the text of the article. User:Linda.m.ward has already indicated that she agrees with me that such edits are clearly non-NPOV. Furthermore, User:Colin at cycling has failed to confirm that he has no conflict of interest with respect to these edits, despite repeated requests. Once again, I ask: do other editors regard such edits by User:Colin at cycling as acceptable with respect to the NPOV principle? Tim C (talk) 21:37, 28 April 2013 (UTC)

---

The following information helps readers to appreciate that the Persaud et al report has added complications, replies published by the CMAJ and is therefore suitable for Wiki.

"A 2012 study by Persaud et al. using coronial records of 129 deaths of cyclists in Canada between 2005 and 2010 found that unhelmeted cyclists in fatal crashes were more likely than helmeted cyclists to have sustained a head injury (adjusted odds ratio [OR] 3.1, 95% confidence interval [CI] 1.3-7.3). [84]The published replies detail that 23% of the bicyclists had been under the influence of alcohol or drugs at the time of death, while the status of 30% was unknown. Of the bicyclists, 16% were carrying unsafe loads at the time of death. Some details are mentioned from other reports, comparing wearers to non-wearers, e.g. A 1993 report from Ontario stated "In teenagers, drinking alcohol (OR: 2.8) and smoking (OR: 4.4) were strongly associated with helmet non-use. In the adult group, female gender (OR: 1.26), higher income (OR: 1.43), higher education (OR: 1.68), non-smoking status (OR: 2.0) and abstinence from alcohol (1.27) were associated with helmet use."

The online critique by Jake Olivier, tended to make statements not supported by evidence in the NZ report.

I consider trying to include one and not the other to be taking a NPOV and fair reporting. I do not see any conflict of interest in taking this approach.

~~Colin at cycling~~ — Preceding unsigned comment added by Colin at cycling (talk • contribs) 10:38, 29 April 2013 (UTC)

That a study has been published does not mean that we have to include it at all and we certainly don't need details that get in the way of a good encyclopaedic article. To my mind this implies that we should remove all the details mentioned by both sides, which strike me as over-burdensome in any account for the general reader. As I've suggested above, we should leave references attached to (a few) sentences that outline the main themes of the debate. Richard Keatinge (talk) 16:54, 29 April 2013 (UTC)

Ding-dong over describing authors

What are you folk up to?

User:Linda.m.ward and User:Colin at cycling are inserting and deleting respectively a qualification "a retired cycling coach and anti-helmet campaigner" on an author's name.

Let's pick at random an author from the other POV camp, say User:Tim.churches as we've a nice list of his output on his user page and he writes on The Conversation. It is clear from his output that he is dedicated to seeking errors, often insignificant, in the output of the other POV camp and publishing it in anything from peer-reviewed papers to self-published articles and web forums. Should we therefore tag all his references with "a dedicated pro-helmet campaigner and seeker of insignificant errors"? Of course not!

Behave folks, this article (and associated ones) has already damaged the reputation of Wikipedia enough. Kiwikiped (talk) 21:06, 26 April 2013 (UTC)

Thank you, Kiwikiped for that appraisal of my work. Numbers in tables that literally don't add up, time-series graphs with data points out by 18 months, statistical calculations performed wrongly with a spreadsheet giving incorrect results in a key helmet meta-analysis - these are indeed all mere nitpicking, and my insistence that quantitative research be arithmetically and mathematically correct clearly makes me a "a dedicated pro-helmet campaigner and seeker of insignificant errors", as you put it. I'm pleased to see that the research oeuvre of Kiwikiped suffers from no such defects. On a more serious note, WP has strict policies on ad hominem comments. Tim C (talk) 07:45, 27 April 2013 (UTC)

Tim you had me there for a while, and then I read your last sentence :-) I'm pleased to see you agree such qualification are inappropriate. Kiwikiped (talk) 04:52, 28 April 2013 (UTC)

talk, I was referring to your ad hominem comments about me and my work and speculation and assumptions about my "POV". Please desist. However, I also do not think that authors should be described in the article text, beyond listing their organisational affiliations where relevant. In article on bicycle helmets, affiliations to organisations which focus on bicycle helmets are relevant, in my opinion. Tim C (talk) 07:08, 28 April 2013 (UTC)

Sigh. Tim, it was an illustrative example wherein I constructed an inappropriate comment about a gander based on an actual inappropriate comment about a goose (see this Wikipedia page), along with a clear bolded statement that such comments are inappropriate. Kiwikiped (talk) 20:03, 28 April 2013 (UTC)

