Dad stops kid from crashing bike into parked car (+ link to video)

davel wrote:

We don't have to.

Being able to explain the ped difference doesn't defeat your argument.

Your fallacious logic defeats your argument.

Your argument rests entirely on the premise that the pedestrian and cyclist rates follow the same trend.

The pre 1995 figures expose the flaw in your argument.

Hence your sad little attempts at deflection.

Rich_cb wrote:

davel wrote:

We don't have to.

Being able to explain the ped difference doesn't defeat your argument.

Your fallacious logic defeats your argument.

Your argument rests entirely on the premise that the pedestrian and cyclist rates follow the same trend.

The pre 1995 figures expose the flaw in your argument.

Hence your sad little attempts at deflection.

You know what is sad? Your insistence on pushing your own agenda against so many other posters who see through your shit.

There's a couple of % difference in Smed's abstract of your report.

Your argument boils down to a 5% increase in helmet-wearing that results in cyclists being 2-3% better off than pedestrians shows that helmets work? When the section of road users who don't wear helmets experienced more than 2-3% over the same period.

Because any rational person, even if they follow your line that helmets account for all the difference, will view that as helmets being poor compared to whatever fucking magic shit that pecs put in their heads in the same period.

davel wrote:

You know what is sad? Your insistence on pushing your own agenda against so many other posters who see through your shit.

There's a couple of % difference in Smed's abstract of your report.

Your argument boils down to a 5% increase in helmet-wearing that results in cyclists being 2-3% better off than pedestrians shows that helmets work? When the section of road users who don't wear helmets experienced more than 2-3% over the same period.

Because any rational person, even if they follow your line that helmets account for all the difference, will view that as helmets being poor compared to whatever fucking magic shit that pecs put in their heads in the same period.

Do you know what statistical significance means?

Rich_cb wrote:

davel wrote:

You know what is sad? Your insistence on pushing your own agenda against so many other posters who see through your shit.

There's a couple of % difference in Smed's abstract of your report.

Your argument boils down to a 5% increase in helmet-wearing that results in cyclists being 2-3% better off than pedestrians shows that helmets work? When the section of road users who don't wear helmets experienced more than 2-3% over the same period.

Because any rational person, even if they follow your line that helmets account for all the difference, will view that as helmets being poor compared to whatever fucking magic shit that pecs put in their heads in the same period.

Do you know what statistical significance means?

Better than you, you troll.

davel wrote:

Better than you, you troll.

Is that honestly the best response you've got?

Pathetic.

Rich_cb wrote:

alansmurphy wrote:

Hypothesis:

 

Quitting smoking causes cancer. 

That data you refer to is from 1993, mine was 2005-2010. You are wrong, my hypothesis is still correct.

 

 

Your hypothesis does not mention the date.

You're right, apologies, there was no clue in what I presented as in your world 1993 could quite possibly have come after the years 2005 or 2010...

alansmurphy wrote:

Rich CB here's some data for you:

2005: 25% of men and 23% of women smoked

2010: 21% of men and 20% of women smoked

2005: 77.9 people per hundred thousand had an incidence of lung cancer

2010: 79.4 people per hundred thousand had an incidence of lung cancer

Hypothesis:

Quitting smoking causes cancer. 

alansmurphy wrote:

Quitting smoking causes cancer. 

No mention of the date in your hypothesis.

Rich_cb wrote:

alansmurphy wrote:

Quitting smoking causes cancer. 

No mention of the date in your hypothesis.

Nor in yours you fucking prick!

alansmurphy wrote:

Rich_cb wrote:

alansmurphy wrote:

Quitting smoking causes cancer. 

No mention of the date in your hypothesis.

Nor in yours you fucking prick!

Now now children.

My hypothesis doesn't need a date.

If your hypothesis relates to a specific period in time then it does.

Not too complicated.

Rich_cb wrote:

alansmurphy wrote:

Rich_cb wrote:

alansmurphy wrote:

Quitting smoking causes cancer. 

No mention of the date in your hypothesis.

Nor in yours you fucking prick!

Now now children.

My hypothesis doesn't need a date.

If your hypothesis relates to a specific period in time then it does.

Not too complicated.

Well now you've stitched yourself up a few times. Firstly, how many cyclists were killed by serious head injuries in, oooh let's just say, before the motor car compared to now?

Secondly, where in your diagram did it say 'develop and test a hypothesis stating dates'?

Thirdly, about 150 posts ago, you refused a challenge to your hypothesis as the dates didn't match your data, in contradicting my VERY simple data set you're too stupid to realise you'd have needed a time machine.

