How?

[lukesdad]

Taking Bens 40% example ( again a wild guess) 90% of commuters wear helmets, 99% of mtbers wear helmets. On the otherhand 90% of 1 hour a week in the local park dont wear helmets, and 99% of old dears pushing (mostly ) their bike to the local shops dont wear helmets.

What can we infer from that in realation to hospital admissions?

Of course you are welcome to provide evidence my guess is wildly wrong.

 

[benb]

I think you do.

Lets say for instance the injuries to helmet users are in a higher risk group to start with, i refer you back to my time question. Would you not expect the figures to be even higher than they are.

I already said you would need to correct for confounding variables.

 

[benb]

On the subject of injuries to helmet users showing no change in countries that have adopted a compulsion law. Why would we expect a change ? Surely it would be the non wearers giving up cycling I would expect the injuries to increase not decrease.

This statement shows you understand nothing about statistics.

The reduction in cyclists in these countries has typically been around 20-30%
Helmet wearing amongst the rest has been very high, as most people comply with the compulsion law.
Yet head the proportion of head injuries remains largely static.

What conclusions can you draw from this?

 

[screenman]

That the stats maybe wrong or made up by poor guess work.

 

[lukesdad]

This statement shows you understand nothing about statistics.

The reduction in cyclists in these countries has typically been around 20-30%
Helmet wearing amongst the rest has been very high, as most people comply with the compulsion law.
Yet head the proportion of head injuries remains largely static.

What conclusions can you draw from this?

Well until you can prove otherwise, my guess would be it is the comitted high risk users in the first group indicated above who have kept cycling, and the park dwellers and old ladies have given up. Of course i could be wrong but....

This proves you know nothing about variables or just choose to ignore them.

 

[benb]

Well until you can prove otherwise, my guess would be it is the comitted high risk users in the first group indicated above who have kept cycling, and the park dwellers and old ladies have given up. Of course i could be wrong but....

This proves you know nothing about variables or just choose to ignore them.

I think you'd have to have a pretty compelling piece of evidence to claim that!

As I have said, the stats are already corrected for confounding variables.

 

[StuartG]

Well until you can prove otherwise

I have to agree with Benb that you have no understanding of statistical inference*. That's fine. I too have no understanding of many things. Time is too short to be a true renaissance man. The issue is realising our ignorance.


* Stats is not about proving things. Its basically about maximising and measuring the level of confidence you can have in a particular hypothesis. Certainency is rare. But knowing when to bet on an outcome is pretty useful. Getting the best odds makes you a winner. Well most of the time

 

[lukesdad]

I have to agree with Benb that you have no understanding of statistical inference*. That's fine. I too have no understanding of many things. Time is too short to be a true renaissance man. The issue is realising our ignorance.


* Stats is not about proving things. Its basically about maximising and measuring the level of confidence you can have in a particular hypothesis. Certainency is rare. But knowing when to bet on an outcome is pretty useful. Getting the best odds makes you a winner. Well most of the time

Educate me then and answer some of my questions or at least humour me and try. Everytime this sort of questioning comes up the only reply is, you don't understand statistics. Now shall we start at the beginning, who is being injured and who is giving up cycling, until you know this your statistics are pretty meaningless

 

[lukesdad]

For the sake of arguement lets say a compulsion law was introduced here. How many mtber's and commuters do you think would give up cycling? Then ask yourself how many "part time cyclists" would give up cycling. Go on take an educated guess.

 

[david k]

Now can I get this point across to some of you once again, I am not for compulsion and doubt I ever will be. There are far to many people who are obese that wearing a helmet would just be another excuse not to exercise.

But these obese people currently dont exercise and there are no helmet laws? So other than taking away an excuse it would make no difference in this example

 

[david k]

For the sake of arguement lets say a compulsion law was introduced here. How many mtber's and commuters do you think would give up cycling? Then ask yourself how many "part time cyclists" would give up cycling. Go on take an educated guess.

i suspect not many but i could be wrong.
I was told that in London the congestion charge saw a reduction in car drivers at first but eventually it returned to normal. I imagine something similar may happen here.
Thats doesnt mean i am promoting compulsion before any takes that approach in any reply

 

[StuartG]

Educate me then and answer some of my questions or at least humour me and try. Everytime this sort of questioning comes up the only reply is, you don't understand statistics. Now shall we start at the beginning, who is being injured and who is giving up cycling, until you know this your statistics are pretty meaningless

Well would you expect a brain surgeon to distill 6 years of formal training and decades of experience into a pithy explanatory post? Or would it be better seeing someone taking an axe to a skull to just say "please don't do that"

There is too much ground to cover before you could take an informed decision on a set of statistics. You do have to rely on people who do understand it to give a fair and objective summary to you. Gaining and keeping trust to do this can be hard in the public policy area where everybody else is pursuing an agenda. We have a continuing issue between Home Secretaries and professionals over drug statistics and policy because facts seldom fit with what people want to hear. Nevertheless a lot of work is done to create reliable statistics and inferences and to try and cry foul when they are misused.

I was lucky - I worked in the private sector where maximising profit and reducing risk rather excludes these problems. You would just say to a decision maker I have an answer for you that will improve the business - or we can't forecast this so don't try. When it comes to helmets and statistics the statisticians are pretty united that it is unclear whether helmets are a benefit or not. That's not what either side wants to hear and many people really can't cope with that uncertainty.

But there it is. You wears a helmet or you do not. You might as well toss a coin to decide as to ask a statistician. It can take a lot of learning to know you don't know ... now when it comes to other actions to reduce casulties we can be of a lot more help. The paper on the effect of urban 20mph zones on KSI is a classic in statistical inference. Its a result I would not have expected.