This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.
After some inactivity there appears to be some fairly active editing on this page again. Can we have a discussion here about what this article should achieve and how it can meet those ends without us constantly undoing each other's changes?
Obviously this is an issue with definite opinion on both sides but the evidence for or against helmet laws is far from overwhelming and in keeping with the guidelines of wikipedia we need to make sure we include unbiased evidence without editorialising.
Thanks to Harvey4931 for including the detailed section about the history of the laws introduction. I was too heavy-handed in my original edits to that section and I think we're closer to reaching a more acceptable compromise now. I still feel the section about the study being ignored should be removed because it is only one study of many that was looked at and singling it out gives one side undue weighting. I'm open to what others thing about this however.
Every study cited on this page has been criticised by someone and then counter criticised. I thought about including rejoinders that have been published to some of the studies here but decided it was more in keeping with wikipedia's guidelines to only include the initial study and not the rejoinders and counter rejoinders. This page is too large as it is without every reference having two or more counter references cited. For this reason I think I'm right in deleting the "this study has been ctiticised" sections. All studies have been criticised.
I also think we have to include every study and reference point rather than deleting ones included by others. If something appears to be an issue it would be great if it could be discussed more and a compromise reached.
I think it is OK to describe and reference responses, rejoinders and other critiques of peer-reviewed published studies, provided those responses and rejoinders and critiques were themselves also peer-reviewed and published in a scientific journal - that is, that they are part of the recognised scientific discourse. To exclude such responses and critiques would mean that, for example, all of the published articles by Bill Curnow would need to be chopped out, and that would be a loss, because they are definitely part of the scientific discourse on the subject. However, critiques and criticisms of research papers which are self-published on the web sites of single-issue organisations should definitely not be cited on WP, and especially when such critiques are anonymously authored. Tim C (talk) 06:31, 26 February 2013 (UTC)
So why was important information deleted, such as the change in pedestrian injuries head injuries in Victoria? It is now generally recognized that other road safety initiatives were introduced at the same time as Victoria's helmet law and that one way to try and separate out these effects is to compare identical statistics for cyclists and pedestrians or other road users. I have also added other important information, e.g. that the estimates of cycle use in the Olivier paper reflect the change in inner Sydney, which differed considerably from other data (e.g. census) for NSW as a whole. Olivier compared injury data for the whole of NSW with cycle use for inner Sydney. It would have been more relevant to compare injury data and cycle use for the same area. Finally, I have added back the explanation about the "estimated 44% changes in adult cycle use in Victoria", which was for the entire period from Dec/Jan 87/88 to 1991, so smooths out the increasing pre-law trends and the effect of the helmet law. The 1990 survey counted adults, which were published; the series of numbers counted from 87/88 to 1990 to 1991 (the latter two surveys using the same sites and observation periods) provide a much more complete picture than simply comparing Dec/Jan 87/88 with 1991. Wiki should be free from the biases created by removing important published information.
If you disagree that this information is correct or relevant, please discuss here. Dorre (talk) 00:06, 19 December 2012 (UTC)
Thanks Dorre and others. There are several problems for this article. The aim is to present an encyclopaedic introduction to all major aspects of the subject, suitable for people who don't know anything about it, based on reliable references, with further reading as relevant. This is really difficult because of the detailed controversy. At present we have, I suggest, too much detail about the research. Not too much research, just too much detail. More specifically, I'm used to reading reports with percentages of this and that, but here we have so many percentages that they really slow me down. I feel that they detract from the story that we should be telling, even though they are important parts of the story. I'd prefer to tell the story in a more readable form, using all the references, but to present in the article only the figures that are essential for understanding the story. Possibly, some of the present detailed text could be presented in the references. I might try a bold edit if I find the time, but in the meantime what does everyone think of this idea? Richard Keatinge (talk) 11:19, 19 December 2012 (UTC)
My suggestion would be to go back to version as at 28 Nov, before the 'editing war' started. One change I noted on 12 Dec was the replacement of "A review of head injuries and cycle use in four Australian states (Victoria, New South Wales, South Australia, Western Australia) and New Zealand concluded that there was no evidence that the laws had reduced head injuries" with "Helmeted cyclists in collision with motor vehicles had much less serious non-head injuries than non-helmeted cyclists". The original quote is a reasonable and accurate description of the findings of the paper. Yet it was replaced by a sentence taken out of context from the discussion, which refer to the results of a completely different paper studying injuries in the USA. So the "revision" is misleading - it implies that the helmeted cyclists in collision with motor vehicles were from Aus/NZ when they weren't and implies that this was a major finding of the review, and omits the main conclusions of the paper, as stated in the title "No clear evidence (of benefits) from countries that have enforced the wearing of helmets." Editing wars start when people see grossly misleading and inappropriate changes such as the above. So let's revert to to 28 Nov, and then see what other changes are necessary to bring it up to scratch. Dorre (talk) 20:04, 19 December 2012 (UTC)
IN the spirit of compromise I would be happy to revert to the Nov 28th edit as long as the graph which is currently on the site was replaced with the recently deleted graph which I feel is far superior and shows a much more complete and unbiased picture. I would then be happy to refrain from any further edits to the page that hadn't been discussed fully on this page and agreed to by several other editors. Dsnmi (talk) 00:43, 21 December 2012 (UTC)
A graph based only on capital cities is neither complete nor unbiased. Proportionately more people cycled to work in regional towns and cities and proportionately fewer wore helmets, so this is where you would expect to see the effects of helmet laws. Concentrating on capital cities also creates additional variation and bias if the weather on census day in a capital city is not conducive to cycling. The cold weather (min 0.8 degrees C) in Melbourne in 1986 could well have reduced the numbers cycling to work. The results also vary depending on how capital cities are defined, e.g. "by 2011 inner Sydney rates of cycling have almost doubled, but almost halved in the greater metropolitan region" http://theconversation.edu.au/more-cyclists-that-depends-on-where-you-live-11154
Unlike capital cities, ABS census data for entire states is readily available, easy to verify, and much less biased. The graph for all states except the NT (which partially repealed its helmet law and had 3.24% of people cycling to work in 2011 - higher than any State on the graph, or even the ACT's 2.82%. Apart from SA and the ACT (with only 10,415 and 2,185 people cycling to work in 1986), there would certainly appear to be an effect of the helmet law, consistent with what is seen in the aggregated data. Possibly the best approach is to show the simple, aggregated data, and refer people to the additional data on individual states, noting that some states have a lot more cyclists than others.Dorre (talk) 05:08, 21 December 2012 (UTC)
Okay I'd be happy if we revert to the nov 28th version with the inclusion of that graph. Thanks for putting in the effort to make it. Can I suggest that once we revert we delete this entire talk page (which is pretty much irrelevant) and agree to discuss any changes here before we make them? 08:33, 21 December 2012 (UTC)
As a temporary measure, I've updated the previous census data graph to include the 2011 census data. The 'editing war' graph was very confusing in that it ignored all data from non-capital cities (47,900 trips to work in 1986), and gave equal weight to Canberra, where only 2,272 people cycled to work in 1986. Individual cities may vary because of weather and levels of enforcement, suggesting that the most realistic picture is not to selectively show capital cities but the entire country. — Preceding unsigned comment added by Dorre (talk • contribs) 22:41, 19 December 2012 (UTC)
Agree to revert to the 28th of November version. I noticed some recent censorship disguised as neutral editing. That is neither healthy nor constructive. Respect and tolerance for others perspective is essential in any collaborative work. Let's not fool ourselves that we have "the facts" and others must be wrong, especially when some of the government-funded studies used to support a view are tainted with a conflict of interest to defend a controversial legislation.Harvey4931 (talk) 04:16, 21 December 2012 (UTC)
I agree some of the editing was definitely verging on censorship and I think we all agree this page should show no bias and simply report facts. Some of these facts will come from government funded studied which must appear here with equal weighting even if some people don't like the results they put forward. To omit them would definitely be a kind of censorship. 08:33, 21 December 2012 (UTC)
For census data, we should probably be quoting directly from the source, i.e., http://www.abs.gov.au/ documents, not interpretations by a third-party website. What specific original government documents, tables, graphs, etc. are they using to get their figures? Nelsonsnavy (talk) 19:28, 28 February 2012 (UTC)
As I understand it this is an area where the Primary source is definitely more appropriate. "Primary sources are often difficult to use appropriately. While they can be both reliable and useful in certain situations, they must be used with caution in order to avoid original research." The census data doesn't contain "original research" which wikipedia defines as " material—such as facts, allegations, and ideas—for which no reliable, published sources exist". In this case the census data is verifiable and reliable and so should definitely be the quote we use and not the third party website's interpretation. — Preceding unsigned comment added by Dsnmi (talk • contribs) 23:57, 28 February 2012 (UTC)
Good point. I stand corrected. My concern is with the neutrality/reliability of the interpretations, given the lack of citations there to specific census data/documents. Then again, one runs into that sort of dilemma a lot with this topic :) Do we have other sources analyzing the same data?Nelsonsnavy (talk) 20:19, 28 February 2012 (UTC)
I agree entirely. Usual practice is a reference to, say "Canadian census data". Actually finding the original source and checking the calculations can be effectively impossible and at best is very very time-consuming. I'd think it good practice for authors to give us the exact reference for their original data, preferably with URLs. In the days of paper, publishers often insisted on saving space in this sort of way, but that scarcely applies now. As a primary source, we can certainly use census data. For interpretations of it (even simple dividing of this by that), however, we need secondary sources. Richard Keatinge (talk) 09:19, 29 February 2012 (UTC)
Primary sources are to be preferred, but sometimes those sources need to be summarised or visualised, as is the case with Census data. That's fine, IMO, provided that the summarisation or visualisation doesn't constitute original research, that it can be independently verified as accurate, and that it comes from or is created by a reliable, trustworthy source. Therefore charts of Census data need to be referenced back to the primary sources in quite a lot of detail, and ideally the means of creation of the chart or graphic (data, computer program code, spreadsheet etc) should be publicly accessible to aid in verification. Finally, the charts and graphs need to be attributed to a trustworthy source. Graphs or charts without attribution, taken or borrowed from anonymous and unattributed web pages don't meet WP criteria, IMO. Tim C (talk) 02:27, 28 February 2013 (UTC)
Jake Olivier's removal of the census data
As everyone knows, census data are recorded every 5 years. Graphs usually show lines to join the points to make the trends easier to see. However, we don't know what happened in between the points, so when some event such as a helmet law happens between two points, the normal convention would be to put it in the middle of the line joining the two points, to demonstrate that the trend between the two points is meaningless.
A graph showing a downward trend from 1986 to 1991 and a line showing the introduction of the law in 1990 would be confusing, because some people might think that the line joining the two points represented a real trend and therefore that the reduction in cycling to work occurred before the helmet law.
I hope that there is general agreement that the census graph and other relevant information should be reinstated. Dorre (talk) 00:53, 25 February 2013 (UTC)
The plot in its current state is clearly wrong and should not be presented as such. The data is presented as a time series, therefore the x-axis is on a continuous scale and putting legislation dates in the middle of observed time points is misleading. The census date in 1991 was in August and many laws came into effect in July only one month prior. This distinction is incredibly important if you reference the NSW, SA reports and others which demonstrate adult cycling did not decline with helmet legislation.
Qld is an anomaly among Australian states and should be treated separately. Although not enforced immediately, the helmet wearing rate for Qld adults was much higher than that for Victoria, NSW or SA pre-law.
For the time being I have reverted Jake's removal of this data, the graph and the paragraph. Jake, I need to understand better what may be wrong with either the graph or the text; your explanation above does not clearly achieve this. It may help if you can draft a revised version here, preferably with references. Richard Keatinge (talk) 16:40, 25 February 2013 (UTC)
I suspect Jake objects to the lines depicting the helmet laws being shown half-way between the two affected points. With no information on what happened between points, mid-way seemed the logical place to put them. But I could probably just remove the lines connecting pre-law and post-law points to show where the law interrupted the trend - that way the vertical lines wouldn't be needed either - the dates could be reported as text.