I previously found you a little stubborn, unable to accept that your 'fact' was an opinion with a little belief thrown in. The more you post the more you seem arrogant and moronic, are you an MP?

alansmurphy wrote:

Well now you've stitched yourself up a few times. Firstly, how many cyclists were killed by serious head injuries in, oooh let's just say, before the motor car compared to now?

Secondly, where in your diagram did it say 'develop and test a hypothesis stating dates'?

Thirdly, about 150 posts ago, you refused a challenge to your hypothesis as the dates didn't match your data, in contradicting my VERY simple data set you're too stupid to realise you'd have needed a time machine.

I previously found you a little stubborn, unable to accept that your 'fact' was an opinion with a little belief thrown in. The more you post the more you seem arrogant and moronic, are you an MP?

You don't understand this at all do you?

Rich_cb wrote:

alansmurphy wrote:

Well now you've stitched yourself up a few times. Firstly, how many cyclists were killed by serious head injuries in, oooh let's just say, before the motor car compared to now?

Secondly, where in your diagram did it say 'develop and test a hypothesis stating dates'?

Thirdly, about 150 posts ago, you refused a challenge to your hypothesis as the dates didn't match your data, in contradicting my VERY simple data set you're too stupid to realise you'd have needed a time machine.

I previously found you a little stubborn, unable to accept that your 'fact' was an opinion with a little belief thrown in. The more you post the more you seem arrogant and moronic, are you an MP?

You don't understand this at all do you?

Sometimes, maybe, when everyone else in the 'room' says that you are wrong, take a look in the mirror!

alansmurphy wrote:

Sometimes, maybe, when everyone else in the 'room' says that you are wrong, take a look in the mirror!

He will argue the toss over anything that supports this argument, or flashing lights, or bright colours. It's exactly what he accuses BTBS of. 

That insurance thread I linked to (the one from 5 months ago that Richie couldn't remember, probably because it didn't suit his argument at that very point) - he knew how financial product management worked then, due to him never having worked in financial product management, but having actuarial mates, or something.

It's exactly the same as me knowing a couple of professional rugby players pretty much makes me a professional rugby player.

If 5 months ago, I was pretty much a professional rugby player, and today, I couldn't remember being pretty much a professional rugby player; if someone told me I was either mentally ill, or a trolly gobshite who argued the toss about so many topics I couldn't remember what I'd made up 5 months ago, I'd probably be a bit concerned and maybe agree. But if you're actually mentally ill or a trolly gobshite, I suppose you just keep arguing the toss over the topic du jour.

That he won't (or will pretend not to) remember this in 5 months tells me all I need to know. 

davel wrote:

alansmurphy wrote:

Sometimes, maybe, when everyone else in the 'room' says that you are wrong, take a look in the mirror!

He will argue the toss over anything that supports this argument, or flashing lights, or bright colours. It's exactly what he accuses BTBS of. 

That insurance thread I linked to (the one from 5 months ago that Richie couldn't remember, probably because it didn't suit his argument at that very point) - he knew how financial product management worked then, due to him never having worked in financial product management, but having actuarial mates, or something.

To be fair in this case it is different as he doesn't need to work in the area,  as there is absolutely nothing stopping him getting the articles he linked to and re-analysing the methodology and data himself.  

He may have an issue because he doesn't have the raw data but if he started by simply looking at the methodology the researchers used to collect their data, then he could understand some of the faults with the research. This basic technique is  actually taught at GCSE level statistics, while at degree level a statistical re-analyst of existing data is something that many degree students in different disciplines do as part of their projects/dissertations.   

Instead he decided he would just parrot bits of the articles to try and suit his theories, or hypothesis as he now calls it, he was trying to make. 

Bluebug wrote:

To be fair in this case it is different as he doesn't need to work in the area,  as there is absolutely nothing stopping him getting the articles he linked to and re-analysing the methodology and data himself.  

He may have an issue because he doesn't have the raw data but if he started by simply looking at the methodology the researchers used to collect their data, then he could understand some of the faults with the research. This basic technique is  actually taught at GCSE level statistics, while at degree level a statistical re-analyst of existing data is something that many degree students in different disciplines do as part of their projects/dissertations.   

Instead he decided he would just parrot bits of the articles to try and suit his theories, or hypothesis as he now calls it, he was trying to make. 

I did critique the methodology.

That's why I mentioned correcting for the participation rate.

I'm still waiting for you to point out the huge number of statistical errors I've made.

