Get your favorite beverage, sit back, and join in the discussion
You are not logged in.
http://news.yahoo.com/s/afp/britainpeoplehistoryoffbeat
I have nothing to add to this
Offline
I do.
UKTV Gold television surveyed 3,000 people.
First off, the people that watch UKTV Gold are morons ANYWAY. Second, 3000 people is not a large enough sample to make any kind of extrapolation onto a population of over sixty million people.
Now, I know you're not really saying we're as stupid as you guys, are you? XD
CSquared
Offline
Um... 3000 / 60000000 = 1:20,000 ratio. That is MORE than an acceptable scientific ratio, so long as the sampling was random.
Which, of course, is where this falls apart, because, as you point out, there is a specific demographic that watch the television network in question. Not knowing the network in question, I'd guess it would be similar to using the ESPN audience as a base for judging intellect in the US.
As to the question of which country is more stupid, I doubt any studies have been done on the actual intellectual capacity (as opposed to the education level) of various countries. Knowledgeable != More Intelligent, and Ignorant != Stupid. Just watch a redneck jerry-rig something when it breaks, and then watch a scientist staring dumbly at a broken shoelace for which they have no replacement, and you'll understand this concept.
PS: For the Programming-Ignorant who read this: != means "does not equal".
Offline
Eric Storm wrote:
Just watch a redneck jerry-rig something when it breaks, and then watch a scientist staring dumbly at a broken shoelace for which they have no replacement, and you'll understand this concept.
I'm a scientist. If my shoelace breaks, I'll splice the two halves together. See, I'm also a sailor. XD
I still say that in order to be really conclusive, a sample should be more than one in twenty thousand. One in ten thousand, maybe - but preferably even more. Or, we should make people that publish these surveys work out the variance. "30% +/- 10% of people said this: ..." That'd really piss them off. XD
CSquared
Offline
Actually, if the survey was done by anyone competent, they DID work out all the statistics. They just didn't clutter up their news blurb with them.
And as to your shoelace, you clearly got my point, so...
Offline
A lot of time when the variances are not included and are significantly large (Say +/- 10% or greater) They only give the number that makes the story more notable. i.e. 30% (+/- 10%) believe in Alien life forms. The news story might read 40% of people or 20% of people... and that is a significant difference. I had an english prof recently who crucified someone for doing that.
Offline
And the variances will only be that large if your sample is significantly non-random.
Which makes me wonder, not having taken statistics class: How the hell do they know what the variance is? In order to know that, wouldn't they need to know what the ACTUAL number was??? It's like those scientists who say, "95% of animal species have yet to be discovered." How the fuck would anyone actually know that???
Offline
The variance is worked out by a horribly complex manner of multiplying, adding, dividing and subtracting inaccuracies. Basically, if you have two pieces of data, and the accuracy of each is (for sake of argument) +/- 5%, then the accuracy of the whole thing is +/- 5% (I think, I can't really remember. Haven't done it in a long time). In other words - the accuracy of the whole depends on the accuracy of the various constituent pieces. You can't have 100% accuracy when measuring something, for example. A ruler measures only to the nearest half a division (scientifically speaking), therefore we should actually write one centimetre as 1cm +/- 0.5 x 10^-1cm. Which, without the evil notation, means 1cm +/- 0.5mm.
But yeah, saying that 95% of animal species are still to be discovered is stupid. XD
CSquared
Offline
Except the problem with that analogy as it relates to this problem is that there are discrete values involved. It's a "digital" rather than an analog question. In other words, people were asked a question. They either got the right answer, or the wrong answer. It's a 0 or a 1. There is no .5 in that situation.
My ultimate question still remains: Since you're polling people, and not something that you can actually physically measure, how do you "guess" at the inaccuracy? Yes, in the case of a physical measurement, it depends upon the precision of the instrument... but we're not talking about physical measurement, but psychological measurement.
If you KNOW your measurement is inaccurate in these instances, you'd also have to know by how much, wouldn't you? There's not a "measurement imprecision" to be introduced, is there?
Offline
True, but most of these people probably gave answers along the lines of "Isn't he that person that... wait, no, wasn't he the... um...". It's not quite binary. More like trinary. Yes, no and null. There's room for inaccuracies, and even if completely fabricated, they could be based on an educated guess. We're not talking physics, after all - it doesn't have to be completely accurate. XD
CSquared
Offline
my head hurts too much math formulae
Offline