It looks like you're using an Ad Blocker.
Please white-list or disable AboveTopSecret.com in your ad-blocking tool.
Thank you.
Some features of ATS will be disabled while you continue to use an ad-blocker.
Originally posted by BrokenCircles
reply to post by EarthCitizen07
How can 1,500 out of millions be an accurate poll sample?
That's why it is called 'A Sample'.
Sample
1: a representative part or a single item from a larger whole or group especially when presented for inspection or shown as evidence of quality: [A specimen]
2: a finite part of a statistical population whose properties are studied to gain information about the whole
Is it an accurate sample? I don't know. Probably is though.
Is it accurate for the still yet to come definitive conclusive results? Nope.
Why not? because it is just a sample.
edit on 10/14/12 by BrokenCircles because: (no reason given)
Originally posted by FurvusRexCaeli
reply to post by EarthCitizen07
There is an online sample size calculator here. There's about 250 million adults of voting age in the US. To accurately model that population with a 4% margin of error and a 95% confidence interval, you need to sample 600 subjects. That means you can be 95% confident the poll will be accurate to within +/- 4%. (Or, it will be accurate to with +/- 4%, 95% of the time.) If you want to increase the confidence level to 99%, you need 1,040 samples. That's just math, of course. If you want meaningful results, you have to design a good survey and execute it well. If you have a bad survey instrument or a bad sampling method, it doesn't matter how many people you sample, you're going to get a biased result.
Originally posted by Laykilla
In conclusion, these samples are propaganda. They'd like you to believe that since they are the experts that they must be giving you accurate data. It couldn't be anymore phony baloney.
And to prove it works, You even said "Is it accurate, Dunno but it probably is."
It's not even CLOSE.
To the contrary, it is impossible for it not to be an accurate sample.
Originally posted by Laykilla
Um, no -- it's impossible to be an accurate sample.
I never claimed that it was. You and the Op are both arguing about just how accurate/inaccurate this small sample size could potentially be, in reference to the eventual end results.
Originally posted by Laykilla
1,500 can never be representative of 311 million.
1% is better than 0.
Originally posted by Laykilla
When what you are testing is extremely variable, a sample size of even 1% is waaaaaaaaaaaaaaaaaaaaaay to low to get any kind of useful data.