Advanced search

Lies, damn lies, misrepresentation, misinterpretation and STATISTICS

(104 Posts)
StealthPolarBear Wed 29-May-13 22:44:55

A thread for anyone who is infuriated by and/or enjoys reading about this type of stuff. I love it, but annoyingly at the moment I can't think of any examples, other than dull work-related ones. For example, if you're looking at something as a proportion of the whole, then you can't consider one thing in isolation. For example, as Trills mentioned on another thread, let's assume heart disease is the biggest killer of adults (which I believe it is). Let's assume 10% of deaths to adults in 1950 were of heart disease, compared to 42% now. Shocking rise? Probably not.

Plus I will attempt to explain the Monty Hall (think that's the name) problem to anyone who is interested and who doesn't already know it.

TheDoctrineOfSnatch Thu 30-May-13 09:15:01

Exactly, Indith, exactly brew

Indith Thu 30-May-13 09:15:26

Oh and understanding the difference between a factor and a cause. Walking into the road without looking was not the cause of my death, it was a factor. The cause was the car that hit me.

StrawColoured Thu 30-May-13 09:15:33

Fascinating thread!

Advertising with statistics is so persuasive though. There's some body lotion or something that claims to "lock in ten times more moisture" - but doesn't explain ten times more than what. Smearing lard on your skin would "lock in" about a hundred times more moisture anyway. And since when is "moisture" so good for the skin? Humans aren't dolphins, they won't die if their skin is dry.

StealthPolarBear Thu 30-May-13 09:17:31

It's all a load of boswellox

TheDoctrineOfSnatch Thu 30-May-13 09:20:05

Is a cinder what's left over when you burn coal etc, but it's a bigger piece than ash?

Hence "cinders and ashes" from Thomas the Tank Engine.

TheDoctrineOfSnatch Thu 30-May-13 09:23:41

There was a long thread about cycling recently where one poster believed that, since the absolute number of cyclists killed in the UK and the Netherlands was the same, UK drivers must be more tolerant of cyclists as in the Netherlands people were more often on cycle paths than the road.

Factor in the regular cycling population in each country? Apparently that was irrelevant....

bigkidsdidit Thu 30-May-13 09:24:40

Reading with interest

I'm a scientist but my stats knowledge is shameful (in my defence i'm a biologist grin )

I don't understand the game show, can you explain?

I do think when discussing death / illness, newspapers should use actual numbers as well as incresaed risk. I was listening to Today today and there was a piece on long term use of painkillers increasing death rates. They did actually use real numbers and said the risk of death increased from (I think) 2 oer 1000 to 3 per 1000. If you have arthritis and were told risk of death increases 50% you might panic, but 2 to 3 per 1000 would probably be worth it to be pain free.

Indith Thu 30-May-13 09:27:14

Yup. All a pile of shite.

I find it sad that I did a presentation on infant sleep including a nice lengthy section on bed sharing on the day that the newspapers published that research done that said co sleeping was incredibly dangerous. My presentation linked all the research done by Helen Ball and mentioned the most recent suggestion that once other risk factors are removed co sleeping is possibly safer than cot sleeping. The research that anyone can access through the links on ISIS is all very well conducted with groups selected and matched and questioned specifically for the purpose of that study. Then we have this research which makes it into the papers and scares everyone silly nad could potentially have a huge impact on women and their babies because nobody wants to do something that will make their babies a huge 5 times more likely to die. Yet this research was done as a retrospective looking at other studies and using their information so there is no control over methodology and they are relying on information gathered while researchers were looking at other things. But this is what the papers have put out there as gospel.

StealthPolarBear Thu 30-May-13 09:27:47

If you stick with your original door, your chance of getting it right is the same as it was originally - one in three. If you swap, your chance of getting it right is the same as it was before he opened the door - two In three. It seems as though it should be 50/50 once a door is open but the key is that the gamesmaster knew what was behind each dpoor

Indith Thu 30-May-13 09:31:09

YEs bigkids, real numbers are needed. If we don't know the baseline then anything like an increase of 67% or a rise by a factor of 5 is meaningless.

TheDoctrineOfSnatch Thu 30-May-13 09:35:06

Contemplates drawing a probability decision tree in ASCII, decides not to...

At the time of your first choice, you had a one in three chance of being right.

Once one choice is eliminated, by changing your choice you have a two in three chance of being right.

It's a bit mind-bending. But it works because the door opener knows the answer. So if you picked the car door, the door opener can open either of the other doors, but if you picked a goat door, the opener has to open the other goat door. As you were more likely to pick a goat door first time, it's more likely that the unopened and unpicked door is the car door.

daftdame Thu 30-May-13 09:35:16

This book is full of this stuff...

Love the title. Amazing how much bad science there is around. (Wha ha ha!!!!)

StealthPolarBear Thu 30-May-13 09:36:04

the big thing about o-cleeping is that they lump in people who fell asleep with their babies on a sofa (which I've done but far from ideal). If they separated out intent then I bet the findings would be very different.

StealthPolarBear Thu 30-May-13 09:36:38

ooh thatnks daftdame that's one I haven't read.
Just finished Naked Statistics, that was good.

bigkidsdidit Thu 30-May-13 09:36:58

I get it! thanks smile

StealthPolarBear Thu 30-May-13 09:37:36

Snatch, I have tried to describe it using a deck of cards, where the idea is to pick the ace of spades. I think it's a lot clearer when the gamesmaster turns over 50 cards.

Indith Thu 30-May-13 09:45:11

not on all studies they don't, tis a bit of a myth I think that they all lump them in together. But usual place of sleep makes a difference. For example a child who normally sleeps in a cot but who comes into bed with parents one night for some reason is at a greater risk thana child who bed shares all the time.

Anyway, I suppose I should start heading towards class. I had ideas about being productive this morning but apparently not. Maybe they need to ban MN from the uni laptops!

StealthPolarBear Thu 30-May-13 09:49:12

Soery good point. I meant on the ones id seen there was never any mention of where the child started the night or, as you say, where they usually slept.
Have a good day!

StealthPolarBear Thu 30-May-13 10:14:31

Just read the sausage one. How can someone "invent jogging"? Thats liks saying someone invented breathing or sneezing.

StealthPolarBear Thu 30-May-13 10:44:13

Just bumping

StealthPolarBear Thu 30-May-13 11:13:39


Witt Thu 30-May-13 11:16:38

Please can you explain the Monty Hall problem to my MIL. My DH and I have tried several times and we hit the same problem - she wants to win a goat!

StealthPolarBear Thu 30-May-13 11:19:59


TheDoctrineOfSnatch Thu 30-May-13 11:23:20

Well, that's ok - she can maximise her chances of winning a goat by not swapping her guess!

StealthPolarBear Thu 30-May-13 11:28:08

Or win a car, sell it and buy as many goats as she wants

Join the discussion

Join the discussion

Registering is free, easy, and means you can join in the discussion, get discounts, win prizes and lots more.

Register now