you ever wonder if one of these studies has ever just had an abormally large amount of stupid people in them?
[QUOTE=Y'all.;41491232]how come they can get away with saying "40% of australians" if they only studied like 1500-2000 people? i guess they technically mean 40% of the australians that they studied but its worded in a way that makes it sound like they're talking about the entire population.[/QUOTE]
You need a test-size of a little less than 1.100 people before your results are broad enough that any more people added won't change it in any noticable amount. Like, the difference on asking 1.100 and 10.000 is almost nill.
It seems like they worded the question as "How long does it take for the earth to travel around the sun" for which the answer can be "a year" or "365 days", but getting that question wrong [I]doesn't mean[/I] that someone doesn't know "how long a year is".
It's sort of understandable because it's a survey on scientific knowledge, rather than common knowledge.
It's all just badly worded and sensationalized.
[QUOTE=Y'all.;41491232]how come they can get away with saying "40% of australians" if they only studied like 1500-2000 people? i guess they technically mean 40% of the australians that they studied but its worded in a way that makes it sound like they're talking about the entire population.[/QUOTE]
Statistics. Usually a large enough sample size will give you a fairly accurate representation of the whole population that it is a subset of.
[QUOTE=Y'all.;41491232]how come they can get away with saying "40% of australians" if they only studied like 1500-2000 people? i guess they technically mean 40% of the australians that they studied but its worded in a way that makes it sound like they're talking about the entire population.[/QUOTE]
most surveys work like that, because it's not reasonable to poll the entire population of a country/continent/planet when the scientific community accepts the results of a smaller scale poll.
[IMG]http://blog.vovici.com/Portals/60483/images//sample-size-table.png[/IMG]
There we go.
[editline]17th July 2013[/editline]
Notice how when you go from 100.000 to 1.000.000, you only need to ask [I]one[/I] more dude to be within a 5% margin of error for the entire population.
[QUOTE=Riller;41491295][IMG]http://blog.vovici.com/Portals/60483/images//sample-size-table.png[/IMG]
There we go.
[editline]17th July 2013[/editline]
Notice how when you go from 100.000 to 1.000.000, you only need to ask [I]one[/I] more dude to be within a 5% margin of error for the entire population.[/QUOTE]
that's a simplistic breakdown, too. it's a bit more complicated because your sample has to reflect different levels of education, people of other races, income levels, etc., depending on what sort of data you have to get. asking 384 random people on the street is not going to give an accurate representation of 1,000,000 scattered across a country, for example.
basically statisticians have been working on this shit for a long time, they know how to find accurate samples, assuming they actually want to get accurate data. just because a sample size might seem small to you does not mean that the poll or survey is unscientific.
[QUOTE=yawmwen;41491382]that's a simplistic breakdown, too. it's a bit more complicated because your sample has to reflect different levels of education, people of other races, income levels, etc., depending on what sort of data you have to get. asking 384 random people on the street is not going to give an accurate representation of 1,000,000 scattered across a country, for example.[/QUOTE]
Well, of course, yeah. Just posted it for the people who claim that 2.000 people is too small of a sample size to give any sort of idea whatsoever.
[QUOTE=Riller;41490634]Depends on if 365 is an acceptable answer, or if they [I]require[/I] you to say 365,25. Because if the latter, it's an easy mistake to make.[/QUOTE]
365.2488
[QUOTE=Riller;41491419]Well, of course, yeah. Just posted it for the people who claim that 2.000 people is too small of a sample size to give any sort of idea whatsoever.[/QUOTE]
the point wasn't to go against you, but to support what you were saying by alluding that statisticians have a pretty complex and comprehensive methodology to gather data. these guys know what they are doing since it's pretty much their job to get as accurate(or in some cases accurately skewed) data as they can possibly get.
I bet this survey was done in <insert local bogan suburb>
hell, nate silver, a statistician, has constructed a system that has predicted several past elections very accurately.
I choose Granville
my merge
statistics are fucking crazy man
ye right ya fuckin cunts, we no this!!
australia has best educatin system in world
Sorry, you need to Log In to post a reply to this thread.