There's one survey in Victoria showing very high helmet wearing rates within in month of the law and considerable reductions in cycling. I've always ignored the data, because the reductions in cycle use are confounded with time of year. The NSW survey of adult cycle use suffered the same problem - the post-law survey was taken at a time of year that was more conducive to cycling, making it extremely difficult to separate the effect of the law from a seasonal increase. However, if Jake is concerned that the laws didn't have immediate effects, the early post-law data from Vic (if I can find it) should put his mind to rest.
It would be even more interesting to compare the early post-law surveys in Vic (enforced legislation) with Qld (non-enforced legislation), where about 50% of adult cyclists disliked helmets so much that they were prepared to disobey the law, because they knew they could get away with it. It gives you a real insight into the potential for helmet laws to discourage cycling. — Preceding unsigned comment added by Dorre (talk • contribs) 21:12, 25 February 2013 (UTC)
As I've mentioned here and other places, the laws are CLEARLY in the wrong positions. The Victorian law was July 1990, NSW was January 1991 for adults and July 1991 for kids, SA was July 1990, the WA in July 1992 and Queensland in July 1991. You say that Qld did not enforce their law for 18 months, yet you also say there is evidence the helmet wearing rate for adults increased to 50% without enforcement yet with legislation which is much higher than pre-law levels for other states. It is misleading to lump Qld with other states that enforced the law when it came into effect.
The current graph puts helmet laws at 1988.5 for Victoria, NSW, SA and Tasmania; and puts the law at 1993.5 for Qld (enforcement only), WA and ACT. The x-axis is on a continuous scale and the graphical objects are clearly in the wrong places. There are 60 months in a five year period and you've put the law near month 30 instead of month 59 where they belong since the 1991 audit was in August.
Your comment about seasonality and riding is immaterial if measurements are taken near the same time of the year. Were they? The City of Adelaide counts found in Marshall and White's 1994 reported 1,747 cyclists in Sept 1990 and 1,797 cyclists in Sept 1991 (one month after the census date). That's a 2.9% increase in ridership pre- to post-law. Why is that not mentioned anywhere? The plot as it stands makes it appear ridership was declining in SA around that time. Isn't that misleading?
Putting the laws midway is not the logical choice and is clearly misleading. It is also unclear why the y-axis stops at 2.5% when ACT goes well above it. This is a common graphical trick to make the variability in the plot seem bigger than it actually is.
What is also missing from all of this is an actual statistical analysis of this data. Are the changes in the proportion of cycling on the census date significant? Are these changes within estimates of natural variability. Although this is called a census, it is clearly only an estimate on days five years apart which therefore has some measure of uncertainty attached to them.
I linked to a plot of transportation mode in Australia from 1900 to 2010. The variability seen in these plots whether up or down is minute compared to the heydays of the 1940's when the percentages were upwards of 9%. The changes in ridership from state to state would be nearly imperceptible if the y-axis extended that far.--JakeOlivier (talk) 21:52, 25 February 2013 (UTC)
The new graphs do not show the position of the law, so cannot be misconstrued. The y axis has also been fixed. It is important not to cherry pick the data. There are clear differences in ridership between capital cities and regional and rural cities. While I don't think it's relevant to go back to the heydays of the 1940s, it is useful to compare pre-law data for regional areas (total of 40,798 cycling to work in 1986) to capital cities (total of 42,318 in 1986). Pre-law, much higher proportions cycled to work in regional areas, as might be expected, given the traffic problems in capital cities. The reduction in cycling in regional areas might be considered an estimate of the effect in areas that are conducive to cycling. This compares to capital cities that are less cycle-friendly so it's much harder to separate the effects of the law from other factors, e.g. difficult traffic situations that also discourage cycling. All graphs have been constructed directly from ABS data that is readily available on the ABS website, so easily verified and checked. Dorre (talk) 08:49, 26 February 2013 (UTC)
Dorre - The new graph is an improvement but it's still not correct. Queensland is clearly different than the other states having a helmet law by the 1991 census date and enforcement by 1996. It is misleading to only focus on enforcement considering helmet wearing increased to 50% with the law. I've mentioned this several times now and you repeatedly ignore this comment.
How exactly is pointing to surveys showing no change from the year prior to the SA law to the next year "cherry-picking"? A fundamental principle of time series analysis is that observations nearer in time are more important than those farther away. Why isn't the SA report presented anywhere on this page? It is clearly relevant. In a report analysing the census data for cities, Mees and Groenhart conclude "Cycling is of negligible importance as a travel mode for work trips in all cities except Canberra". The small proportion of those cycling to work varies very little from state to state from 1976 to 2011. This variability only appears large through the use of misleading plots. Why isn't this report mentioned anywhere? You can find it here.
http://mams.rmit.edu.au/ov14prh13lps1.pdf
What is concerning about ignoring the 1940's peak is that the strong inferences being made about changes from a bit over 1% in 1986 Victoria to just under 1% in 1991 is an exaggeration. Using data in the Mees report, the percentage using a bike as primary transport for all capital cities was 1.14% in 1986 and 1.13% in 1991. There is no reasonable statistical analysis that would indicate that's an important difference. Importantly, most Australians were subject to helmet legislation by the 1991 census date. By the time the 1996 census was taken there are many other factors that could be influencing "trends". I emphasise "trends" because a lot of information is missing for observations taken on a single day five years apart. What is clear from this data is that cycling is not a major form of transportation in Australia from 1976 to 2011 (see Mees report).
You state "These states all show departures from the trend suggesting that the law discouraged cycling to work?" How do you know this? You give no citation and this does not appear to be a neutral point. For instance SA peaks in 1981 followed by a decline in 1986 with a decline of similar magnitude in 1991 after the helmet law. How exactly does a steady decline support your point? Changes in time series happen for many reasons (especially for observations so far apart) and alternative explanations need to be explored before drawing inferences. The sentence should be deleted as it is unsubstantiated by any analysis.
You state "As in some other states, the law may not have been enforced immediately -- the trend in SA..." This is not substantiated by any analysis or given a citation. Considering the law came into effect prior to the 1991 census, is there no other explanation for the small decline more than five years after? It is unreasonable to assume helmet legislation had a huge impact by 1996 when there was no evidence of an immediate impact in 1991.
You state "In Qld, no provision was made for penalties, so the law was widely ignored (about 50% of adults wore helmets) until 1 Jan 1993, when penalties were introduced." Helmet wearing surveys usually put helmet wearing at no more than 20% pre-law for adults in Australia. How is an increase to 50% justify your statement that the law was largely ignored? That amount is larger than any Australian state pre-law for adults. It is not as high as other states post-law, but an increase of 30% in helmet wearing is not small.--JakeOlivier (talk) 01:55, 27 February 2013 (UTC)
Dorre - In addition to my comments above, the use of the wording "The trend in SA shows a greater reduction..." is not substantiated. The pre-1991 trend in Adelaide is non-significant by linear or log-linear regression (with only 3 time points it would difficult to do anything else). This remains true even when the 1991 data is included. This is also true for Melbourne and the others. This enhances my earlier point that these graphs in their present state are incredibly misleading. There only appears to be great variation because the y-axis range is so small. In statistical terms, changes in percentage points less than 5% (in this case even smaller) is inconsequential. This wording is simply not true.
You now use the terms "widely ignored" to describe the helmet law in Queensland. How is that true? 50% is half of adults and is a large increase in helmet wearing when starting from 20%. The other states with enforcement top out at around 80% or so. The point is there was a clear increase in helmet wearing in Queensland after the law among adults even without enforcement. Your wording is therefore misleading.
In the next section, there is the statement "...regional areas such as small country towns, which are generally more conducive to cycling...". Where is the evidence to support that statement? Cities have more motor vehicles but they are also more likely to have cycle ways. Are there differences in injury rates relative to exposure? I'm unaware of any such study.--JakeOlivier (talk) 22:06, 27 February 2013 (UTC)
Phrase removed. Most people would think it blatantly obvious that places with nearly 3 times the percentages cycling to work are more conducive to cycling. But it's not worth arguing. The same for the definition of trends. Removing the offending word does not change the data or the interpretation.
As for "widely ignored", about half of cyclists were not wearing helmets in 1992 - most people would consider "widely ignored" appropriate for those circumstances. Alternatively, assuming (as in NSW) a 26% pre-law wearing rate, then twice as many cyclists were ignoring the law as were persuaded to wear helmets when the law was introduced.
Cherry-picking is when you focus on a small area which may not be representative of the whole country, or even the whole state. In 1986, more than 8,000 people cycled to work in Adelaide, - the 1850 counted cycling through the CBD represents a small proportion of the total. It would be very interesting to compare the trends for a longer period to see how they fit in with the census data.
Michael Walker's trends for Sydney from 1990 to 1996 were quite different to central Adelaide - http://cyclehelmets.org/jpg/1241_11.jpg The drop in cycling in Perth also coincided with the helmet law, according to counts on the Causeway and Narrows bridges.
Moreover the declines did not stop in 1996. "Unfortunately, the general upward trend since 2000 is simply reversing a rapid downturn that occurred between 1996 and 2000 when the percentage of commuters travelling by bicycle plunged 73% nationwide (from 1.9% to 1.1%)." - See Fig 6 of http://www.infrastructureaustralia.gov.au/publications/files/Cycling_Infrastructure_Background_Paper_16Mar09_WEB.pdf Of greater concern is that these apparent declines in cycling were not mirrored by declines in injury rates, implying a substantial deterioration in safety. It's good to see that this has started to reverse, and that by 2006 we were back to 1996 levels, if not to pre-law levels of cycling. Dorre (talk) 08:40, 28 February 2013 (UTC)
Dorre - As I've tried to get across to you before, you continue to make statements without citation or evidence. You cannot call something a trend when it is indistinguishable from a horizontal line (i.e., pattern not changing over time). Your comment that "most people" would agree with you is more indicative that most people don't understand statistics, which is unfortunately probably true.
50% wearing helmets post-law is not anywhere near "most" or "widely" of anything. It's half of cyclists. If you randomly chose a cyclists in Queensland in August 1991, the probability they would be wearing a helmet would be equivalent to flipping a coin. That's not as high as other places but clearly different than jurisdictions without helmet laws. Your plot is still misleading.
Cherry-picking is actually choosing data or results that support your conclusions while ignoring evidence to the contrary. That is what you have done. The SA report data is clearly relevant here and points to misleading information through selective citation/cherry-picking on this Wikipedia page.
Differential patterns in different places indicates heterogeneity. So something happened differently in Sydney than Adelaide. It wasn't helmet laws since they both had them. What happened to one and not the other?
It is curious you're attaching very small changes in cycling to work as being due to helmets. The attached report did not mention helmets at all. On the other hand, there's a nice plot of cycling as percent of transport mode. Australia is above the UK, US and Canada (countries most similar to Australia on many, many levels), yet none of these countries have adult helmet laws (some US cities do though). Your hypothesis of helmet legislation as a deterrent to cycling doesn't match here.
You appear to be insisting that helmet legislation is a deterrent but only after the census date in 1991. How can that be remotely possible? Most of Australia was subjected to helmet legislation by the 1991 census date but cycling to work in the capital cities changed from 1.14% to 1.13% for 1986 and 1991 respectively. That means 1 less cyclist for 10,000 commuters. Not very convincing. Your hypothesis then depends on cyclists not being initially deterred by helmet legislation yet somehow had a change of mind over five years later. Why would that be the reason? Were there no other changes that could explain such minor declines in some places? You appear to have chosen your conclusion and then moved the goalposts when the data didn't match.--JakeOlivier (talk) 05:56, 1 March 2013 (UTC)
My definition of "cherry picking" is when you report data for a small subset, instead of the whole lot. For example, The Conversation - http://theconversation.edu.au/more-cyclists-that-depends-on-where-you-live-11154 states: "while the proportion of people cycling to work in inner Sydney has increased markedly since 2006, the proportion cycling to work in outer Greater Sydney has gone down consistently since 2001." Anyone who cherry picks the inner Sydney data and compares it with injury data for the whole of NSW cannot expect to end up with unbiased estimates.
I don't understand your argument about non-enforced laws being anything like the deterrent of enforced laws. Which is the more likely option for someone who doesn't like helmets - to follow the example of 60% of teenagers and 50% of adults and ignore the law, or do something drastic like give up cycling? Why give up cycling when you can just ignore the law?