I strongly endorse this warning and add, as I have before, that this article now needs significant abbreviation (and better writing) and not more text. Richard Keatinge (talk) 21:25, 26 April 2013 (UTC)

I agree that personal descriptions of authors in the text are not generally required, although authorial affiliations with organisations with an exclusive or significant focus on bicycle helmets do need to be made clear to readers. Tim C (talk) 07:45, 27 April 2013 (UTC)

An appropriate reference is normally considered sufficient. Richard Keatinge (talk) 10:08, 27 April 2013 (UTC)

[Advance apologies on length, I should make it shorter but both skill and time elude me, the research on the WWF was time-consuming enough.]

Oh dear, Tim after your previous response I thought this was sorted, but now we have "not generally required" followed by a complete backward step with an insistence on "authorial affiliations". I'll even disagree a little with Richard Keatinge on this one, though he is obviously a Wiki-expert, and say that some of the references added to this article are an outright abuse of "authorial affiliations" to add NPOV editorial.

There are now 14 references to footnote [a] in the article! To claim that is NPOV is absurd and is a serious disservice to Wikipedia and its readers. I expect a lawyer might suggest it was downright defamatory. What are you folks playing at?!

Now I know I'm wandering into a minefield surrounded by quicksand, and seriously considered just blocking Wikipedia at the routers instead, but let me explain to see if we can rescue this mess.

As the two POVs fighting over this article hold very strong views I'll use as an example a completely different, often controversial, area: conservation. I'll also stick with Australian examples.

I've no idea how many POV camps there are on conservation; but I know some accuse others of putting animals before the need of people for jobs and livelihood, while some accuse others of destroying the planet for our children for short term profit, etc. The warring POVs in this case end up in Court, on picket lines, and who knows what else.

The Worldwide Fund for Nature (WWF) (note: I have no affliation with them or their opponents) is an international conservation organisation with an Australian branch. Given they are a conservation organisation they are bound to have opponents. They are evangelists for conservation, which could be phrased as "they have made their minds up" - a phrase thrown out during the debates on this article.

Professor Arthur Georges, Professor in Applied Ecology and Dean of Applied Science at the University of Canberra (note: I have no connection with Prof Georges and as far as I know he has no opinion either way on the subject of this article) is clearly an academic of some standing with many peer-reviewed publications to his name.

He is also a WWF Governor.

Should any reference to his work in Wikipedia have an attached footnote: "Arthur Georges is a WWF Governor, a pro-conservation organisation"? Well maybe yes if you are a business who wants to destroy turtle habitat (just a made up example, many of his recent papers are to do with turtles), but certainly not if you are being NPOV. That Prof Georges is a WWF Governor is irrelevant when considering his peer-reviewed research. It is not a negative, indeed it is more likely a positive recognition by his peers of his standing - random folk without standing among their peers are not usually asked onto editorial or governing boards!

Only if Prof Georges publishes a paper as "Prof Georges, WWF" is his membership of the WWF relevant; then he is speaking as a representative of the WWF.

Back to the WWF. They produce reports, let's pick one at random Kimberley Region: Searching for Wallabies. It is by the WWF, there is no author named at the end. If this is referenced in Wikipedia should the reference be: "the March 2012 WWF report 'Kimberly Region: Searching for Wallabies'"; or should it be: "the anonymously authored March 2012 WWF report 'Kimberly Region: Searching for Wallabies'"? It is frankly astounding that I should even need to ask the question, the answer of course is the first.

So let's return to the article in question and its two POV camps. Keeping to Australia, as above, I think it is fair to say that referenced authors Robinson and Curnow are in one camp and referenced authors Church and Olivier are in the other (and one of each pair is from University and the other not - balance is good).

It turns out that Robinson and Curnow are both on the editorial board of the international organisation the BHRF - the parallel with Prof Georges and the WWF is obvious (though I'm sure the WWF is of higher standing in its field than the BHRF is in its that is not relevant here). If a paper by either is written as "Robinson/Curnow, BHRF" then the affiliation is relevant, and obvious in the reference. If they, for example, write as "Robinson, University of New England" then their association with the BHRF is irrelevant just as Prof Georges' association with the WWF is under the same circumstances. Note that for some mentioning the BHRF/WWF might be seen as a positive, as others a negative, but that is irrelevant because the mention itself is irrelevant. Such a reference would only be inserted in a Wikipedia article if the editor was not being NPOV, however it might well backfire as they might add it as a positive attribution while the reader might see it as as negative or vice-versa - another reason, though none is necessary, not to do it!