Quick reminder:
I've never said correlation proves causation.

OK. let's try this another way.

Your Hypothesis is:

Rich_cb wrote:

Cycle Helmets reduce deaths (from head injuries)

You have stated that:

Rich_cb wrote:

I've never said correlation proves causation

The hypothesis quite clearly states a belief that cycle helmets cause a reduction in fatalities from head injuries. The only data you have presented is related to correlation (and none of those being cyclist death rates from head injuries) so therefore, following your own comments and logic, we can agree that the results of your 'study' have to be labeled as inconclusive.

ClubSmed wrote:

OK. let's try this another way.

Your Hypothesis is:

Rich_cb wrote:

Cycle Helmets reduce deaths (from head injuries)

You have stated that:

 

Rich_cb wrote:

I've never said correlation proves causation

The hypothesis quite clearly states a belief that cycle helmets cause a reduction in fatalities from head injuries. The only data you have presented is related to correlation (and none of those being cyclist death rates from head injuries) so therefore, following your own comments and logic, we can agree that the results of your 'study' have to be labeled as inconclusive.

I'm pretty sure (Davel will hopefully confirm) that a few hundred posts ago I said that I couldn't prove the hypothesis with the data available.

Without further large scale research or access to data about the cause of death for all cycling fatalities we will never be able to prove the hypothesis.

The challenge has therefore been to test the hypothesis as far as possible with the available data.

The data I've presented fits the expected pattern if the causative factor(s) were protective equipment.

The data also indicates a specific factor reducing the risk of head injury for adult cyclists.

Do the data patterns prove that helmets are that factor?

No.

There could be other cyclist specific factors at work alongside helmets or there could be a combination of factors that don't include helmets.

Whatever the factor(s) is/are we would expect to see a strong correlation in time with the decline in fatalities and head injuries.

Increased helmet use does correlate exactly with the decline but I'm sure other things do too, possibly including sharks, ice cream sales and smoking cessation.

Rich_cb wrote:

Do the data patterns prove that helmets are that factor?

No.

There could be other cyclist specific factors at work alongside helmets or there could be a combination of factors that don't include helmets.

Whatever the factor(s) is/are we would expect to see a strong correlation in time with the decline in fatalities and head injuries. Increased helmet use does correlate exactly with the decline but I'm sure other things do too, possibly including sharks, ice cream sales and smoking cessation.

OMG I think he finally agrees with everyone else. 

I've fixed the one flaw: A small increase in helmet use does correlate roughly with the larger decline (see earlier post quoting from peer review of BMJ study as to why this is unsafe to draw any conclusions from).

CygnusX1 wrote:

Rich_cb wrote:

Do the data patterns prove that helmets are that factor?

No.

There could be other cyclist specific factors at work alongside helmets or there could be a combination of factors that don't include helmets.

Whatever the factor(s) is/are we would expect to see a strong correlation in time with the decline in fatalities and head injuries. Increased helmet use does correlate exactly with the decline but I'm sure other things do too, possibly including sharks, ice cream sales and smoking cessation.

OMG I think he finally agrees with everyone else. 

I've fixed the one flaw: A small increase in helmet use does correlate roughly with the larger decline (see earlier post quoting from peer review of BMJ study as to why this is unsafe to draw any conclusions from).

The decline in fatalities and serious head injuries occurred at the exact same time as the rise in helmet use.

The correlation is exact.

You can't draw conclusions about the magnitude of effect without proving causation.

Rich_cb wrote:

CygnusX1 wrote:

Rich_cb wrote:

Do the data patterns prove that helmets are that factor?

No.

There could be other cyclist specific factors at work alongside helmets or there could be a combination of factors that don't include helmets.

Whatever the factor(s) is/are we would expect to see a strong correlation in time with the decline in fatalities and head injuries. Increased helmet use does correlate exactly with the decline but I'm sure other things do too, possibly including sharks, ice cream sales and smoking cessation.

OMG I think he finally agrees with everyone else. 

I've fixed the one flaw: A small increase in helmet use does correlate roughly with the larger decline (see earlier post quoting from peer review of BMJ study as to why this is unsafe to draw any conclusions from).

The decline in fatalities and serious head injuries occurred at the exact same time as the rise in helmet use. The correlation is exact. You can't draw conclusions about the magnitude of effect without proving causation.

So you have a correlation in one case.  You haven't controlled for any possible confounding variables, as you just have one dataset, in one place at one time with one set of conditions.   You can't control for anything because you only have the one, strongly-contaminated, real-world dataset. 