Nobody has ever said that helmet laws are the only deterrent. I remember reading that 50-60% of London peak-hour commuters wear helmets voluntarily, presumably because of perceived dangers. In areas where cyclists feel threatened by traffic, many cyclists are likely to think that wearing a helmet is well worth the inconvenience. I would imagine that helmet laws are much more likely to deter cycling in areas with low helmet wearing rates and low perceived dangers. If the proportions cycling to work are an indication of the how pleasant it is to cycle, you'd have to say that the 1986 data of 3.14% cycling to work in regional areas beats the 1.14% in capital cities by almost a factor of 3. This is where you'd expect to see the effect of helmet laws, not in areas where it's too scary to cycle without a helmet. Anyone there who didn't like helmets wouldn't want to cycle there anyway!!!
I've no objection to reporting the Adelaide data, as long as it is balanced by the longer series of counts at 25 sites in Sydney, the counts on the Causeway and Narrows bridges in WA, and the information on commuter cycling - the "rapid downturn that occurred between 1996 and 2000". The problem with reporting every single piece of available data is that it becomes harder to see the wood for the trees. Should we do the same for head injury rates? The current article compares cyclists with other road users in WA - should we add all the comparisons of cyclists and pedestrians in Vic and NSW and for all road users in SA?Dorre (talk) 08:56, 1 March 2013 (UTC)
Harvey4931's deletion of the summary of peer-reviewed references
Harvey4931 has repeatedly removed summaries of peer-reviewed work assessing changes in head injury hospitalisations around the NSW helmet law. The quoted work also gives reference to a criticism by Chris Rissel which was also removed. His/her comments state "The quoted studies are tainted with a conflict of interest and misleading. Wikipedia is not the place to debate controversial studies", "Wikipedia should not be used by academics to promote themselves, particularly when the description is one-sided." and "removed POV material. This study does not demonstrate effectiveness of helmets. A number of other factors, including lower speeds & bicycle lanes, could explain". None of these criticisms have been substantiated and appear to only be Harvey4931's opinions. Although I am a co-author on two of these references, I did not write these passages. Although, I made one minor change to correct the year the study was conducted/published.--JakeOlivier (talk) 21:40, 27 February 2013 (UTC)
I agree entirely with JakeOlivier (disclaimer: I am also a co-author of one of the studies). These studies are directly relevant to the topic of this article, they have been undertaken by experienced researchers at reputable academic institutions, and are published in a reputable, peer-reviewed scientific journal. The amount of Australian research on the topic is not large, and thus every study is noteworthy and deserves to be mentioned and referenced in this article. Similarly, critiques of these studies published in scientific journals, and responses to those critiques, are also worthy of mention and reference. It is not acceptable for an editor to delete such material on the basis of unsubstantiated personal opinions about the quality of the research. I also note the unsubstantiated allegation of conflict of interest in an edit summary. Please note the WP guidelines on Wikipedia:Conflict of interest and in particular WP:SELFCITE and Wikipedia:SELFPROMOTE. Nothing in the description and citation of these papers is in conflict with these policies. We are very happy to debate and negotiate the form of words used to describe these studies, in order to arrive at an accurate description with a NPOV, but we do not accept attempts to suppress mention of and references to these studies in this article. That said, the current descriptions of these studies are terse and only mention factual results of these studies. WP articles are not intended to be a vehicle for criticism or critique or discussion or debate about scientific studies. WP articles may mention, summarise or refer to such criticism or critiques or debate published elsewhere, provided the sources are reliable as per WP guidelines, but WP is not a place for primary criticism of and debate over studies. WP is an encyclopaedia, not a forum (Talk pages excepted).Tim C (talk) 03:08, 28 February 2013 (UTC)
Harvey4931 inserted the commentary
"These figures are cycling injuries per population, they do not reflect the risk of injury per cyclist. According to data from this study, between 1991 and 2000, the number of arm injuries doubled, while the number of head injuries increased by 40%. 1991 was a transitional year where the helmet law was partially implemented. Cycling numbers kept declining in the early 1990's, a 1996 cycling survey in Sydney revealed that cycling counts were 48% below the level before the law.[1]. There is no evidence of cycling levels recovering in the late 1990's. Despite a much lower number of cyclists, head injuries and arm injuries increased significantly for a decade, suggesting a higher risk of accident."
which I have deleted. The original study did, in fact, compare head and arm cycling injuries per cyclist as estimated from bicycle imports, cycling counts into the Sydney CBD and ERASS participation surveys (no other yearly estimates of cycling exist for NSW during that period). When compared to cycling numbers, head injury rates always declined -- this pattern was not always observed for arm injuries. Figure 2 shows standardised values over time for injuries and estimates of cycling numbers.
note entirely true. Between 1991 and 2000, none of the data above was available. The estimate of cycling relies largely on population growth.Harvey4931 (talk) 08:52, 28 February 2013 (UTC)
Wording has been enhanced to include injury per cyclist and the years involved. The second statement is not even nearly true. Cycling participation in NSW increased 50% from 2001 to 2010 (about 400K to 600K), the NSW population increased by only 10% from 2001 to 2011 (about 6.6M to 7.2M). This is easily verifiable from Table 2 in our paper.--JakeOlivier (talk) 03:59, 1 March 2013 (UTC)
If cycling numbers continued to decline, why would a CENSUS of hospitalisations over that period show no trend for head injuries and an increase in arm injuries? Harvey4931's reference is to a cycling survey that does not appear to be anywhere online nor does it appear to have been peer-reviewed. There are no distinctive changes in either time series with the exception of the helmet law and the increase in cycling infrastructure expenditures. If cycling declined by 48%, as Harvey4931, then head injuries would have increased by an equivalent amount. That would be beyond highly unlikely.
1991 was not a "transition year" for NSW. The law came into effect instantaneously. No infrastructure needed to be built. There are issues with kid and adult laws being effective six months apart. But our 2011 paper analysed injuries relative to the helmet law dates (be the injured a kid or an adult). This paper is primarily focused on the impact twenty years hence and 1991 is the starting point.--JakeOlivier (talk) 02:29, 28 February 2013 (UTC)
Of course 1991 was a transition year. The child law only came in July. Enforcement was initially minimal.
Either you accept valid criticism of misleading claims related to your study, or the misleading claims will be removed.03:39, 28 February 2013 (UTC)
Harvey4931 - this has nothing to do with accepting criticism. You comments are simply incorrect (this has been demonstrated before). As it relates to our analysis, 1991 is the starting point. It is certainly true that helmet legislation was introduced at different times for kids and adults, but the comparison is between head and arm injuries. So, when we compared these values in 1991, for say kids, helmet legislation, if influential in deterring cycling, would affect them both during the same time period (these two injury mechanisms are clearly dependent as cyclists cannot remove their heads from their arms). There is clear evidence in the data, when compared with subsequent years, 1991 fits the overall post-law time series pattern.--JakeOlivier (talk) 04:08, 1 March 2013 (UTC)
The "List of Concerns" section
I deleted this because, apart from being improperly formatted, it appeared to be a set of demands for changes in helmet safety labelling, complaints about governments not heeding advice or not following due process, and a set of personal concerns about helmet effectiveness and safety. Many of the points lacked references. One used another WP article on helmets as a reference - that is against WP policy (WP can't be used as a reference for itself). WP is an encyclopedia, not a soapbox. It is not OK to include sets of demands in an article, nor complaints about government processes. Descriptions of government processes that are documented and referenced elsewhere are OK, but not whinging that due process wasn't followed or that advice from lobby groups or petitions was not heeded.
Tim C (talk) 02:27, 6 March 2013 (UTC)
'List of concerns' Basically probably 3 groups of people exists with regards helmets, those in favour or either helmets or laws, those opposed to either and those not bothered. Those opposed will probably have most concerns. After 20 years of debate anyone accessing the wiki site may wish to have a clear idea of any concerns that people have. Providing a list of concerns gives then direct access to precise information with references. Some concerns on a list could include reasons why helmets are promoted or the concerns people had for introducing helmet laws.
Concerns have been expressed about helmets/helmet laws across a number of articles. Providing a 'list of concerns' allows readers to quickly assess the main concerns and in the articles that they where published. This directly benefits the public in understanding the issues and examining the evidence if needed. My view is that a list should be provided to assist the public and bring together issues that may need consideration. it is really part of free speech allowing people to discuss issues that concern them.
Colin at cycling (talk) 22:58, 7 March 2013 (UTC)
I'm sorry, Colin, but your list of concerns is really not suitable for Wikipedia in anything like its current form. We are here to provide a neutral, readable summary of the facts and of all sides of the various arguments, based on secondary sources wherever possible and on consensus. The present state of this article offers a lot of room for improvement, but none for adding polemic. Free speech is not a Wikipedia policy; we may have to include primary sources to some extent, but in most articles which describe an ongoing controversy, we do no more than note its existence. Here, when we have achieved a neutral summary of reliable sources, we will certainly find that the controversy speaks for itself. Richard Keatinge (talk) 08:22, 8 March 2013 (UTC)
Non-enforced laws
I undid Dorre's removal of the statement " It is important to note, however, that bicycle usage in all capital cites was 1.14% in 1986 when no helmet laws existed and 1.13% in 1991 when most Australians were mandated to wear helmets." without being signed in. Therefore a UNSW IP address is given instead. I am confused why Dorre indicates "Need ref showing non-enforced laws discourage cycling". My inserted comment is a factual statement regarding the Ride to Work census data and the fact that most of Australia had a helmet law by the 1991 census date. JakeOlivier (talk) 04:59, 7 March 2013 (UTC)
Mainly because you haven't come up with one logical reason why anyone who disliked helmets would do something drastic as give up cycling when they could just follow the example of every second cyclist and disobey the law. "Most Australians" seems misleading when there was no helmet law in WA or the ACT, and anyone in Brisbane knew perfectly well that they didn't have to obey the law if they didn't want to, because the police had no powers to enforce it. That leaves Sydney, Melbourne, Hobart (with 385 people cycling to work on census day in 1991), and Adelaide where the law came into effect a few days earlier. If you want to know the effect of enforced laws on cycle use, you have to look at the data for places with enforced laws, rather than bulk up the data by including places that did not have enforced laws. Dorre (talk) 08:51, 7 March 2013 (UTC)
Helmet wearing for adults in Queensland increased to 50% with a helmet law but no enforcement. That is much more than any Australian state pre-law. How does that not set Queensland apart? There is clearly a third category separate from "no law" and "enforced law". Lumping Queensland with other states without legislation is not correct as it is unlike every other state.
"Most Australians" is used because the percentage of Australians with a helmet law by the census date of 1991 is greater than 50%. I'm not sure why that needs defined. This is true whether Queensland is included or not. South Australia had a helmet law over month by the time of the census date. That's clearly more than a few days and is more than twenty commuting days. Plenty enough time for someone to stop cycling because they don't like the helmet law.JakeOlivier (talk) 04:15, 11 March 2013 (UTC)
Proposed clarification and improvements
While some considerable work has been put into this page, it could still benefit from some clarifying changes. In order to improve this page I am floating a number of ideas in the Talk section before making any edits. It seems that the main purpose of this page is to discuss the various aspects of mandatory helmet legislation in Australia, but this is not entirely obvious at first glance. If that is the case the page would benefit from this being made clearer in the titles. By way of clarification the main title could be changed to something like “Bicycle helmet Laws in Australia”, the “injury rates” section could be changed to “Bicycle helmet law efficacy” and the “Bicycle Usage” section could be changes to “Helmet laws and ridership”.
It seems there are two fundamental questions around helmet laws: 1) did they reduce head injuries at a population level, and 2) did they change levels of ridership? These questions seem to approximately correspond to the “injury Rates” and “Bicycle Usage” sections. It seems the aim of the “injury Rates” section is to summarise assessments of the effectiveness of the various state laws. For readers to get a clear idea of the evidence it would help to separate rigorous studies from opinion, possibly through some kind of hierarchy of evidence. For example subsections within “injury Rates” might be something like 1. Peer-reviewed analyses, 2. Peer-reviewed commentary, 3. Grey literature (e.g. reports from government or universities), and 4. Other (e.g. website, blogs, opinion pieces). The last option could be omitted. A similar hierarchy of subsections could also be applied to the “Bicycle Usage” section.