Now Church and Olivier are not members of an international organisation, one does not to my knowledge exist for their POV - if one did they may well be invited by their peers to be on its editorial board. However being of like-mind they participate together in The Conversation as itinerant evangelists of the pro-helmet-law POV. (Note: The Conversation used only as an example as it is mentioned elsewhere, they may well evangelise their POV in other forums, both real and electronic, as well.) Should all references to their peer-reviewed work carry the footnote "A pro-helmet-law advocate vocal on The Conversation"? Of course not! That would be as none NPOV as referencing the WWF or BHRF.

To try to argue that one (an affiliation with an evangelising organisation) should be mentioned while the other (itinerant evangelism) not based on Wikipedia rules would be an attempt to twist the letter of the rules to usurp the intention of the rules and an unworthy activity for any editor. Would anyone seriously try to exclude or brand the WWF on the same argument?

Indeed to reference the WWF, BHRF or The Conversation-hosted evangelism would be to reverse the causal relationship. Prof Georges' affiliation with the WWF is undoubtedly not the source of his professional output, it is professional output which will have seen him appointed by his peers as a WWF Governor. Likewise Olivier's research does not derive from his musings on The Conversation but vice-versa, and to suggest otherwise would be to devalue his work.

And what of the output of the BHRF itself? Well, just as with the WWF output referenced above, that is what it is. It is not "anonymously authored", it is by the BHRF. (Whether it is appropriate to reference WWF or BHRF output in a particular context is an orthogonal question entirely, and one I won't enter in to here.)

So folks, of both POV camps, please stop being juvenile. You are meant to be collaborating on producing an NPOV article which presents both POVs. Remove those childish footnotes, the "anonymously authored" jibes, and whatever other nonsense you've both added - frankly the article is by now so unreadable I've no idea what else is in there.

This article is an embarrassment to Wikipedia. Kiwikiped (talk) 04:52, 28 April 2013 (UTC)

Kiwikiped, it is unfortunate that you find this article to be, as you put it, "...an embarrassment to Wikipedia". Fortunately Wikipedia empowers you to improve it. BTW, "anonymously-authored" is not a jibe. The referenced articles have no authors listed. Richard Keatinge has told us that the BHRF Editorial Board takes collective responsibility for unattributed web pages on the BHRF web site. OK, but in that case it is important that whenever an a paper by a member of the BHRF Editorial Board is mentioned or referenced in the article, that readers are informed of what other references in the article that author may also have written. As an alternative to annotating mentions of papers written by BHRF Patrons and/or Board Members, it might be better to annotate every reference to anonymous BHRF web pages and graphs with a list of the BHRF Patrons and Editorial Board? Tim C (talk) 07:00, 28 April 2013 (UTC)

Sigh... Tim, sorry that my direct analogy helped not one wit. Changing where you place the inappropriate sauce does not make it appropriate. Remove the sauce! Kiwikiped (talk) 20:14, 28 April 2013 (UTC)

Bicycle Helmet Research Foundation (BHRF) Patrons and Editorial Board members either take responsibility for otherwise anonymously-authored content on their web site, or they don't. The BHRF web site states that they do. However, I am happy to remove the footnotes about BHRF Patrons and Board members and substitute an appropriate authorship attribution to otherwise anonymously-authored BHRF material cited in this and related articles. Tim C (talk) 21:22, 28 April 2013 (UTC)

However, before I make those changes to address Kiwikiped's complaint about identification of the affiliation of cited authors with the Bicycle Helmet Research Foundation in footnotes, could Kiwikiped confirm a lack of any conflict of interest with respect to the Bicycle Helmet Research Foundation please. Tim C (talk) 21:27, 28 April 2013 (UTC)

Sigh... Kiwikiped (talk) 21:45, 28 April 2013 (UTC)

Exactly the same thought occurred to me when I looked that the case-control studies section. The idea was to have one or two examples, then use the meta-analyses to give an overall summary of the results. We should be writing for the average reader who doesn't know what an odds ratio is, let alone an adjusted odds ratio. Adding details of individual studies (e.g. Persaud) leads to other editors adding details of published criticisms, perhaps even references to cyclists who died when wearing helmets - http://members.shaw.ca/jtubman/deadhelmet.html

