Cycling UK Forum

thirdcrank wrote: I am no statistician, but I know enough about statistics to question some doubtful conclusions and that is all I am doing here.

Against my better judgement I find myself drawn back into this. On p2 I pointed out that the proposition of the OP is entirely fallacious and without some basic statistical information no conclusions or sensible speculations can be made. But it seems some would rather cherish their "Counterknowlege" than test the issue scientifically. You know there are people who believe Diana was murdered or that 9/11 was a US govt plot. I've a good mind to report this thread to Bad Science.

Science may just be the wrong tool to unpick this one, though visual perception, saccades and so on may have some mileage.

thirdcrank

I think we're probably on the same page. I just want to know why it's wrong - I agree that instinct says it probably is, although the DRL business illustrates that instinct can sometimes give you tunnel vision.

I see what you mean about direct routes too. As a student in Leeds about 1980, I rode to Newcastle on direct roads (but not, by then, the A1 - what would I use now?) Having moved down here to Hertfordshire a couple of years later, I rode home to Cheshire and back a few times via (mostly) the A6. When I did it again a decade or so on from then, the road had been "improved" (read "made useless to cyclists"). I used the A5 in the same way several years back to get to Staffordshire, and again "improvements" had made it difficult - but what route are you supposed to use?

OTOH the A1 near here blocked a direct but non-A-road route to the north, but recent improvements did consider cyclists and provide a bridge, reconnecting the former route. Here to Bath in a straight line is an easy route to follow, with few A roads in the way, and so are many other longer journeys - Skegness is fine, except for one five-mile stretch.

And there's always the GPS as a way of making a direct-but-complex route easier to handle without checking navigation on every street corner.

It's not all black

Ben Lovejoy wrote:I share your optimism, Drossall. I see more and more of my friends getting back into cycling.

I think there's a threefold effect:

First, one person in a social group returning to cycling is enough to have them remind the rest what fun it is.

Second, we're most of us of an age where we can afford to indulge ourselves with nice leisure toys.

Third, once people own a bike and start using it for pleasure, they also get reminded of how useful they can be as a form of transport too.

I think the £2 per litre of leadfree will be a significant step toward more cycling too .

As for cycling without lights,has anyone done the ground work and by that I mean ride for a few weeks in the dark with and a few weeks without lights bags not being part of the scientific experiment .

I don't see the fuel hikes having had much effect. In my mind, it's much more about carrot than stick.

Most of the information collected by statisticians using questionairs are seriously flawed. Almost every time I complete one I find the question forbids me from giving the answer that the questioner needs.
eg. Would you be happy for your child to cycle to school?
I had this on a questionaire from my son's school. I would love my son to be able to cycle to school but the road to his school is far to dangerous to walk or cycle on. So how do you answer that with yes or no.
In the case of lights, when you ask a Police Officer was the bike lit and the bike had exceedingly good lights which were not carrying a BS mark, how does the Police Officer answer?
Also many of us know that a good reflectors and reflective jackets give far greater visibilty than ordinary cycle lights. The law wants the lights to be less than a metre high which generally leaves the lights obscured by the nearest car.
So maybee instead of the narrow question of do you have lights. You should ask how can I make myself so obvious that they notice me.

Another possibility is that a motorist driving along will see a lit cyclist and drive past with their normal generous 6" of safety gap but will be suprised when they see an unlit cycle at the last second and in their haste have to give the cyclist more space than normal.

Unfortunately the survey gives no answers. However the manufacturers of Reelights make quite definate claims with actual figures about the decrease in risk of collision due to being lit.

Precisely meic.

meic wrote:Most of the information collected by statisticians using questionairs are seriously flawed. Almost every time I complete one I find the question forbids me from giving the answer that the questioner needs.
eg. Would you be happy for your child to cycle to school?

These are "consumer surveys", (and the one from your child's school was probably designed by a teacher with no knowledge of how to construct a questionnaire) and quite different from questionnaires properly constructed, tested and analysed by serious researchers.

To get anywhere with this issue we would need a carefully designed observational study sampling different road types, geographical areas, times of night, etc to record the numbers riding lit and unlit, as a starting point.

To demonstrate the absurdity of the present argument, suppose it was found that in every 100 accidents 90 were found to wearing fluorescent yellow clothing while ten wore tweed jackets, plus fours, a deerstalker and were smoking a pipe. Or ten were found to be women and 90 were men and since we know around 50% of the population are women it must be safer to ride a bike if you are a woman. How many would vote for the deerstalker and a sex change?

Auchmill wrote:

meic wrote:Most of the information collected by statisticians using questionairs are seriously flawed. Almost every time I complete one I find the question forbids me from giving the answer that the questioner needs.
eg. Would you be happy for your child to cycle to school?

These are "consumer surveys", (and the one from your child's school was probably designed by a teacher with no knowledge of how to construct a questionnaire) and quite different from questionnaires properly constructed, tested and analysed by serious researchers.

To get anywhere with this issue we would need a carefully designed observational study sampling different road types, geographical areas, times of night, etc to record the numbers riding lit and unlit, as a starting point.

Wouldn't a true survey test the respondent as well. 'Do you feel road cyclists have, every right/ some right/ should be taxed/ off the road entirely? My complete guess would be the 'every right' replies rarely engage with cyclists negatively, however they are illuminated.
We may be mixing science, the objectively knowable, with the influence and expectations of the observer.

Serious surveys (questionnaires) are complex and time-consuming to create, involving, among other things, a bank of questions, pilot studies to test the questions, their reliability and validity and questions to test for internal consistency (to spot the respondents who are "taking the mickey") and trained administrators of the survey. Unfortunately, they are often seen as a cheap and cheerful way to get some data, which is usually meaningless, but is presented as cutting edge science.

Some of the recent headlines about the "hazardous" drinking habits of the middle classes seems to suffer from this defect.

If some survey were to be conducted, it seems pretty obvious to me that the issue is not really the one posed by this thread about lights. This is just a proxy for the wider issue of the vulnerability of cyclists on the road. Once you've got the lighting question out of your system, you are back in the area of policymakers paying lis service only to the the promotion of cycling

thirdcrank wrote:you are back in the area of policymakers paying lip service only to the the promotion of cycling

That's always the central point, which comes up frequently. The issue is: how do we influence policy makers, aka politicians, who have 10K other things on their mind, notably, setting their own pay?

It is said the answer lies in lobbying.

This is traditionally of two sorts: EITHER you pester your elected representatives, demonstrate, write to newspapers, petition, etc. It is generally accepted that this does not work.

Or: You employ retired decisionmakers, both elected and appointed as consultants. Even more so if they have been 'elevated to the peerage' as a reward for their endeavours the so-called 'Lord on the board' gives respectability to all sorts of projects. You then pay them well for their advice, which is a polite way of saying that you demonstrate to today's decisionmakers that if they toe your line, their turn on the gravytrain will come.

Questionairres properly constructed, tested and analysed by serious researchers arent available for any problems in real life. They are available only for theoretical problems where the variables can be controlled by the researchers.

How did we get onto questionnaires? They are useful (if properly designed) for finding out what people are thinking. Use of lights is a matter of observation. You "just" go out and look to see what is happening.

Now of course a proper experiment design using observation is just as difficult as a proper questionnaire design, and legitimate questions have been raised about whether current observations (accident statistics and ah-hoc reports of unlit cyclists) are adequate for the current purpose. However, that should be addressed by better observations, not questionnaires.