The “Bicycle usage: opinions in relation to the helmet law” subsection would be better combined with the “Public attitude to helmets” section.
The references do not appear to provide a comprehensive list of the literature on this subject. There seems to be many references to two websites that appear to provide opinion only and do not seem to be entirely impartial. Also papers by Robinson are well represented, while others are not cited at all (e.g. Cameron, 1992; Carr, 1995). The article could certainly be improved by making the reference list more comprehensive, and by removing POV website references. A link to the original reports by Walker, and by Smith & Milthorpe would be helpful for the “bicycle usage” section. S.Walter.1 (talk) 23:56, 28 February 2013 (UTC)
I am in broad agreement with these suggestions, although I don't think the name of the article should be changed, at least not yet. Completely agree that all relevant peer-reviewed scientific papers on bike helmets and helmet laws in Australia should be referenced in this article - the Australian literature is not so vast as to make that impractical. Likewise, key reports and source documents should be directly referenced as primary sources, rather than referred to via commentaries on them. The commentaries and papers which report on an analyse these primary sources should also be cited, but clearly flagged as secondary sources in the text. To that end, I plan to add direct references to the various Victorian and NSW helmet surveys from the early 1990s. However, the current section headings are confusing at best (eg what is meant by "control groups" in this context?), and the following changes are proposed:
What is meant exactly by "control group" in the context of this section? It doesn't correspond to the usual epidemiological or statistical senses of "control group". However, if what is intended is the presentation of changes in cycle commuting rates (in Census and other data) in the context of concurrent changes for other commuting modes ("control groups"?), then that is a good idea. It is very useful to see how cycling has changed in relation to car use and public transport use. All these data are available in the Censuses. Tim C (talk) 23:10, 1 March 2013 (UTC)
Bicycle usage: capital cities vs all other areas -> combine with above
I wonder if this is suitable for inclusion in this article at all - I'd suggest a well-referenced sentence or two at most, see below. Richard Keatinge (talk) 12:09, 1 March 2013 (UTC)
There are quite distinct differences in cycling participation between metropolitan areas and regional towns in Australia, and indeed, within metro areas. Likewise the evidence from the 1990s NSW and SA helmet and ridership surveys was that there were differences in effects between cities and regional towns.Tim C (talk) 23:10, 1 March 2013 (UTC)
Bicycle usage: opinions in relation to the helmet law -> Opinions on cycling in relation to helmet laws
Bicycle usage: changes without concurrent control groups around the time of helmet compulsion -> Surveys of helmet use and cycling participant before and after introduction of helmet laws
The problem with the current subsection title is that it presumes (or at least suggests) hat trends are due to helmet laws. They may be, but they may not be, and the reality is that there are many influences on bicycle ridership, helmets being just one of them.Tim C (talk) 23:10, 1 March 2013 (UTC)
Comments sought. However, I will go ahead and add the references to the Vic and NSW surveys under the current "Bicycle usage: changes without concurrent control groups around the time of helmet compulsion" for now. Tim C (talk) 05:37, 1 March 2013 (UTC)
I agree with both Scott and Tim. The sections given are inadequate and don't accurately portray the issues relevant to a page titled "Bicycle helmets in Australia". This discussion is definitely moving things in the right direction. Another valid point is the over-reliance on secondary analyses, re-analyses and commentaries. Much of the primary source material relevant to this page is missing.--JakeOlivier (talk) 07:24, 1 March 2013 (UTC)
I agree that this page needs some rethinking. My main worry for some time has been about its un-encyclopaedic nature, as I have mentioned above. My own judgement, and that of my Wikipedia consultants (any innocent victim I can ask to have a look at the article, usually family and friends with no particular interest in the subject at hand), is that the sections on injury rates and bicycle usage are so over-detailed and full of percentages etc that they are simply not manageable for the usual encyclopaedic reader. This should be an encyclopaedic article for the casual reader and should not try to be a formal review of the literature, nor should it necessarily try to give details of all the primary literature (though we might choose to include some at least). I remind everyone that we should look for secondary literature wherever possible, including peer-reviewed websites. Primary literature is less suitable, though here we can certainly make a case for including it in some form because of the general lack of good quality up-to-date reviews.
I have recently done, I think, quite a good job on an article with analogous problems of strong points of view in the primary literature and a shortage of really focused reviews, Fall of the Western Roman Empire; the main issues are illustrated rather than formally reviewed, while giving enough references that the reader can pursue any major hypothesis that takes their fancy. Here, I am mulling over a bold rewrite which would remove most of the percentages and associated details of individual studies; it will outline the major issues and major controlled comparisons and will mention (without detail, either as a list or as references) that there is a significant primary literature expressing strong points of view. I feel that this will give us a far more readable article with far less argument. As a fringe benefit, it may tend to protect editors from hints of WP:CONFLICT. Comments? Richard Keatinge (talk) 12:09, 1 March 2013 (UTC)
Source documents should always be given prominence by referring to them first. For example, for the Victorian, NSW and SA helmet wearing and ridership surveys, the source reports of these surveys (from which everyone else has drawn their data, and their conclusions) are readily available online (the SA report soon will be), and thus must be mentioned and clearly referenced first, to allow interested readers the opportunity to examine them and form their own conclusions. Then secondary sources should be mentioned. However, where secondary are available online, there is no need to copy chunks out of them into the WP article, nor to reproduce endless percentages and other figures from them. At the most, a very terse summary should be included in the WP article. It is also important to make clear, in the text, the source of the commentary (which is often criticism of studies or results, it seems). For example, in reference to the numerous case-controls studies on helmet efficacy, there was an assertion that "...all these studies have been criticised" followed by a reference to an article by Dorothy Robinson. It is important that the reader is aware that it is a single author who has apparently criticised all the case-control studies, and that that author's name appears in very many places in this article.Tim C (talk) 23:10, 1 March 2013 (UTC)
In terms of structure of the page, can I suggest re-organisation in the following major sections:
Intro
History
Effects of helmet laws on helmet wearing and numbers of cyclists
Effects of helmet laws on cyclist injuries (or injury rates)
Helmet laws and public bike share schemes
Public opinion on helmet laws and cycling
Within each section there can still be the subsections as discussed above. Such a re-organisation could be done by just moving existing subsections around as a whole and a bit of section and subsection re-labelled, as discussed above. Further refinement of subsections could then proceed. At the moment, the article jumps all over the place.Tim C (talk) 23:10, 1 March 2013 (UTC)
The utility of the source documents is aptly demonstrated by Tim's citing the NSW survey dates as "September 1990 and September of 1991, 1992 and 1993". If he hasn't read them and can't even cite the dates of the survey correctly, what hope is there for the average reader, who has to come to grips with other issues such as the confounding with time of year? I would suggest broad summaries first then references to the details.Dorre (talk) 00:44, 2 March 2013 (UTC)
Thank you, Dorre, for the gratuitous comment on my competence. The mistake was a typo, one which I would have corrected had you not done so first. In fact, I have read all the reports in detail, having taken the trouble in 2010 to visit the State Library of NSW to retrieve and photocopy the reports for 1990, 1991 and 1992, and the Fischer Library stack at the University of Sydney for the 1993 report. Alan Davies (The Urbanist blog on Crikey) also remarked on how difficult it was obtain a copy of these reports, which contain the raw data that Dorothy Robinson in particular has relied upon for the analyses in her oft-cited articles. Because they are important documents, but so hard to access, we asked NSW Roads and Maritime Services staff to have PDF copies of them placed on the RMS bicycle information web site, to allow other scholars and members of the public to easily examine the reports if they so wish. Tim C (talk) 03:29, 3 March 2013 (UTC)
Dorre - The NSW government has posted these reports under the names "Bicycle Law Compliance and Helmets - 1990", "Bicycle Law Compliance and Helmets - 1991", "Bicycle Law Compliance and Helmets - 1992", "Bicycle Law Compliance and Helmets - 1993" and "Bicycle Law Compliance and Helmets - 1994" found here http://www.bicycleinfo.nsw.gov.au/tools_and_resources/cycling_research.html. Calling them that certainly makes it easier for the uninitiated to follow, so I'm not sure what you're complaint is really about. You've inserted a reference that helmet wearing surveys (which are clearly different than ridership surveys) ran from 1990-1996. Beyond the 1994 report, nothing appears to exist anywhere except for a conference presentation that doesn't appear anywhere on the net. Why?
Why would you start with other people's interpretations first? It's like jumping right into the middle of the novel. It would also overemphasise the work of Robinson which appears to be exclusively critical of helmets. The original reports are much more balanced and, while being the source material, often draw their own conclusions about helmet wearing and ridership. These documents are clearly the starting point.
A comment was given above about the criticisms of case-control studies which appears to be exclusively attributed to Robinson (and possibly others in the non-peer-review blogoshpere). This seems incredibly strange that it is mentioned at all considering all case-control studies have shown a benefit. Even Rune Elvik's often misquoted re-analysis of Attewell's paper demonstrated a clear benefit of helmets to protect the head. This criticism is eerily similar to Fisher's criticisms of case-control studies for the association of smoking and lung cancer. He was never convinced in the end and I wonder if the same is true for Robinson. There are only "hypothetical" studies in Elvik's work and there's no evidence any of them actually exist--JakeOlivier (talk) 05:55, 2 March 2013 (UTC)
Here's some entries from the 'Blogosphere' Jake talked about:
Lawlor, D.A., G. Davey Smith, and S. Ebrahim, Commentary: The hormone replacement-coronary heart disease conundrum: is this the death of observational epidemiology? Int. J. Epidemiol., 2004. 33(3): p. 464-467.
Ioannidis, J.P.A., Contradicted and initially stronger effects in highly cited clinical research. JAMA, 2005. 294(2): p. 218-228.
Lawlor, D.A., et al., Those confounded vitamins: what can we learn from the differences between observational versus randomised trial evidence? Lancet, 2004. 363(9422): p. 1724-7.
I added the Walker 1996 reference because it provides some information on the difference between counts in spring and autumn in NSW. Do you have a better reference? Or do think that it's simpler pretend it doesn't matter? Or OK to claim, as Tim did, that they all took place in September?
I wasn't aware that Robinson criticized helmets, only helmet laws, because they lead to reduced cycling, lost health and environmental benefits, and reduced safety in numbers. — Preceding unsigned comment added by Dorre (talk • contribs) 06:38, 2 March 2013 (UTC)
Dorre - Please stay on topic. None of the articles you've attributed to me are in any way related to my comments. It is well known that case-control studies are weaker than reandomised controlled trials, but that argument by itself is not a sufficient criticism of such studies related to helmets. As I mentioned, case-control studies consistently find a benefit to helmet wearing and head injury. This is not remotely equivalent to the infamous M&M link to cancer, for instance.
My point about Walker's 1996 presentation is due to the fact no one seems to have a copy to read. After reading Robinson's interpretation of his earlier studies, I was surprised to come away with a different impression of his data. This highlights the importance of the primary work in this area. Knowledge should be given firsthand if possible followed by second and third hand resources.
I'm incredibly confused by your statement Robinson is not critical of helmets. The statement given that all helmet case-control studies are contested comes from her work. Much of her helmet theories aren't backed by much evidence anyway. I once thought safety in numbers may be real until analysing the NSW hospitalisation record. As cycling participation has increased, so have non-head related cycling injuries. So, more cyclists = more injuries. Strange.JakeOlivier (talk) 10:15, 2 March 2013 (UTC)
Per capita cycling cycling participation in NSW is still well below pre-law levels. Gillham & Rissel's analysis of travel survey data showed a very small increase in the number of daily cycle trips (from 420,700 in 461,929) compared to a much larger increase in population, representing a 28% reduction in per capita cycling. I accept the argument that some of this may be due to changes in population structure, but it's still the correct denominator with which to compare per capita injury rates.
That comparison shows per capita head injury rates are about the same as in 1991, implying that they are about 28% higher, after adjusting for the reduction in per capita cycling. The really surprising statistic is that arm injuries have gone through the roof.