For the same reason, I wouldn't necessarily want to go into anecdotal 'my helmet saved my life' stories (at least on in the same section as the accidental hangings), unless you also want to discuss the sheer number of such stories compared to the number of cycling fatalities before helmet laws. There's no way of telling what would have happened without the helmet - even what proportion of cyclists would have actually hit their heads. However, there's no doubt about the cause of death in the case of accidental hangings. The references indicate the scale of the problem, without adding additional words to the text. Dorre (talk) 22:42, 26 April 2013 (UTC)

The point is that if anecdotal material about helmet disbenefits is to be included, then anecdotal material about helmets benefits ("it saved my life") also needs to be included, in at least equal quantity. For this reason, my feeling is that all such anecdotal material and individual case reports should be removed entirely from the article. Tim C (talk) 07:45, 27 April 2013 (UTC)

One way in which this and related articles could be improved would be to discuss each study or paper in the text just once. Some papers by some authors are currently mentioned and cited in the text literally dozens of times (scroll down to the references and look for papers with long series of superscripted letters in front of them - that needs to be corrected because it represents undue emphasis, which is something WP guidelines warn against. Tim C (talk) 07:45, 27 April 2013 (UTC)

We have been heading somewhat in that direction and the result is an article rapidly degenerating into unreadability. So far as I can make out, we are all here in good faith to write a good encyclopaedic article, but instead of telling a story (with whatever references are needed, whether many for one sentence or one to make a repeated point) we are adding more and more details. Wikipedia is not an appropriate place for ding-dong argument and NPOV is not measured by the number of references.

Apart from addressing the general problem of increasing unreadability, I'd like to recommend to all parties the essay Wikipedia:Writing for the opponent. Richard Keatinge (talk) 10:08, 27 April 2013 (UTC)

Qualification "a retired cycling coach and anti-helmet campaigner" - It appears to be a non NPOV and the article and others by the author states his qualification, and do not include "anti-helmet campaigner". Wiki is about providing a quality text on a topic, not allowing selection and tagging of individuals with make shift descriptions.

~~Colin at cycling~~

Update:

Tim suggested above that I should maybe go an clean up some of the mess. I've made a start but I doubt it is complete. Coincidentally I found an author for an "anon" paper as I went through (declaration: I've not read the paper and have no opinion on it, I just saw the anon, followed the link and found the author). I think "anon" is now eradicated, which can't be bad.

If anyone, from either POV, would prefer to point out stuff I've missed (rather than clean up whatever you've both inserted yourselves) then I'll agree to go in again and do some more cleanup.

However these fixes are only a small step to bringing this article up to encyclopaedia standard.

Now please all behave and try to get this article to something readable and worthy! Kiwikiped (talk) 02:09, 30 April 2013 (UTC)

Stated, noted or argued?

When cleaning up (see Ding-Dong section) I spotted a lot of "X stated", "Y noted" etc. when maybe it is "X argued" - or is that just the academic speaking? Now one might state that grass is green, but argue licensing hours are too restrictive; but maybe one wouldn't argue grass is green or state licensing hours are too restrictive. Maybe if there was more argue and less state/noted this article would appear less adversarial?

I must admit I changed one or two as I happened across them. But then I reached the footnote which reports that errors have been found only to then say they await confirmation; so they've actually been possibly found surely? However I realise such a change might not go down well on this particular page, so at that point I refrained from making an more stated/noted -> argued kind of changes and am placing this Talk item instead. (If anybody wishes to undo such a word change I won't object, they just caught my eye as I looked for the stuff covered under Ding-Dong.)

But ultimately there are bigger issues with this article than this one (it is is an issue at all)... Kiwikiped (talk) 02:28, 30 April 2013 (UTC)

Reversion of the description of the Rodgers 1988 study by User:Harvey4931

User:Harvey4931 reverted the description of the Rodgers 1988 study from:

A study by Rodgers in 1988 examined the relationship between estimates of annual bicycle-related injuries and fatalities, and estimates of annual bicycle helmet sales in the US between 1973 and 1985. It showed no association between estimated helmet sales and estimated numbers of serious injuries, and a small but statistically significant association between annual helmet sales and annual cyclist fatalities.

to:

To examine claims that growth in the use of hard shell cycle helmets had been successful in reducing cycle-related injuries and death, Rodgers studied over 8 million cases of injury and death to cyclists in the USA over 15 years. He concluded: "There is no evidence that hard shell helmets have reduced the head injury and fatality rates. The most surprising finding is that the bicycle-related fatality rate is positively and significantly correlated with increased helmet use".

