Here's another great one for you GI's out there. From the metro Wed Oct 19: What do you think folks?!
Metro How can the head of risk suggests that correlation? It might be an interesting fact, but that's all there is to it surely. Some people...
What a load of nonsense So does that mean if one those 'risky' names changes his/her name, they suddenly become a safer driver overnight?!!? It's probably that these names are frequently occuring ones and hence more accidents are attributed to people with these names....
well you hadly get people called Peggy or Damon anymore that's why there are the safest driver's cos the frequency of claims reported them will be low. However, my mum is callled Helene and she drives like a nutter.....Could there be some truth..............
I'm a Statistician, and when I read things like this it really drives me crazy I'm sure that it's just a wind up anyway, but far too often things like this are actually being used because, well, they look like they work don't they... I believe that it is actually highlighted in the CT3 notes that correlation isn't always an indication that the variables are linked, but this is sadly ignored far too often. It is a fundamental Statistical rule that you *cannot* simply fit patterns to results. It is possible to infer anything from anything (Hence the old adage "you can prove anything with Statistics"), and spotting patterns like this leads to series mistakes. Slightly unrelated, but an example of using Statistical methods with know understanding of what you're doing: Theyt predicted in all the national newspapers that August 12th this year would be the hottest day of the year. Betting shops took money on it, and most of my colleagues booked days off work. Do you know how they came up with this fantastic piece of analysis?? They averaged when the hottest days were over the last few years... Genius. And, of course, complete rubbish... And it was much to my delight that it was cloudy all day before raining... leafy
I wouldn't be so quick to call it complete rubbish - if its been the hottest day on average in the past then its probably as good an estimate as any other, and better than most. The temperature isn't completly random - it is hotter in summer than in winter for example! I would be slightly worried about your colleagues who took days off because of this 'advice' though. And how do you average days anyway? 'The average of July 10th and August 30th is the 4.5th of August'. Clearly.
I am thinking of all the people I know of by these names and actually I can believe this is true. It could be due to the parents. For example a parent who is more likely to raise a risky child could be more inclined to those types of names.
Yes, I completely agree there is strong statistical and scientific evidence that summer is hotter than winter (I'm being diliberately flippant). However predicting the hottest point in a year with any seriousness 6 months in advance with a 24hour degree of certainty based solely on past experience is complete rubbish when weather systems from one moment to the next behave in such a chaotic and unpredictable fashion. I do understand your point, I know you weren't lending your weight behind the claim, and if they had been estimating the hottest month, or even couple of weeks I might have let it pass. Past experience would then be a good guide of an underlying trend that might signify a period of the year at which we are closer to the sun, or a particular sunspot makes an appearance... But making an outright claim that you have located the future hottest 'day' is extremely flawed. Weather patterns are extremely chaotic, and while they generally average out over periods of weeks even the most advanced scientists would admit to being unable to be certain of the temperature it will reach next Monday, let alone next August... And again, you are quite right when you say that averaging from past experience is the best method we have...but when has "It's the best we can do" ever been good enough for rigorous Statistical analysis?? Having said that, "The Hottest Day of The Year Will Be August 12th +/- 12days" would hardly make the front page of a Newspaper, so if people want to pay these people to come up with the claims they do then fair play to them. All I know is I was glad I knew better than to listen to it and thus stayed warm and dry and retained a day's holiday!! leafy
Leafy - if u get the chance - read Freakonomics by Steven Levitt. Quality book. reading your post, I think you'd enjoy it. Nadeem