Not so reported road injuries. On a per capita basis, reported pedestrian fatalities in NSW are 27% of their 1990 values, and reported injuries about 38%. Reported per capita fatalities and injuries for NSW cyclists are about 45-47% of 1990. The Bambach paper reports 639 head injuries from bike/motor vehicle crashes between 2001-2009 and a tie-up rate of 71%. Coincidentally, 639/9 = an average of 71 per year. Assuming similar tie-up rates for head and non-head injuries, that means about 100 head injuries per year from collisions with motor vehicles. But your paper shows 706 head injuries and 1620 arm injuries in the year 2010. What's causing them? My guess (a guess based on personal observation) is that sports cycling and mountain biking have displaced transport cycling. I'm also going to guess that these activities result in many more falls of the bike and many more arm injuries than transport cycling. Some data on hospital admissions in the ACT show that 51% of cyclists in non-motor vehicle collisions have arm fractures, compared to 13.5% in collision with motor vehicles.
So the big increase in arm injuries has probably nothing to do with increased cycling but a switch from transport cycling to sports and mountain bike cycling, where it's not uncommon to fall of the bike and end up with a broken arm. I also suspect that the helmet law has been partly responsible, by frightening people off their bikes and making transport cycling a whole lot more inconvenient. To quote the current 'Bicycle helmets in Australia' page, for WA "Shopping trips fell by 55%, from about 5.2 to 2.3 per weekday per 100 people; trips for education by 79%, from about 8.2 to 1.7 per weekday per 100 people.[56]" All this is very interesting, perhaps it's even worth a mention on the page?Dorre (talk) 03:40, 3 March 2013 (UTC)
Dorre -- Gillham & Rissel's analysis in a non-peer reviewed journal made the analytic mistake of not accounting for changing demographics. It is well known that cycling rates can be vastly different among age groups. Australia, like many Western countries, has a glut of baby-boomers who were near or at retirement age for the 2011 survey yet much younger for the 1985/86 survey. When age is properly accounted for, there is an estimated 8% increase in cycling from the 1985/86 survey to the 2011 one. Our analysis can be found here (http://www.abc.net.au/environment/articles/2012/07/18/3546884.htm). However, I would caution anyone who wants to draw inferences about surveys so far apart using different data collection methods. The increase in cycling is evidenced in many surveys and cycling counts. The rest of your arguments fall apart from there and don't really require a response.
You're reference to my work is misleading as the Bambach paper analysed all hospitalisations in which a cyclist was involved in a motor vehicle collision over set time frame (2001-2009) while the other paper in which I'm lead author is a time series analysis for all cycling related head injury hospitalisations for years 1991-2010. Comparing your "estimate" for cyclists in a motor vehicle collision with all cycling head injury hospitalisations in 2010 is meaningless with regards to helmets as you've presented no connection with helmet wearing and/or helmet legislation. Remember this is data two decades after the law. The increase in non-head injury hospitalisations for cyclists increased at a steady rate post-helmet law. You claim this is due to an increase in sport cycling resulting in more arm injuries. Do you have any "real" evidence of this? I've searched the literature and turned up nothing. There does exist evidence that cycling in general is on the rise.JakeOlivier (talk) 01:42, 19 March 2013 (UTC)
I'd like to remind everyone of a few points. First, we really are supposed to base ourselves on reviews in preference to primary sources. Primary sources are occasionally useful to illustrate the points made by secondary sources, but that's about it. It might be possible to gain consensus that this article should present a complete list of primary sources in this area, but that case has not yet been made. I've argued already, and strongly opine, that this article already has far too many details from primary sources. I'm busy at the moment (far too much paperwork has to be in for my yearly review this week) but I do propose to try out a bold edit fairly soon.
Second, with all due respect for the expertise of our newly-joined editors and their justifiable pride in their own peer-reviewed primary publications, it is deeply inappropriate for any editor or group of editors to declare as above "we do not accept attempts to suppress mention of and references to" their own primary studies. I don't want to bite the newbies, but if we are to work constructively together to build a better encyclopedia, we can only do so by careful and tactful attempts to reach consensus on presenting relevant points, in accordance with Wikipedia guidelines. We should work hard at the difficult, perhaps un-natural, art of reaching agreement on presentation with people who do not agree with all of our own views on the meaning of what's being presented. It can be done - look at the Israeli–Palestinian conflict article for example. Not doing it is a waste of time at best.
I think that there may be some semantic misunderstanding going on here. As explained in the WP policy article Wikipedia:Identifying_and_using_primary_and_secondary_sources, there is a distinct difference between what is mean by primary, secondary and tertiary sources in historiography and the use of the term (or cognate terms) in science - the latter is obviously more relevant in the context of this article. In particular, the WP policy article notes: "Uses in fields other than history: In science, data is primary, and the first publication of any idea or experimental result is always a primary source. These publications, which may be in peer-reviewed journal articles or in some other form, are often called the primary literature to differentiate them from unpublished sources. Narrative reviews, systematic reviews and meta-analyses are considered secondary sources, because they are based on and analyze or interpret (rather than merely citing or describing) these original experimental reports." In the case of the helmet and cycling helmet survey reports, the documents contain the compiled data but also analyses and summaries of the data. Surely you are not suggesting that a WP article such as this should not refer to such documents? With respect to my comment about not accepting suppression of mention of and reference to our studies, that was in the context of an editor deleting the references to them shortly after they were added, with an edit summary alleging bias and conflict of interest on the part of the studies' authors (allegations which are totally without foundation and which we reject completely). These studies are the most recent of a just a handful of Australian studies in which actual data is analysed, as opposed to the very many opinion articles and commentaries,into the effects of helmets laws in Australia. Surely you agree that the existence and results of these studies warrant inclusion in this article? We added the description of and references to the studies because no other WP editor had done so, despite the papers having been published so time ago. Tim C (talk) 04:11, 3 March 2013 (UTC)
McDermott and Klug 1982, "Difference in head injuries of pedal cyclist and motorcyclist casualties in Victoria", reported 73 skull fractures for pedal cyclists compared with 31 for motorcyclists and concluded that pedal cyclists had a significant greater incident of fractured vault of the skull. They were mainly comparing adult motorcyclists (96%) to cyclists aged less 17 years of age (73%). Adult skull stiffness is higher than for children therefore they were not quite comparing like with like. They reported 181 pedal cyclist fatalities compared with 451 for motorcyclists. The travel data available for 1984/5 (about 7 years after their study period) detailed bicyclists spending 114,500 hours per day cycling in Victoria compared with 17,500 hours per day for motorcycling. Relating time of travel to skull fractures shows motorcyclists incur nearly three times that of bicyclists, a factor of 278% and have a fatality rate 16.3 higher than bicyclists and the overall injury rate for motorcyclists was 16.1 times higher. Motorcyclists generally wearing helmets were 16 times more likely to be killed or injured and nearly 3 times more likely to suffer a skull fracture compared with bicyclists who were generally not wearing helmets. With hindsight it was a mistake for McDermott and Klug not to relate injury and death to time spent travelling, making their findings unsuitable for considering overall safety. In addition they reported having no information on the cause of death. Their recommendation for a coordinated campaign, involving the Royal Australasin College of Surgeons, road safety and traffic authorities, the Educational Department, school principals' and parents' councils, and the media to increase helmet wearing rates was not based on reliable methods. Both McDermott and Klug were members of the Road Trauma Committee, Royal Australian College of Surgeons that requested the Government of Victoria to introduce bicycle helmet legislation.
In the History section, the article currently states:
"McDermott and Klug 1982 [9] reported on data from 1975 to 1980 for Victoria, with 73 fractures of the skull for pedal cyclists compared with 31 for motorcyclists and concluded that pedal cyclists had a significant greater incident of fractured vault of the skull. Their study reported the number of fatalities as 181 for pedal cyclists compared with 451 for motorcyclists but did not include information regarding exposure. The travel survey data available for 1985/6 [10] details bicyclists spending 114,500 hours per day cycling in Victoria compared with 17,500 hours per day for motorcycling. Per hour of travel motorcyclists generally wearing helmets were more likely to be killed or suffer a skull fracture compared with bicyclists who were generally not wearing helmets. Both McDermott and Klug were members of the Road Trauma Committee, Royal Australian College of Surgeons that requested the Government of Victoria to introduce bicycle helmet legislation."
Reporting the results of the McDermott and Klug study is fine, as is reporting the 1985/86 travel survey data for cyclists and motorcyclists (although relevance is marginal IMHO). However, the next sentence, "Per hour of travel motorcyclists generally wearing helmets were more likely to be killed or suffer a skull fracture compared with bicyclists who were generally not wearing helmets." is unreferenced and appears to be a synthesis of the two preceding sources in order to imply that unhelmeted cycling is safer than helmeted motorcycling. That may or may not be the case, but drawing such a conclusion in the article is synthesis, and given the following sentence about McDermott's and Klug committee membership, the intent of this synthesis seems to be to cast doubt on McDermott and Klug in some way. I think the synthesis should be removed, because it violates WP policy. Tim C (talk) 20:59, 7 March 2013 (UTC)
The information was included in the 2007 Velocity reports on cycling. Appendix
'Mistakes in helmet assessments from Australia' B3
[2][3]
McDermott and Klug were authors of a helmet report that by not relating to exposure gave the impression that cyclists were more at risk of head injury than motorcyclists. They were also part of a key group who had contact with the Government. Including these details in the History section helps to explain how the law appeared to be justisfed.
Adding ref to artcile.
Colin at cycling (talk) 09:11, 8 March 2013 (UTC)
I have checked the references given. One points to the programme page for a European cycling conference. I will remove that because it is irrelevant and probably a mistake. The other points to a paper delivered at that conference by a Colin Clarke. The relevant section in that paper titled "Safety comparison" gives two references in support of the cyclist-to-motorcyclist fatality and head injury rate comparison which appears in this WP article. One of those references is the 1996 paper by Dorothy Robinson, which doesn't mention motorcyclists at all, but does cite cyclist data. The other is a report from the Australian Federal Office of Road Safety, which is conveniently available online. I will substitute these references for the reference to the cycling conference paper. Tim C (talk) 21:06, 8 March 2013 (UTC)
I have added Citation needed tags to two of the Census data charts currently in this article. Clicking on these charts reveals that they were created by WP user Dorre, who does not have a user profile page. A WP user name does not constitute an identified and reliable source, as WP policy requires for all material which appears in WP article. The source of the data used in these graphs is provided by Dorre, but that is insufficient - a reference to where the graphic itself has been elsewhere published is required - otherwise these charts must be construed as original research, which is not permitted in WP articles. The graphs present a particular interpretation of the source data, as evinced by discussion about them by their author and Jake Olivier elsewhere on the Talk page. Therefore, their source must be made clear, or they should be removed. Tim C (talk) 02:25, 9 March 2013 (UTC)
It has been 11 days since citations were requested for the two graphs added by and attributed to User:Dorre in the "Bicycle usage" section of this article. User:Dorre has made several edits to the article and to this Talk page since then, but has not supplied references of where these graphs have been published external to Wikipedia (Wikipedia cannot be used as an authority or source for Wikipedia). It is proposed to remove the graphs if no external source which meets WP reliable source criteria is supplied in the next few days. Please note objections, if any, and reasons here. Tim C (talk) 02:30, 20 March 2013 (UTC)
The Report of a gaoling of a female minor in NT for failing to pay fines for failing to wear a helmet.
I have edited the statement added by User:Colin at cycling regarding the gaoling of a female minor in the NT for failing to pay helmet fines. The problem is that there appears to be just a single source for this, and that source is a 1998 report in a UK newspaper for teachers, not in Australian mainstream news media. Therefore there must be some doubt about the accuracy and verifiability of the report, and I have tagged it accordingly as a single source. — Preceding unsigned comment added by Tim.churches (talk • contribs) 21:37, 15 March 2013 (UTC)
I would like to draw attention to the WP policy
'Self-published as sources on themselves'
Self-published may be used as sources of information about themselves, usually in articles about themselves or their activities, without the self-published source requirement that they be published experts in the field, so long as:
the material is neither unduly self-serving nor an exceptional claim;
it does not involve claims about third parties;
it does not involve claims about events not directly related to the source;
there is no reasonable doubt as to its authenticity;
the article is not based primarily on such sources.