You can't say what would have happened had helmet use not increased, still less what might have happened had there been wider social and legal changes that might-or-might-not be helped or hindered by a helmet-promoting culture.

What's the point supposed to be, here?  I zoned-out during the last few pages of toing-and-froing so don't know what point you are trying to make.

In general it seems to me there is a big problem with drawing conclusions from 'case studies' when you have no ability to perform a controlled trial.  As I think I said already, that's why we keep getting these dodgy stories (that the Daily Mail particularly loves) about this-or-that causing-or-preventing cancer.

To really say anything convincing without a controlled experiment you need a lot of repeated cases, across different conditions, not just one graph and dataset.

FluffyKittenofTindalos wrote:

Rich_cb wrote:

CygnusX1 wrote:

Rich_cb wrote:

Do the data patterns prove that helmets are that factor?

No.

There could be other cyclist specific factors at work alongside helmets or there could be a combination of factors that don't include helmets.

Whatever the factor(s) is/are we would expect to see a strong correlation in time with the decline in fatalities and head injuries. Increased helmet use does correlate exactly with the decline but I'm sure other things do too, possibly including sharks, ice cream sales and smoking cessation.

OMG I think he finally agrees with everyone else. 

I've fixed the one flaw: A small increase in helmet use does correlate roughly with the larger decline (see earlier post quoting from peer review of BMJ study as to why this is unsafe to draw any conclusions from).

The decline in fatalities and serious head injuries occurred at the exact same time as the rise in helmet use. The correlation is exact. You can't draw conclusions about the magnitude of effect without proving causation.

 

So you have a correlation in one case.  You haven't controlled for any possible confounding variables, as you just have one dataset, in one place at one time with one set of conditions.   You can't control for anything because you only have the one, strongly-contaminated, real-world dataset. 

 

You can't say what would have happened had helmet use not increased, still less what might have happened had there been wider social and legal changes that might-or-might-not be helped or hindered by a helmet-promoting culture.

 

What's the point supposed to be, here?  I zoned-out during the last few pages of toing-and-froing so don't know what point you are trying to make.

 

In general it seems to me there is a big problem with drawing conclusions from 'case studies' when you have no ability to perform a controlled trial.  As I think I said already, that's why we keep getting these dodgy stories (that the Daily Mail particularly loves) about this-or-that causing-or-preventing cancer.

 

To really say anything convincing without a controlled experiment you need a lot of repeated cases, across different conditions, not just one graph and dataset.

You were right to zone out.

He's still trying to make the same point.

Bluebug wrote:

He's still trying to make the same point.

He's also still waiting for your statistical critique.

Rich_cb wrote:

Bluebug wrote:

He's still trying to make the same point.

He's also still waiting for your statistical critique.

I doubt it would sway you even if it came from the emeritus professor of statistics @ Oxford / Cambridge / Yale / Harvard, anyway. 

But it’s my last day in the office this year, things are quiet and although I should be working on finalising the architecture of an advanced analytics workbench (R, Python etc.) capability on a Hadoop big data platform for my esteemed data scientist and actuarial colleagues to use in the new year, I’ll bite.

Sorry folks, but its...

(more follows)

FluffyKittenofTindalos wrote:

So you have a correlation in one case.  You haven't controlled for any possible confounding variables, as you just have one dataset, in one place at one time with one set of conditions.   You can't control for anything because you only have the one, strongly-contaminated, real-world dataset. 

 

You can't say what would have happened had helmet use not increased, still less what might have happened had there been wider social and legal changes that might-or-might-not be helped or hindered by a helmet-promoting culture.

 

What's the point supposed to be, here?  I zoned-out during the last few pages of toing-and-froing so don't know what point you are trying to make.

 

In general it seems to me there is a big problem with drawing conclusions from 'case studies' when you have no ability to perform a controlled trial.  As I think I said already, that's why we keep getting these dodgy stories (that the Daily Mail particularly loves) about this-or-that causing-or-preventing cancer.

 

To really say anything convincing without a controlled experiment you need a lot of repeated cases, across different conditions, not just one graph and dataset.

Increasing helmet use is correlated with decreasing fatalities, decreasing serious injuries, decreasing serious head injuries.

In these sort of situations you usually can't do a randomised controlled trial or a cohort study as the numbers involved are far too small so the studies would have to be enormous.

Approximately 100 deaths a year in a population of about 55 million (NI is not included in the stats).

Case control studies are probably your best bet and even then you have to make some compromises.

When comparing adult cyclists to adult pedestrians you do see a larger decrease in the head injury rate for cyclists.