I was intrigued by the earlier description of the Rodgers study, so obtained a copy of the original paper, and found that Rodgers did not "study over 8 million cases of injury and death to cyclists in the USA over 15 years" - 8 million was the cumulative estimate of cycling injuries over 15 years, derived from sampled reports from a network of hospital emergency rooms and physicians' offices. Therefore data was not collected on 8 million injuries, 8 million was just the estimated number of cycling injuries (all injuries, and very rough estimates as well). Helmet wearing was estimated from sales of one model of Bell cycling helmet. Copyright restrictions prevent me from posting the PDF of the paper online, but I did re-construct the entire data used in the Rodgers analysis, using the figures provided in teh Rodgers paper and an earlier paper to which it referred, and then checked with Rodgers by email that these data were an accurate re-construction of what he analysed. Here are the entire data as analysed by Rodgers:

year, total.injuries, head.injuries, lower.trunk.injuries, deaths, 
          bicycles.7yr.cpsc, bicycles.10yr.cpsc, cpsc.compliant.7yr, cpsc.compliant.10yr, 
                    helmets, neiss, riders.millions, ped.deaths, vehicles.millions 

1973,453000,59700,14700,1100,  64,76.4,0,0,                 0,0,67.8,10200,129.8 
1974,520000,70000,15200,1000,  71.3,85,0,0,                 0,0,77,8500,134.9 
1975,520300,68900,15400,1003,  70.8,86.2,0,0,            4000,0,77.5,8400,137.9 
1976,503300,63600,18700,900,   70.4,87.5,0,0,           16000,0,78,8600,143.5 
1977,546500,71500,17300,900,   70.8,89.6,0.13,0.11,     64000,0,85.8,9100,148.8 
1978,491600,62600,16400,892,   71,91.3,0.26,0.21,      300000,0,79.5,9600,153.6 
1979,557700,71500,21900,932,   72.5,94.1,0.39,0.31,    600000,1,73.1,9800,159.6 
1980,503400,63200,17700,965,   72.1,94.8,0.49,0.39,   1000000,1,73.5,9700,161.6 
1981,550000,68200,23200,936,   71.4,95.2,0.58,0.47,   1500000,1,73.9,9400,164.1 
1982,573700,68100,25400,883,   68.7,93.3,0.65,0.52,   2000000,1,74.3,8400,165.2 
1983,571200,69000,24500,839,   68.2,93.5,0.72,0.58,   2600000,1,78.1,8200,169.4 
1984,556700,67300,25700,849,   69,94.7,0.78,0.65,     3200000,1,79.5,8500,171.8 
1985,581800,68800,28000,890,   71.1,97.2,0.83,0.7,    4000000,1,80,8500,177.1 
1986,564400,69100,25100,941,   74.2,100.6,0.87,0.75,         ,1,,8900,181.4 
1987,561800,69400,24700,949,   77.3,104.1,0.9,0.8,           ,1,,7500,183.9

Yes, that's the entire extent of the data which Rodgers analysed. Note in particular how obviously approximate the helmet sales estimates (as a proxy for helmet wearing - he assumed every helmet sold was worn) from 1973 to 1985 are:

0
0
4000
16000
64000
300000
600000
1000000
1500000
2000000
2600000
3200000
4000000

Note that the (obviously very approximate) estimate of helmet use by cyclists in 1985 is 4 million. But the estimated number of cyclists in that year was 80 million i.e an estimated 5% helmet wearing rate. It is drawing a very long bow indeed to base any conclusions about the impact of helmets on injury and death in cyclists based on these data. For these reasons, I believe that the description of the Rodgers study which User:Harvey4931 replaced with an earlier version is more accurate, and I am restoring it to the article. Tim C (talk) 12:28, 28 April 2013 (UTC)

---

This information seems to be orginal research in drawing its own conclusions. "A study by Rodgers in 1988 examined the relationship between estimates of annual bicycle-related injuries and fatalities, and estimates of annual bicycle helmet sales in the US between 1973 and 1985. It showed no association between estimated helmet sales and estimated numbers of serious injuries, and a small but statistically significant association between annual helmet sales and annual cyclist fatalities." It appears to not make clear the findings of Rodgers with regards to helmets.

Rodgers says There is no evidence that hard shell helmets have reduced the head injury and fatality rates. The most surprising finding is that the bicycle-related fatality rate is positively and significantly correlated with increased helmet use This is a clear statement from the report.