Primary source tags removed from statements by Sue Abbott and Alan Todd, as requested - agree that the these instances meet the WP guidelines listed above. Tim C (talk) 21:31, 26 March 2013 (UTC)
Erke & Elvik or Høye & Elvik?
User:Dorre has questioned an edit I made to a reference to a Norwegian report:
Personally, I'm at a loss to understand some of his [ User:Tim.churches ] edits, for example he "Corrected reference to Hoye and Elvik (not Erke and Elvik as stated)" I clicked on the link to the reference, a pdf file which lists the authors as Alena Erke and Rune Elvik on the front page.
The correct reference is indeed Erke and Elvik. I was mislead by this page about the report in question, , on their institition's web site, which clearly gives the authorship as Høye & Elvik: https://www.toi.no/article19378-29.html Presumably Alena Erke changed her name to Alena Høye at some stage. I trust that this adequately explains the rationale for the edit. Tim C (talk) 04:01, 23 March 2013 (UTC)
Edit by [[User:Kiwikiped] for non-NPOV reasons
User:Kiwikiped recently made this edit: (Deleted " but did not provide any case law to support this assertion" in relation to a quote from an article in a peer-review journal. This seems to be a distinctly non-NPOV comment on the article as so inappropriate, but correct me if I am wrong.)
This is just a note about why I consider that text appropriate and not in violation of non-NPOV. An article in the NZ Medical Journal is given as the authority for the text in question. However, on investigation of that source, an article written by Colin F Clarke, the only information contained in it on possible compensation discrimination is as follows:
Discrimination can occur in accident compensation cases where a cyclist was not wearing a helmet, compared to pedestrians or indeed motor vehicle occupants who received head injuries. The helmet laws result in unfair compensation and a biased legal process.
Unfortunately, no references at all are given in support of this assertion in the NZ Medical Journal article - no case law, nor other documented examples. Thus this appears to be an unsupported assertion by the author, and thus, I felt, warranted the qualification I added in the WP article. The principle is that just because an assertion is made in an article in a peer-reviewed journal doesn't mean it is automatically reliable - there is an expectation that such assertions, if repeated in WP articles, are themselves supported by authority to reasonable external references (or by reported data as the case may be). For these reasons, I don't think the qualification represents a non-NPOV and I propose to undo the edit. It is also worth pointing out that the article referenced was in a NZ medical journal and is about the NZ bicycle helmet laws, whereas this WP article is about bicycle helmets in Australia. Thus it is a stretch to use such an article as an authority for assertions about possible discrimination in compensation cases for road traffic accidents in Australia. Tim C (talk) 19:59, 17 March 2013 (UTC)
→Colin at cycling (talk) 09:41, 18 March 2013 (UTC)
The NZ article was an evalaution of their helmet law and included several aspects. Mention of discrimination was inserted in the text, summary and conclusions. To decide on discrimination requires a full consideration of the issues and the report provided for this level of consideration. The discrimination part was a general statement applying if compensation was being reduced because helmets were not worn. In part the example given was from the UK showing that other road users also had similar proportions of head injuries to cyclists. Australian legal opinion states that compensation can be reduced by 25% if not wearing a helmet. The NZ report mentions “Erke and Elvik (Norwegian researchers) 200710 stated: ‘There is evidence of increased accident risk per cycling-km for cyclists wearing a helmet. In Australia and NZ, the increase is estimated to be around 14 percent. It assessed a range of issues in deciding Discrimination can occur in accident compensation cases where a cyclist was not wearing a helmet, compared to pedestrians or indeed motor vehicle occupants who received head injuries. The helmet laws result in unfair compensation and a biased legal process. The statement was supported by the previous details with approximately 4 times more pedestrians dying from head injury than cyclists. Evidence to support the statement was provided in the details and the whole report. The article was peer reviewed and did not require additional references to support the above statement, it was considered acceptable.
In the UK there has been discussions about the reduced compensation aspect over a number of years, so it was widely known without providing detailed references.
Adding
" but did not provide any case law to support this assertion" questions the statement that was not done in the peer reviewed article. If anyone wants to challenge the statement they need to write a suitable article and have peer review or perhaps refer to a Human Rights court hearing.
The article coming from New Zealand who also have an all age helmet law (following the Australian example)and have similar proportions of cyclist to pedestrian deaths, indicating that it is suitable for including in the discussions. On balance leaving out " but did not provide any case law to support this assertion" is the more appropaite action. Colin at cycling (talk) 09:41, 18 March 2013 (UTC)
For what my opinion may be worth, I feel that Colin has provided sufficient support that we could reasonably include the comment if we think it's beneficial to an encyclopaedic article. I don't think that it is. I'd like to abbreviate a lot of the argumentation including this comment. Richard Keatinge (talk) 13:25, 18 March 2013 (UTC)
I see two distinct issues here (a) should the quote & reference be in the article and (b) should the added editorial be in the article? I expressed no view on (a).Regarding (b) I still see the added editorial as exactly that. My first thought, without reading the paper, was that the reference might follow something along the lines of "we observe A, B, C... and conclude ..." and maybe none of A, B, etc. is "case law". Given that the reference comes from a refereed journal then, without evidence (or even assertions) that the journal is an unreliable source, the content, argument and conclusions of the paper have been judged to be a reasonable academic presentation. To add the "but did not provide any case law to support this assertion" in this light is editorial, appears to imply that the quality of the paper is lacking, and appears to assert the opinion that such a statement could only be justified by case law. As such adding those words appears to be comment/editorial in an article which I understand is meant to present information - in this case that such and such a statement occurs in a paper Tim as clearly indicated he thought adding the editorial was useful, and that is not questioned, but however good intent doesn't stop it being POV and it certain seems to be. I still think it therefore has no place and should be removed, or maybe go Richard's way and remove the whole thing? Kiwikiped (talk) 18:01, 18 March 2013 (UTC)
Yes, point taken. My concern was merely to signpost the fact that the assertion, even though it appeared in a peer-reviewed journal, was opinion, unsupported by referenced facts. However, the text in this WP article already indicated that it was an assertion by an individual author, and the added qualification of that statement was probably unnecessary. However, the scope of the source article and thus the context of the assertion does need to be made clear, given that this WP article is about Australia, and the referenced source article is about New Zealand, which has quite different accident compensation arrangements than the various States and Territories in Australia. As such, I also agree that the whole sentence could be removed without detracting too much from this article (perhaps move it to the page on bicycle helmets in New Zealand, if there is one?). Tim C (talk) 20:32, 18 March 2013 (UTC)
Comment Suit yourselves; the topic is pure ho-hum to me, but I think it is nuts to remove every reference that someone somewhere might argue that we might take exception to. We should do the reader the courtesy to credit him/her with a little good sense, and supply the information, the citation, and the context if that is all there is available. If the citation is patent garbage, and patently without merit or authority, of course we would chop it, but if there is reason to pay it some serious attention in the absence of rival claims, equally certainly it should be considered seriously. Perhaps future harder support or refutations should materialise, fine, but till then... JonRichfield (talk) 17:48, 24 March 2013 (UTC)
Perhaps a problem here is that the website appears to be biased towards helmets, and they don't seem to give both sides to the debate. That is what I think anyway. Numbermaniac - T- C00:32, 2 April 2013 (UTC)
The de Jong and Newbold models.
User:Colin at cycling added mention of the de Jong health benefit model of mandatory helmet laws into the section in this article titled "Health implications of bicycle helmets legislation". I removed this mention and reference because de Jong, although he is an Australian researcher, specifically did not evaluate the Australian case in his paper (he only looked at North American and European countries), and also his work and related work is aleady covered in the main WP article on bicycle helmets Bicycle helmet. The mention of de Jong was restored by User:Colin at cycling, and so I made the description of de Jong's work slightly more precise and added mention of the Newboald model, which was published in the same issue of the same journal (Risk Analysis), on the very next page to the de Jong paper. In fact, the Newbold paper was a constructive critique of the de Jong model - which came to a slightly different conclusion than de Jong with respect to the health benefit of mandatory helmet laws in the US case. Thus, the de Jong paper should not be mentioned in a WP article without also mentioning the Newbold paper too. My view is the neither paper need be discussed in this article on helmets in Australia, for the reasons set out above. But if they are discussed here, then BOTH must be mentioned in order to maintain a NPOV. Tim C (talk) 23:46, 28 March 2013 (UTC)
I'm all for getting rid of both. If one is included then the other should be as well but this page has become an out of control shambles and we need less information not more. Everything on this page should be directly relevant to bike helmets in Australia and anything that isn't has no place on this page.Dsnmi (talk) 04:34, 29 March 2013 (UTC)
Agreed. I have removed references to both and substituted a hatnote pointing to the discussion of these mathematical models in the main WP article on bicycle helmets. Tim C (talk) 05:18, 29 March 2013 (UTC)
RfC: Can the Bicycle Helmet Research Foundation web site be considered a reliable authority and source of references for this article
Seeking to remove these references seems a bit strange, but that doesn't mean doing so is wrong as I am no authority. This article seems to be a fair attempt to present a topic in a NPOV in which there are very strongly held opposing POVs. In doing so it references material from sources from both camps, it surely needs to do this to back up the material presented. For example it references material from Governments which passed the laws, which unsurprisingly supports the laws, as well as referencing material from those who oppose the laws; that would seem to be the balanced approach. Lobby groups; be they for the protection of animals, or against some law or other, etc.; tend to be focussed by nature, to reject references to them because of this inherent nature seems unbalanced. In an article discussing, say, battery farming, I'd expect to see references to material by both those supporting such farming and those opposing it - to include either one to the exclusion of the other would be a non-NPOV article surely? Kiwikiped (talk) 15:43, 17 March 2013 (UTC)
Complete removal of the references to the BHRF web site is not proposed, although some pruning of them (and the assertions in the article that they are used to support) may be reasonable, given the heavy reliance on that web site (13 references at last count). However, flagging of the references to the BHRF website as a potentially unreliable source may be justified and in accord with WP policies. Please see my rationale for this on the Reliable Sources noticeboard. Tim C (talk) 20:09, 17 March 2013 (UTC)
We will see what emerges from the discussion on the RS noticeboard. Meanwhile I continue to opine that we have achieved an article with many virtues, but which is practically unreadable. I'll make proposals, or a bold edit, when I have time. I hope for a concise narrative account of the main assertions and arguments, graphical presentations of the main data series, far fewer figures especially percentages, and keeping most or all the current references. Richard Keatinge (talk) 20:42, 17 March 2013 (UTC)
It has been more than a week since it was noted that citations are needed for the (2) graphs in the 'Bicycle usage: capital cities vs all other areas' section. Dorre has made 11 contributions since then, but not added the requested citations. The whole section should be removed.--Linda.m.ward (talk) 02:50, 19 March 2013 (UTC)
Reliable sources are usually required to have "editorial oversight". Just glancing through their editorial board members shows that it is about 50:50 split between experts (university professors, researchers or consultants in the field and the like) and laypersons (activists, GPs etc). I'm not overly familiar with the finer points of determining reliability, but given their obvious POV, it is a bit of a red flag. That said, the data they quote seems to be legit, and their list of sources is immense. Perhaps it would do, where possible, to quote the source material direct rather than the BHRF, since most of the sources seem to be NPOV. Certainly better than citing an interest group 13 times. --Yeti Hunter (talk) 11:12, 21 March 2013 (UTC) Edit: I came here from your request at the sciences RFC noticeboard
This article currently has 126 references, of which 6 cite BHRF pages.
Reference 30 is that "while in the NT since March 1994 there is an exemption for adults cycling along footpaths or on cycle path."
Reference 41 is an article by D L Robinson, author of several peer-reviewed journal articles on bicycle helmet laws.
References 42 and 44 are cited using the text: "Several précis of and commentaries on these surveys have appeared on websites and blogs.[42][43][44][45][46]"
Reference 81 is an article with tables and graphs of Australian Bureau of Stats census data.
Reference 101 is cited as evidence that "Nonetheless, this is still low by international standards at 0.4 trips per bike per day in July and 0.8 trips per bike per day in January,[102]" The BHRF webpage provides a convenient summary of information on usage rates.
