This article, if you delve into it, gives an insight into the difference between the Oxford model approach and the Imperial Model.
Its behind a paywall.
It's behind a paywall :-(
You could try this:
I do often wonder you know if you could using statistics, prove that one make of chainsaw has caused moor fatal accidents than another, or one colour of car is involved in more collisions than another etc. Brian
Statistics don't lie, only misleading presentation of partial data does. If one colour of car causes more accidents in a statistically significant way, then either it is a false result due to chance, in which case repeating the experiment on a separate set of data is most unlikely to show the same effect again; or it is a real finding, in which case the experiment is repeatable and we have to go and look for the reason. If statistical tests are honestly applied then they are a very powerful tool for confirming or excluding a real correlation. But they don't show the reason; for instance, in your car colour case it could be dangerous drivers are more likely to choose a particular colour rather than a direct effect of the colour.
Unfortunately that version doesn?t give the outline of the assumptions - just the sleaze story, which is disturbing in itself.
Prof Gupta is being ignored for the wrong reasons.
Meanwhile, Ferguson- despite his history with modelling- was listened to and here we are.
Thank you the media.
There is another factor you?re missing, it applies in particular to the car colour example, number in population.
Car colours go in and out of fashion. I tend not to follow it but, for example, I recall in the early 90s a Wine Red seemed popular. Around 2000 it was Silver.
Some years there are more car sales etc. , so more of the popular colours.
There are companies who, it is claimed, adjust premiums based on colour but whenever I?ve asked it hasn?t been a factor.
In message <1oodpln.c2iyvh1q3lg6qN% snipped-for-privacy@hayter.org, Roger Hayter snipped-for-privacy@hayter.org writes
I choose white coloured cars on the basis that they might be more visible and hence less likely to be hit in error. They also don't show khaki coloured dirt readily:-)
Sadly, on the two near miss occasions, the other driver was clearly distracted, mobile phone/children in car, and simply not paying attention.
Pile ups, caused by driving conditions, seem unlikely to relate to car colour unless impatient drivers have some preference.
Bright yellow should stand out but, I would imagine, could be considered aggressive and, if insurers do consider it, add to premiums.
I pick on practical and taste reasons.
I had a silver Honda, lovely car, until you needed to match the paint. Even supermarket trolley prangs turned into nightmares. I was told metallic silver is one of the worst to get right.
Reds seem to fad.
Keep it simple, go with Henry Ford on colour for cars, if you can. ;-)
You raise the important point that assumptions about the distribution of variables should be tested. That kind of non-random distribution of colours is something that would need to be allowed for in the design of the study in order to make the statistics valid.
He didn't miss it, he simply didn't *try to* list all the potential instances.
I wonder if your wording is actual arrogance or just clumsy?
Maybe it's just another example of a left brainer view of the world where everything *they* see and understand of the world is the one and only way, no matter what any other person says?
And believe in experts when their guidance suits their case / cause but not when it doesn't?
Cheers, T i m
It is EXACTLY the kind of thing those who misuse statistics overlook.
One of the ?life lessons? I added to my teaching was to always treat statistics with suspicion. I had a pet phase an OFSTED inspector loved.
ROTFL
Tim, you do realise, you fit your ?left brainer? description perfectly, don?t you.
I don't think I've ever chosen the colour of a car because the option has never arisen ... ;-)
I think I have 'accepted' a car colour (even what it wouldn't have been my choice) because the alternatives would have taken longer or been more expensive (company cars).
Our current car is silver because it was my Dads and Mum gave it to us when he died (8 years now I think). Before that it was blue because a mate sold it to me for £100 (lasted 7 years I think). Before that it was my ex Company car and that was Red because I could have a red 2GL that day, rather than waiting 6 weeks for whatever colour I fancied and a 1.6L (the 2L lasted 23 years). ;-)
Cheers, T i m
Bright yellow should stand out but, I would imagine, could be considered aggressive and, if insurers do consider it, add to premiums.
I pick on practical and taste reasons.
I had a silver Honda, lovely car, until you needed to match the paint. Even supermarket trolley prangs turned into nightmares. I was told metallic silver is one of the worst to get right.
Reds seem to fad.
Keep it simple, go with Henry Ford on colour for cars, if you can. ;-)
===
Mine is silver ..... no one has bumped it .. yet!
The Honda, although a lovely car, seemed 'jinxed' - trolley 'dings', all kinds of things. We traded in, not because of that, it was very expensive to insure when we added eldest as a driver, plus Senior Management (who had 'inherited' it by then) wanted a small car. We bought a new Picanto, which was great. We replaced it with the Smart Car we now tow behind the motorhome.
Be careful, you might not be able to get up. ;-(
The sad thing is you would never be able to see why you aren't right.
Or maybe not sad, 'ignorance is bliss' eh, especially when it comes to Veganism! ;-)
Cheers, T i m
The Honda, although a lovely car, seemed 'jinxed' - trolley 'dings', all kinds of things. We traded in, not because of that, it was very expensive to insure when we added eldest as a driver, plus Senior Management (who had 'inherited' it by then) wanted a small car. We bought a new Picanto, which was great. We replaced it with the Smart Car we now tow behind the motorhome.
===
Sounds good! Mine is a SWB Mitsubishi Shogun. The Boss has the same but LWB :)
I had sports cars for years, then I tried an SUV and I was sold :)
Colours of cars yes, although I can't remember which one and indeed it may change over time.
It was an article on data mining, many years ago. They basically programmed the computer to look for correlations and set it off, without giving is specific correlations to look for. The result was that it found some that no-one would think of looking for - I remember one being that you were more likely to die in a car accident if you were born on a Tuesday! I think that was the day mentioned. I can't understand why it could make any difference, but that's what was found in the data.
SteveW
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.