This indicates a cyclist specific factor.

There is more than correlation at play here but given the limitations mentioned causation is almost impossible to prove.

Rich_cb wrote:

Increasing helmet use is correlated with decreasing fatalities, decreasing serious injuries, decreasing serious head injuries.

Let's be clear, when you say correlation, what you actually mean is "fairly similar overall general trend". There is not any direct correlation for example for the spike in cyclist fatalities in 2002 (that does not return to the same level until 2008) with the trend in cycle helmet wearing rates.
I also don't recall seeing the data about the decreasing serious injuries, I remember it being either static or increasing slightly (but I am still off with a temperature, so could be mistaken).
As for decreasing serious head injuries, there was a significant decrease amongst pedestrians too, the largest decrease was amongst child cyclists who did not show any increase in helmet wearing rates. So no definitive correlation here

ClubSmed wrote:

Let's be clear, when you say correlation, what you actually mean is "fairly similar overall general trend". There is not any direct correlation for example for the spike in cyclist fatalities in 2002 (that does not return to the same level until 2008) with the trend in cycle helmet wearing rates.
I also don't recall seeing the data about the decreasing serious injuries, I remember it being either static or increasing slightly (but I am still off with a temperature, so could be mistaken).
As for decreasing serious head injuries, there was a significant decrease amongst pedestrians too, the largest decrease was amongst child cyclists who did not show any increase in helmet wearing rates. So no definitive correlation here

We've covered the child cyclist data.

Correct for participation rates and the difference between the decline in pedestrian and cyclist rates disappears for children but not for adults.

The correlation doesn't have to continue indefinitely to be correct.

A new factor could easily have become significant at that time. The fact that both pedestrian and cyclist rates plateau and then decline again simultaneously indicates a shared factor like traffic increasing or mobile phone ownership (correlation is not causation).

The KSI rates followed pretty much the exact same pattern as the fatality rates.

Rich_cb wrote:

[quote=

The correlation doesn't have to continue indefinitely to be correct.

A new factor could easily have become significant at that time.

Like the forming of one direction or the launch of the iPhone?

This is where you don't need the cleverer people on this board to prove you an idiot, I can do it. You are finding a piece of data that doesn't meet your desired result and openly just throwing out 'a new factor' whereas everything else is helmet.

How can you prove the other things weren't due to infrastructure and the anomalies be helmet related, or blue cars or Santa or sharks?

Rich_cb wrote:

 

Increasing helmet use is correlated with decreasing fatalities, decreasing serious injuries, decreasing serious head injuries. In these sort of situations you usually can't do a randomised controlled trial or a cohort study as the numbers involved are far too small so the studies would have to be enormous. Approximately 100 deaths a year in a population of about 55 million (NI is not included in the stats). Case control studies are probably your best bet and even then you have to make some compromises. When comparing adult cyclists to adult pedestrians you do see a larger decrease in the head injury rate for cyclists. This indicates a cyclist specific factor. There is more than correlation at play here but given the limitations mentioned causation is almost impossible to prove.

The fact that you can't do good studies, doesn't make the bad ones any more convincing.  Believing that it does seems to be the main fallacy you are perpetrating here.

FluffyKittenofTindalos wrote:

The fact that you can't do good studies, doesn't make the bad ones any more convincing.  Believing that it does seems to be the main fallacy you are perpetrating here.

The fact you can't realistically do gold standard studies doesn't mean you shouldn't examine the evidence at all.

All I've done is present the evidence available.

The evidence suggests an adult cyclist specific factor that is reducing head injuries.

Rich_cb wrote:

CygnusX1 wrote:

Rich_cb wrote:

Do the data patterns prove that helmets are that factor?

No.

There could be other cyclist specific factors at work alongside helmets or there could be a combination of factors that don't include helmets.

Whatever the factor(s) is/are we would expect to see a strong correlation in time with the decline in fatalities and head injuries. Increased helmet use does correlate exactly with the decline but I'm sure other things do too, possibly including sharks, ice cream sales and smoking cessation.

OMG I think he finally agrees with everyone else. 

I've fixed the one flaw: A small increase in helmet use does correlate roughly with the larger decline (see earlier post quoting from peer review of BMJ study as to why this is unsafe to draw any conclusions from).

The decline in fatalities and serious head injuries occurred at the exact same time as the rise in helmet use.

The correlation is exact.

You can't draw conclusions about the magnitude of effect without proving causation.

You can, you just need to use the word "if" to clarify that it is not a fact and purely based on a supposition