I would suggest that the BHRF it is probably one of the most reliable authoritive sources for the information cited above. Their editorial board contains many prominent cyclists, e.g. John Franklin, author of 'Cyclecraft' - http://en.wikipedia.org/wiki/Cyclecraft and many others with links to cyclists organisations such as the CTC and the ECF. The BHRF provides links to original papers and sources, where available on the web, so in order to maintain its reputation, has to be very careful about ensuring the information provides is as accurate as possible.
Readers of the Bicycle Helmets in Australia talk page will know that I personally used the census data (available on the BHRF website) on numbers cycling to work, and total numbers of people who travelled to work on census day by state to calculate the percentage of single-mode journeys by bike for a) individual states, b) capital cities and regional areas according to whether there was an enforced helmet law at the time of the census. Tim objected to these graphs because I put vertical lines to indicate the separation between data points with and without enforced laws. To avoid any interpretation of the vertical lines as dates of legislation, I therefore replaced the vertical lines with dotted lines. Tim then objected because he considered drawing graphs from published data was 'original research', so I asked the BHRF to include them on their page of census data. The data cover censuses in 1976, 81, 86, 91, 96, 01, 06, and 11 in 8 different states & territories - Vic, NSW, Tas, SA, WA, ACT, Qld and the NT - so 64 individual documents or downloads were needed to compile the data. I downloaded them all and found only one discrepancy compared to the previoulsy-published data on the BHRF website. I contacted the ABS and they confirmed it was their mistake - the scan for one state included a page on cycling to work for another state. The document on the ABS website has now been corrected.
Perhaps Tim can explain why he thinks there is any inaccuracy or unreliability in the information presented?
Even peer-reviewed literature contains inaccuracies and silly mistakes. For example, one of the early papers [Cameron, MH; Vulcan AP; Finch CF; Newstead SV (June 1994). "Mandatory bicycle helmet use following a decade of helmet promotion in Victoria, Australia--an evaluation.". Accident Analysis & Prevention 6 (3): 325–337.] reports that Melbourne, with a population of 3 million, averaged about 60 million hours of cycling per week (see Fig 7) - an average of 20 hours per week for every many, woman and child in the city! As far as I can determine, the journal has never published a correction to this clearly ludicrous information. If Tim have ever seen equally incorrect on the BHRF website, I expect he would have complained about it immediately.
There is an interesting web page http://www.cbdbug.org.au/2011/06/helmet-research-paper-released/ that provides documents obtained under right to information legislation illustrating the process of government oversight into the commissioning of bicycle helmet research in Australia. Some questions remain about earlier research. For example, a study of the first 3 years of legislation in Victoria found that head injury rates were no different to pre-law trends. The following year, the researchers used a new model to show a decline in the number of head injuries, while admitting that this models cannot distinguish between reductions in head injury because of reduced cycling and reductions because of increased helmet wearing. Research reports provide details of numbers of both adult and child cyclists in 1990, 1991 and 1992 at the same 64 sites and observation periods used to estimate the change in children's cycle use. Indeed, for children, the changes in cycle use are pretty similar to the changes in numbers counted. Yet, instead of using the numbers of adults counted (or estimating cycle use from the highly correlated measure of numbers counted) the published paper by Cameron et al. (1994) claimed that "adults were not included in the 1990 surveys". This sort of inaccuracy can escape peer review - the original research reports were not readily available at the time - so the reviewers would have no reason to know that adults had, indeed, been counted in the 1990 surveys.
Tim is clearly in favour of helmet laws. He was co-author of a paper that found a small reduction in the ratio of head to arm injuries, while noting that "the contribution of factors such as risk compensation and safety in numbers has not been incorporated in this study." Despite not knowing whether there was an increase in injury rates per km cycled, or any discussion of whether the losses from head injury that could be prevented by helmet wearing compare with the lost health and environmental benefits of cycling, the paper nonetheless concluded '...repealing the law cannot be justified." Personally, I'm at a loss to understand some of his edits, for example he "Corrected reference to Hoye and Elvik (not Erke and Elvik as stated)" I clicked on the link to the reference, a pdf file which lists the authors as Alena Erke and Rune Elvik on the front page.
So in summary, I believe the BHRF is a reliable source, that is needed to balance the research funded and sponsored by Australian government agencies. A NPOV cannot be achieved without presenting both sides of the argument. The 6 citations to BHRF pages improve this article. I cannot see why anyone would want to argue against the reproduction of graphs produced from data that can be downloaded and verified from the ABS website. Their removal would represent a considerable loss, given that census data on cycling to work is one of the most longest reliable and consistent data series that covers a 35 year period that can illustrate long-term trends before and after helmet laws were introduced. Dorre (talk) 01:48, 23 March 2013 (UTC)
Content: Material form this organisation's web site is referenced some 13 times (as at 16 March 2013). All these references have been edited (by me) to clearly identify the source, and thus a search the article or article source for "Bicycle Helmet Research Foundation" will show the references.
The question I am seeking opinions on is whether the pages on this organisation's web site can be considered a reliable and appropriate source for the article in question?
The organisation appears to be a company registered in Anglesey, Wales, UK, to a Dr Richard Keatinge, who is the sole director listed.
The BHRF policy statement says "cyclehelmets.org is administered by the Bicycle Helmet Research Foundation (BHRF), an incorporated body with an international membership, to undertake, encourage, and spread the scientific study of the use of bicycle helmets.The purpose of cyclehelmets.org is to provide a resource of best-available factual information to assist the understanding of a complex subject, and one where some of the reasoning may conflict with received opinion. In particular we seek to provide access to a wider range of information than is commonly made available by some governments and other bodies that take a strong helmet promotion stance. It is hoped that this will assist informed judgements about the pros and cons of cycle helmets."
The BHRF site claims that its material is peer-reviewed but no details are given on the process. The organisation does list its and Editorial Board. Of concern is that three of the members of the Editorial Board, which it says is responsible for the content of the web site, themselves operate or are associated with organisations and/or web sites which appear to be actively lobbying against mandatory cycle helmet laws, or promote such lobbying. This is potentially problematic because a great deal of the contention in the Bicycle helmets in Australia article is about mandatory helmet laws.
In addition, I have been unable to find any statements, comments or publications attributable to any of the other members of the BHRF Editorial Board which are not critical of bicycle helmet laws (and in some cases, of bicycle helmets themselves). If the BHRF promoted a balanced view of bicycle helmets and helmet laws, I would have expected at least some members of its Editorial Board to have publicly voiced opinions favourable to bicycle helmets or bicycle helmet promotion campaigns and laws. Of course, my search was necessarily not exhaustive, and I may have overlooked such material somewhere on the Internet.
Just to make it clear, complete removal of the references to the BHRF web site is not proposed, although some pruning of some of those references (and the assertions in the article that they are used to support) may be reasonable, given the heavy reliance on that web site (13 references to it in the article at last count). However, flagging of the references to the BHRF website as a potentially unreliable source may be justified and in accord with WP policies.
The other thing worth pointing out is that the vast majority of the pages on the BHRF site seem to be anonymously authored - I can't find any attribution to any named individuals except for a few pages (out of many on the site) attributed to Dorothy Robinson. Their Editorial Boardpage states that "The Editorial Board is responsible for the content of cyclehelmets.org" but that's not the same as putting by-lines on pages. I have never seen a peer-reviewed publication (as they claim to be) that doesn't attribute articles to the individuals who authored them.
The nub of the problem is that the BHRF site is about a single subject. The Bicycle Helmet Research Foundation would have a lot more credibility if it were an organisation concerned with all aspects of cycling safety, or an organisation interested in the biomechanical and engineering aspects of all types of helmets. But it is only about bicycle helmets, and furthermore, it seems to be exclusively concerned with what it holds to be the negative aspects of bicycle helmets. Yes, the site does have a page titled "Published evidence supportive of helmet effectiveness or promotion", but nearly every study listed on that page links to a sub-page on cyclehelmets.org which presents a critique of the study which concludes that the study is invalid or fatally flawed. All studies have flaws, but the critiques appear to be relentlessly negative. There's also a page titled "Published evidence sceptical of helmet effectiveness or promotion", but the links on that page don't lead to critiques of these papers, they just point to full-text copies of these papers or to their PubMed records. Thus there would appear to be an imbalance here.
Of even greater concern is that some of the critiques on the BHRF cyclehelmets.org site appear to attack and impugn the professional integrity of researchers. For example, the site has this to say about one particular study: "Despite attempts to manipulate the results, one of the largest reviews of the evidence has not been able to find any reliable evidence that helmets have benefited cyclists". Further accusations of dishonest and unethical behaviour are made against the authors of that study on this page (see the section labelled "Ethics"). Those are not the sort of words one expects to find in a balanced, unbiased and professional publication.
Anonymity of the articles on the BHRF site is one of the main problems I have with it being considered a reliable source. No other peer-reviewed publication that I am aware of permits anonymous articles. The issue is one of potential undue weight. Articles published in peer-reviewed journals by Bill Curnow and Dorothy Robinson, are already cited 4 and 22 times respectively in the WP article in question. Both Bill Curnow and Dorothy Robinson are listed as members of the BHRF editorial board. BHRF web pages are referenced 13 times in the article - 3 references to a BHRF page attributed to Dorothy Robinson, and 10 references to anonymously-authored BHRF pages. It is important for WP readers to be able to know who is responsible for those anonymous web pages, which are used as authorities for statements in the article - particularly when three members of the BHRF editorial board, including Curnow, appear to be running lobby organisations and/or campaigns opposing mandatory bicycle helmet laws, as noted above.
Here is another example of the problems posed by the anonymous and undated (and unversioned) nature of most of the articles/pages on the BHRF web site: two graphs of Australian Census data were added some time ago to the Bicycle usage section of the article by a pseudonymous WP editor who has contributed a great deal of material and made many edits to the article. Although the graphs are based on Australian Bureau of Statistics Census data, they combine and present that data in a particular manner and are thus a synthesis, or possibly original research, and as such they are required to be referenced to an external source. The source files for these graphs are in WikiMedia Commons and are listed there as that pseudonymous editor's own work. I added citation needed tags to these graphs, and the editor was asked via the Talk page for the article to supply suitable external references for the graphs. This was done: the references given are to an anonymous and undated BHRF web page, where copies of these graphs now appear. Thus it would appear that there is some connection between the pseudonymous user in question and the BHRF organisation. This is potentially problematic given that most of the BHRF web pages are anonymously authored, including the one given as the external reference for these two graphs. Thus a circular chain of anonymous authorship has been established. Pseudonymity of WP editors is permitted (and in fact sacrosanct), but anonymity of authorship of external references is just not acceptable in my opinion. Someone has to put their name to and take personal responsibility for web pages on the cyclehelmets.org site if they are to be used as authority for assertions and statements (and sources of graphs) in this and other WP articles. Tim C (talk) 02:22, 23 March 2013 (UTC)
I asked for evidence of inaccuracy or unreliability in the information presented - like, for example, counting 29% fewer adult cyclists at the same sites and observation periods in 1991, compared to 1990, then claiming "adults were not included in the 1990 surveys" and "an estimated increase in adult (cycle) use of 44%", or that "In Melbourne adult cyclist numbers doubled after the helmet legislation was introduced".
As you know, these comments refer to surveys (at the same 64 sites and observation times and time of year) in which 29% fewer adult cyclists were counted in the post-law survey than the pre-law survey. I can see why an organisation that used these results to claim that "adult cyclist numbers doubled after the helmet legislation was introduced" might be considered an unreliable source. Your comments sound like an ad hominem argument - you can't find any inaccuracies with the cited information so instead attack the organisation. If I were a cynic, I'd point out that if you manage to have the BHRF labelled an "unreliable source" maybe people won't see the references to the organisations that claim numbers of adult cyclists doubled, when the surveys actually counted 29% fewer cyclists?