The data appears to suggest that pedestrians had a 26% reduction in fatalities, 10200 to 7500 and cyclists had a 14% reduction, 1100 to 949. Rodgers does not detail the number of pedestrian deaths per year.

I would suggest the following may be suitable.

Rodgers reported on injuries and deaths to cyclists in the USA over a 15 years. He considered bicycle use levels, bicycle standards, helmet use estimates and road safety trends using pedestrian data as a guide. For helmets he stated: "There is no evidence that hard shell helmets have reduced the head injury and fatality rates. The most surprising finding is that the bicycle-related fatality rate is positively and significantly correlated with increased helmet use".

Colin at cycling (talk) 18:00, 28 April 2013 (UTC)Colin at cyclingColin at cycling (talk) 18:00, 28 April 2013 (UTC)

This is not original research, it is just presentation of the actual data used by Rodgers in his research, abstracted directly from his paper, from an earlier paper by Petty which Rodgers is re-analysing in his paper, and from a US National Safety Council report which Rodgers cites. However, from these data one can assess what importance should be attached to the Rodgers study. After reading the Rodgers paper and assembling the data he used, it became clear that quoting one sentence from Rodgers' conclusion overstated what could validly be drawn from the Rodgers analysis of the data shown above. The following text would be an acceptable compromise:

"Rodgers analysed annual (1973 to 1987) national time-series estimates of cycling injuries, annual counts of cycling and pedestrian road accident fatalities, estimates of bicycle use and bicycle standards compliance, and estimates of helmet use based on sales of one particular brand of (now obsolete) hard-shell helmet (helmet wearing rates used by Rodgers rose from 0% in 1973 to 5% in 1985). After fitting a linear regression model to these data he concluded that: "There is no evidence that hard shell helmets have reduced the head injury and fatality rates. The most surprising finding is that the bicycle-related fatality rate is positively and significantly correlated with increased helmet use".

Tim C (talk) 21:02, 28 April 2013 (UTC)

I suggest that it is not useful to describe the Rodgers study in the main body of the text at all. As Tim points out, it is not a strong piece of evidence; rather than go into nitpicking descriptions in an encyclopaedic article, I would attach it as a reference, possibly with a brief quotation as part of the reference, to some appropriate sentence that outlines the fact that many time trend studies do not find a useful effect of helmets. I hope this helps. (Beyond Wikipedia I would find Tim's reconstructed data and comments on this paper useful as a basis of a BHRF commentary, and I intend to put such a proposal to the editorial board. I don't know if Tim would want to work up his commentary for this purpose, or if he'd want any specific credit on cyclehelmets.org, but in any case the usual offer of anonymity would apply.) Richard Keatinge (talk) 10:11, 29 April 2013 (UTC)

All content contributed to wikipedia, including content on Talk pages, is licensed under a CreativeCommons Attribution-ShareAlike-Remix license, is it not? Therefore the Bicycle Helmet Research Foundation can copy the Rodgers data I have provided above and place it on or in a web page on its own web site. However, the terms of the wikipedia license MUST be complied with to the letter if it does so, and that includes licensing the contents of that or any BHRF web page or article which uses or includes wikipedia content under the same CreativeCommons license, and giving full and clear attribution to wikipedia, including listing the exact version of the WP article from which the material was copied. I'm not making this up - check the wikipedia pages on copyright and licensing. However, personal attribution is not required, nor would it be requested in this case. Tim C (talk) 05:20, 30 April 2013 (UTC)

I'm quite familiar with the copyright situation, thanks. As a supererogatory point of courtesy I'd prefer to have the explicit and active support of any contributor; possibly the best approach would be for someone else to repeat your data extraction with any attribution being entirely at your discretion. Anyway, kudos for doing the work. Back to talking about Wikipedia; I repeat my recommendation that the Rodgers study, like many others, should be used simply as one of many references for an outline of the situation. Richard Keatinge (talk) 07:23, 30 April 2013 (UTC)

Thank you for your kind offer, but no, I do not wish to contribute to the Bicycle Helmets Research Foundation web site, and no, you do not have my consent to use my name nor any data, words or other content which I have contributed to wikipedia on or in any Bicycle Helmet Research Foundation web page or article. Tim C (talk) 08:42, 30 April 2013 (UTC)

Thanks for the detailed reply Tim. The revision suggested as a compromise seems a more neutral reflection. In general, I tend to favor quoting directly from a study as it is less prone to interpretations, and thus disagreements about interpretations. I will update the description accordingly.Harvey4931 (talk) 03:39, 2 May 2013 (UTC)