#21
|
|||
|
|||
Safety in Numbers
On Sep 9, 12:03*am, "Adam Lea" wrote:
But if the Gaussian model is wrong...... If the authors are using statistical tests that assume normality and the data is not normally distributed then it should be picked up in the peer review. *Rolls around laughing* *picks himself up off the floor* *falls over laughing again* The peer review process is far from perfect. Verification of stats is getting better but there are still a lot of crap studies getting through. ...d |
Ads |
#22
|
|||
|
|||
Safety in Numbers
On Sep 9, 9:49*am, Toom Tabard wrote:
When cycling safely and considerately I constantly had near-misses which could easily have had serious consequences. That rarely if ever happens when walking. I have never heard of a near miss having serious consequences. A hit, yes, but a miss, no. Many people over estimate the severity of a near miss (close pass). It is unpleasant and annoying, but is rarely dangerous. Obviously the margin for error is reduced, increasing the overall number of not quite near misses, so it is worthwhile encouraging vehicles to leave more space. ...d |
#23
|
|||
|
|||
Safety in Numbers
On Tue, 9 Sep 2008 01:49:41 -0700 (PDT), Toom Tabard wrote:
On 8 Sep, 12:27, "wafflycat" wrote: Without fail, the number one reason folk give to me as to why they won't cycle on roads is that it is "unsafe". Of course, those of us who do cycle know that in the great scheme of things it is no more dangerous than walking and a zillion other activities deemed as safe to do- Hide quoted text - I walk in town, a lot more than and a lot further than the average person. I no longer cycle in town, because I find it is unsafe, and it certainly many times more hazardous than walking. When walking, you only mix with traffic when crossing, and can freely choose, monitor and control how you do that. When you are cycling and permanently interacting with traffic you have no such control. When cycling safely and considerately I constantly had near-misses which could easily have had serious consequences. That rarely if ever happens when walking. Perception is not reality. Anecdote is not proof. Cycling is not more dangerous than walking. |
#24
|
|||
|
|||
Safety in Numbers
On 8 Sep, 13:18, judith wrote:
The number of seriously injured cyclists being treated in London hospitals is soaring, official figures have revealed. They show twice as many riders are being admitted for treatment as six years ago. The biggest rises were in Kensington and Chelsea, Wandsworth, Camden, Kingston and Havering. Cases of cyclists admitted to hospital jumped from 422 in 2000/01 to 819 in 2006/07. For children under 16, the number rose from 145 to 184, and for under-11s it went from 70 to 75. In Kensington and Chelsea, cycling casualties increased from 38 to 69; in Wandsworth from 43 to 84; in Kingston from 23 to 55; in Havering from 11 to 48; and in Camden from 22 to 53. Looks like cycling is dangerous to me. So casualties have doubled. If the number of cycle-miles has trebled (for example), then cycling has actually got safer for the individual cyclist (which is what I want). If the roads become so dangerous that the only cyclist left in London is Boris, and by some miracle he avoids being hit, then there will be zero cyclists injured. That does not equate to cycling being safe. Peter. |
#25
|
|||
|
|||
Safety in Numbers
On Tue, 9 Sep 2008 04:45:30 -0700 (PDT), al Mossah
said in : If the roads become so dangerous that the only cyclist left in London is Boris, and by some miracle he avoids being hit, then there will be zero cyclists injured. That does not equate to cycling being safe. One of the fundamental problems with the way road "safety" is managed is that if a road becomes so frightening that no pedestrian or cyclist will go near it, the statistics show it as being unusually "safe". Conversely, unthreatening roads are counted as dangerous because they attract more of the most vulnerable. Guy -- May contain traces of irony. Contents liable to settle after posting. http://www.chapmancentral.co.uk 85% of helmet statistics are made up, 69% of them at CHS, Puget Sound |
#26
|
|||
|
|||
Safety in Numbers
Toom Tabard wrote:
I walk in town, a lot more than and a lot further than the average person. I no longer cycle in town, because I find it is unsafe, and it certainly many times more hazardous than walking. So if it's "certainly" more hazardous than walking, why doesn't that show up in the DfT's stats? When walking, you only mix with traffic when crossing, and can freely choose, monitor and control how you do that. Up to a point: when you get to the place you have to cross then you have to cross. The fact that you can wait is a bit moot if the gaps aren't there, and you have much less influence on the traffic than if you are a part of it, as you are on a bike. When you are cycling and permanently interacting with traffic you have no such control. You do, OTOH, have a lot more control over the traffic. When cycling safely and considerately I constantly had near-misses which could easily have had serious consequences. That rarely if ever happens when walking. So how safely were you actually cycling? You "constantly" had near misses, yet they all missed, so how near were they really? I rarely have "near misses": might be my cycling, might be the local driving, might be my perception... ISTM you are possibly confusing actual safety with perceptions of safety, and possibly perceptions of how safely you were cycling. I see quite a few people adding to their chances of danger when cycling "sensibly", for example hugging the kerb. Pete. -- Peter Clinch Medical Physics IT Officer Tel 44 1382 660111 ext. 33637 Univ. of Dundee, Ninewells Hospital Fax 44 1382 640177 Dundee DD1 9SY Scotland UK net http://www.dundee.ac.uk/~pjclinch/ |
#27
|
|||
|
|||
Safety in Numbers
In message