My suggestion is that references to anonymously-authored and undated pages on the BHRF web site be tagged with [unreliable source?] due to doubts about the impartiality of the Editorial Board of the organisation, for the reasons set out above, and the inability to assess the expertise and good-standing of the author(s) of such pages if they are not identified. Tim C (talk) 04:37, 23 March 2013 (UTC)
Due to the nature and tone of this particular debate, certain experts have well-founded worries which only anonymity will allay. In the circumstances I don't think that affects reliability for our purposes. Richard Keatinge (talk) 13:55, 23 March 2013 (UTC)
User:Dorre muses (above): "If Tim have ever seen equally incorrect on the BHRF website, I expect he would have complained about it immediately." Well, I can say that your expectation is incorrect - I am not in the business of correcting the BHRF site - that is the job of the site owner and its Editorial Board. My only concern is whether the BHRF web site is a suitably reliable reference for this and related WP articles. But since you ask, here is something that the BRHF Editorial Board might wish to correct: on this page, research by Voukelatos and Rissel is described: http://www.cyclehelmets.org/1207.html?NKey=22 - with a footnoting noting that the study contained serious errors, but failing to mentioned that the study was formally and completely retracted by the journal which published it. The study is also listed (with broken web links to the paper) here: http://www.cyclehelmets.org/1160.html and here: http://www.cyclehelmets.org/1146.html - in both cases without any mention that the paper contained serious errors and had been retracted (indeed, citation of papers after they have been formally retracted is a highly unusually practice in peer-reviewed scientific discourse). Yet these pages on the cyclehelmets.org site are clearly maintained, because references to much more recent studies have been added to them. Tim C (talk) 04:37, 23 March 2013 (UTC)
You have a useful point in there and I hope that a much fuller account of this paper and its aftermath will soon be on the site. (We'd appreciate any other corrections and updates you or indeed anyone may be able to offer.) We should however, despite your respected opinions above, wait for a clear policy-based consensus on the subject of reliability. Richard Keatinge (talk) 13:55, 23 March 2013 (UTC)
Here is another example of the selectivity of the information offered on the BHRF cyclehelmets.org site: there is a link titled "Cutting through the controversy about helmet effectiveness" at the very top of the cyclehelmets.org home page, leading to this page: http://www.cyclehelmets.org/1249.html which is about the de Jong cost-benefit model of hypothetical all-age mandatory helmet laws. The reference for this study is: de Jong, Piet (May 2012). "The Health Impact of Mandatory Bicycle Helmet Laws". Risk Analysis 32 (5): pp782–790. Fine, but there is no mention whatsoever of the response to the de Jong model by Newbold, which was published on the very next page of the same issue of the same journal (Newbold, Stephen C. (May 2012). "Examining the Health-Risk Tradeoffs of Mandatory Bicycle Helmet Laws". Risk Analysis 32 (5): pp791–798.). Newbold critically examined the de Jong model, and far from dismissing it, he suggested various improvements to it and identified parameters in it which require further research in order to quantify accurately (such as exercise type substitution behaviour). Newbold used published US data in his refined version of the de Jong model and found that mandatory bicycle helmet laws would seem to have positive net public health benefits in the US, although he warned that results from both his model and that of de Jong needed to be treated as very provisional due to uncertainty around values of key parameters to them. In fact, there doesn't seem to be any reference to Newbold's paper at all on the BHRF cyclehelmets.org web site. It is this sort of selectivity in the evidence presented that calls the reliability of the BHRF as a source for WP articles into question. Tim C (talk) 01:31, 24 March 2013 (UTC)
Tim Churches has previously pointed out that the studies that appear on the BHRF web site as supportive of helmet effectiveness or promotion are accompanied by critiques claiming that the study is invalid or fatally flawed, but that the studies listed as being sceptical of helmet effectiveness of promotion are not accompanied by negative critiques. I too had noticed this lack of balance when I first stumbled on the site a couple of years ago, and have consequently not spent a lot of time on the site, but have encountered much evidence of inaccuracy or unreliability in the information presented.
The 'Published evidence sceptical of helmet effectiveness or promotion' page (http://www.cyclehelmets.org/1146.html), contains 2 articles that most certainly do NOT 'cast doubt on the effectiveness of cycle helmets in reducing head injuries (at least serious head injuries), or on the wisdom of helmet promotion or laws', and one article (critical of the helmet law) that was retracted more than a year ago.
With respect to an article by Elvik (Accident Analysis & Prevention, 2011;43(3):1245-1251. 2011), the page states that 'Attewell et al influenced by publication and time-trend biases. When controlled for, the protective effects of helmets are smaller. Adding new studies, no overall benefit of helmets found.' Elvik noted that 'the re-analysis shows smaller safety benefits associated with the use of bicycle helmets than the original study', and reported the following odds ratios for fatal, brain and head injuries
- fatal injury 0.27 in Attewell's original study, 0.23 in Elvik's re-analysis
- brain injury 0.42 in the original study, 0.47 in the re-analysis
- head injury 0.40 in the original study, 0.58 in the re-analysis
Re head injuries, 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.'
As already noted by Tim Churches, an article by Voukelatos and Rissel (Australasian College of Road Safety, ACRS 2010;21(3):50-55. 2010.) that was retracted in early 2011 is still listed on the page. The page described the study that was retracted 2 years ago as showing that 'Helmets were not the main reason for the drop in head injuries in Australia since helmet laws were introduced. General improvement in road safety from random breath testing and other measures were probably the cause.'. Many studies published after the retraction have been added to the page (7 in 2012 alone), yet this study has still not been removed from the page.
In describing an AIHW article (Serious injury due to land transport accidents, Australia, 2003-04), the page states that 'Wearing a helmet seems to have no discernible impact on the risk of head injury'. The AIHW report does not even mention bicycle helmets.
I have scanned only a few other pages on the BHRF site, and would not consider any of them to be reliable . . .
'Changes in cycle use in Australia' (http://www.cyclehelmets.org/1194.html) cites an SA study as showing a reduction in cycling to school, notes that the study 'likely' under-estimated that actual decline, and fails to note that the decline in cycling to school in the SA study was very similar to the declines in cycling to school in NSW and Victoria. The article also fails to note that the SA study found that there was no (statistically significant) evidence of a reduction in adult cycling, that prior to the helmet law cycling to school comprised about 20% of cycling in that age group, and that after the helmet law there was an increase in cycling to/around other venues of similar magnitude to the drop in cycling to school.
'Helmet laws: Northern Territory' (http://www.cyclehelmets.org/1114.html) claims that 'Helmet use is now low in the NT . . . Estimates in 2004 suggested that 15% - 20% of cyclists continue to wear helmets, mostly "serious" cyclists'. I live in Darwin and have been cycling almost every day (commuter and utility) for about 10 years. The helmet wearing rate is much higher than the 15-20% alleged, probably at least 80%, ie. it the number of cyclists NOT wearing helmets would be at most 15%-20%.
'Head Injuries and Helmet Laws in Australia and New Zealand' (http://www.cyclehelmets.org/1241.html) contains numerous examples of 'overlooking' evidence that runs counter to the claim that the helmet law was/is a bad thing. With respect to the most obvious examples, the article
Claims that the Melbourne surveys were done 'in similar weather'. The authors of the study in which the data was collected noted differences in weather conditions, and concluded that after taking those differences into account, there was no reduction in adult cycling as a result of the helmet law.
Notes downward trends in injuries for all road users; that in Victoria, head injuries fell almost as much as non-head injuries; and suggests that the reductions were due to 'large reductions' in cycling. The article fails to note that the injury data in the Victorian study showed that pedestrian head injuries dropped by about 20%, cyclist non-head injuries dropped by about 25%, and serious and severe cyclist head/brain injuries dropped by 40%.
Cites injury data from an SA study, yet fails to note that the SA study also found that there was no decrease in overall cycling as a result of the helmet law.
Claims that there was 'no obvious effect' of the law in SA with respect to reduced cyclist head injuries, yet fails to note that the injury data in the study also showed that
cyclist concussion admissions dropped by 54%, and that admissions for all other causes of concussion dropped by 27%
preventable injuries other than concussion reduced by 41%
This sort of debate is as I've argued before not really suitable for an encyclopedia; on a deeply contested subject we should not rehash minutiae but should give an overview of the main points of the argument. It's also not really relevant to the subject of reliability. You may disagree with everything a source says, but that is not the point at issue. If we leave out every source that makes a contestable comment we will be left with a remarkably short article. Richard Keatinge (talk) 15:52, 24 March 2013 (UTC)
With respect, User:Dorre, who has stated on Wikipedia:Reliable_sources/Noticeboard#Bicycle_Helmet_Research_Foundation that s/he is a member of the BHRF cyclehelmets.org Editorial Board, specifically asked (above) for "evidence of inaccuracy or unreliability in the information presented", and that is exactly what User:Linda.m.ward has provided. Debating on Talk pages the accuracy and reliability of sources referred to in WP articles is entirely appropriate, BTW. Such debate should not appear in the article itself, of course. Tim C (talk) 21:30, 24 March 2013 (UTC)
I could say that the 1992 MUARC study was conducted in better weather than the pre-law survey (the 1991 survey, was slightly worse, but on average the two post-law surveys had similar weather to the pre-law survey. Yet, despite the better weather, and had a bicycle rally passing through one of the sites (451 cyclists in 1992, 72 in 1991), they sill managed to count 83 fewer adult cyclists in 1992 than 1990. But I'd be falling into the trap of arguing about minute details, rather than allowing the data and the graphs to paint the broader picture.
As they say, a picture is worth 1,000 words. If we want to illustrate the effect of helmet laws we have to show the data, as honestly and reliably as possible, in a way that people can understand, so that they can see the effect of the helmet law compared to other effects such as safer roads and pre-law trends. Dorre (talk) 22:22, 24 March 2013 (UTC)
Indeed, if we are to achieve a good article then good graphical summaries of the main data series are essential. Per WP:SYNTHNOT this is not a problem, indeed it's what Wikipedia editors are supposed to do. Richard Keatinge (talk) 10:25, 3 April 2013 (UTC)
It is wrong to attempt to dismiss a particular information source on the basis of an alleged bias, rather than on the basis on the accuracy of the information itself. For example, CRAG openly states where it stands. Not everybody agrees with that position, however some of the information provided on the site is still useful. For example, this page summarising surveys of cycling after the helmet law. (http://crag.asn.au/?p=174). This is informative, described in the neutral tone, and properly referenced. To dismiss it on the basis of not agreeing with CRAG's position seems biased.
The same applies to BHRF. Attempts being made to dismiss material from BHRF on the basis of not agreeing with the position of some of its members seems biased. This discussion should be based on facts, not on disagreeing with some members of a group.
Linda Ward long list of statements she disagree with on BHRF is interesting. However, it is not proper to attempt to denigrate it without providing BHRF the opportunity to reply. It would be more appropriate to contact BHRF directly and let them know of statements believed to be incorrect on their site, providing supporting evidence. For example, claiming that the helmet wearing rate in the northern territory is 80% based on personal observations is no more generalisable that somebody else perception. A survey is needed to support such statements.
It is odd that an avid helmet advocate attempts to dismiss references from CRAG on the basis that CRAG is open & honest about their position, and now attempts to do the same about BHRF. Particularly while quoting studies without disclosing relevant conflict of interests nor disclosing that they have been conducted by helmet advocates. This is deceitful. These studies could be critiqued in a similar manner as the way BHRF material is being dismissed.
Effectively, adopting the approach suggested means punishing honesty (ie. discounting information from sources that have the honesty to state their position), and rewarding deceit (to present as neutral information from sources that attempt to conceal their position, or to pretend to have a neutral position while being helmet advocates).
I do not claim to speak on behalf of Wikipedia, but it seems dangerous for Wikipedia to tolerate such an approach. Rewarding misleading & deceptive conduct can only encourage it, leading Wikipedia to becoming a platform hijacked by well-organized, well-funded deceitful special interest groups.
The key is to disclose conflicts of interest when relevant. For example, it would be inappropriate to dismiss the summary of survey on the CRAG site based on CRAG stated position. The same goes for BHRF. If believe that the information posted is inaccurate, then explain why based on facts. If you are aware of an undisclosed conflict of interest, then state it (rather than try to censor the information), so that people can make a fully informed judgement. Dismissing information on the basis that you don't agree with the source stated position is a form of bias.
Arguments must be based on facts, not on disagreement with parties stated position. If you disagree with a statement referenced by BHRF, then argue based on facts, rather try to dismiss it based on the source being BHRF.Harvey4931 (talk) 21:28, 5 April 2013 (UTC)
^Walker M. Bicycling in Sydney: law compliance and attitudes to road safety. Velo Australis; 1996; Fremantle, Western Australia.