"Adam Lea" wrote: "Mike Clark" wrote in message .uk... Tell me what you mean by "significantly reliable prediction" and perhaps it would be possible to come up with an answer to the question. Significance as used in academic publications often assumes that the parameters of the model follow a Gaussian distribution. Then the probability that the observed results fall within the expected limits of the predicted Gaussian are quoted (eg 95% confidence limits, or 99% confidence limits etc). However all that tells you is that if the Gaussian model is accurate then you might have got similar results by chance 1 in 20 times or 1 in 100 times respectively. But if the Gaussian model is wrong...... If the authors are using statistical tests that assume normality and the data is not normally distributed then it should be picked up in the peer review. You'd like to think that peer review would always pick this up but sadly experience has taught me that such things can often slip through the net. If your peers have a similar view to your own about how to model the data they may not be entirely independent in their reviewing of your work, or put another way, ideas can be fashionable and unfashionable amongst academics, but that doesn't necessarily make them right or wrong. I'd recommend you to read "The Black Swan" by Nassim Nicholas Taleb. He has in part a background as an active trader on the stockmarkets and thus singles out economics, and in particular econometrics for special attention. Within the book he has several diatribes about the awards of the Nobel prizes for Economics for mathematical modelling that completely fails to take into account the visible evidence of the occurrence and dominating effects of "Black Swan" (unforeseen) events within the stock markets. Indeed I was reading in the Times today that the failure of the computer systems on the London Stock Exchange yesterday, when share trading hit high levels, was a "Black Swan". Really? Computer failure at a crucial time considered to be outside of normal expectations? But returning to your basic point about picking up whether the data fits to a normal distribution, the problem could be that even if the data collected so far is consistent with a normal distribution, that doesn't mean that the data collected in the future will be consistent with a normal distribution. There are other distributions that appear normal for much of the sampling. In other words whilst it is possible to disprove a normal distribution, it is much harder to prove a normal distribution for the data. So returning to the specific point about "safety in numbers", even if it were shown, that in previous examples where there was an increase in cyclists and cycling, that this was accompanied by a decrease in the rate of accidents, this doesn't mean that if we increased the numbers of cyclists in a specific location in the future that this will necessarily be accompanied by a similar drop in the accident rate. We might hope it does, but are we so confident in our models that we couldn't conceive of getting an unexpected result? Mike -- M.R. Clark PhD, Reader in Therapeutic and Molecular Immunology Cambridge University, Department of Pathology Tennis Court Road, Cambridge CB2 1QP Tel +44 (0)1223 333705 Web http://www.path.cam.ac.uk/~mrc7/ |
#28
|
|||
|
|||
Safety in Numbers
william lawson wrote:
I quite agree - I also think that many of the "problems" are made worse by cyclists riding through red traffic signals - switching from car lane to car lane, riding on pavements and not using cycle lanes. Cyclists riding through reds only have themselves to blame if they end up flatter than they usually like. But switching lanes... bikes are legally defined as vehicles, and those are *vehicle* lanes, not "car lanes". Cycle lanes are not all created equal: some are good, many impede cyclists and add to their danger, so it's pretty daft to use them all. I now walk to college as although it takes longer - it is safer How do you know? Why don't the DfT's figures for relative safety show that walking is typically safer? Pete. -- Peter Clinch Medical Physics IT Officer Tel 44 1382 660111 ext. 33637 Univ. of Dundee, Ninewells Hospital Fax 44 1382 640177 Dundee DD1 9SY Scotland UK net http://www.dundee.ac.uk/~pjclinch/ |
#29
|
|||
|
|||
Safety in Numbers
_ wrote:
Cycling is not more dangerous than walking. Given certain measures over certain populations it isn't, but it's a too-sweeping statement not to benefit from further qualification. I would suspect there could well be particular locations where cycling is more dangerous than walking, and certain individuals for whom cycling is more dangerous than walking. Pete. -- Peter Clinch Medical Physics IT Officer Tel 44 1382 660111 ext. 33637 Univ. of Dundee, Ninewells Hospital Fax 44 1382 640177 Dundee DD1 9SY Scotland UK net http://www.dundee.ac.uk/~pjclinch/ |
#30
|
|||
|
|||
Safety in Numbers
On Tue, 09 Sep 2008 10:08:55 +0100 someone who may be william lawson
wrote this:- I quite agree - I also think that many of the "problems" are made worse by cyclists [snip] and not using cycle lanes. Which bit of, "use of cycle lanes is not compulsory and will depend on your experience and skills," do you have difficulty understanding? Highway code, Rule 63 "Cycle Lanes. These are marked by a white line (which may be broken) along the carriageway (see Rule 140). Keep within the lane when practicable. When leaving a cycle lane check before pulling out that it is safe to do so and signal your intention clearly to other road users. Use of cycle lanes is not compulsory and will depend on your experience and skills, but they can make your journey safer." -- David Hansen, Edinburgh I will *always* explain revoked encryption keys, unless RIP prevents me http://www.opsi.gov.uk/acts/acts2000/00023--e.htm#54 |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
What are these BB numbers telling me? | TomYoung | Techniques | 7 | October 16th 06 06:03 PM |
Numbers to think about | CowPunk | Racing | 107 | August 2nd 06 10:48 AM |
Safety in Numbers. | Simon Mason | UK | 11 | April 23rd 05 09:34 PM |
bicycling - safety in numbers | Paul R | Social Issues | 7 | April 20th 05 03:51 PM |