> You're depending on 1018 people who actually may not be statistically relevant, since America is a big, big country, to drive your thinking on that country.
It's a problem if they're badly sampled, but the sample size you need is doesn't really change as you increase your population, as long as your population is a lot bigger than your sample size. If 1000 people would be OK for a population of 1M, then it should be just as good for a population of 100M.
I don't think this really invalidates your point, it's just a misconception I see a lot.
It's a problem if they're badly sampled, but the sample size you need is doesn't really change as you increase your population, as long as your population is a lot bigger than your sample size. If 1000 people would be OK for a population of 1M, then it should be just as good for a population of 100M.
I don't think this really invalidates your point, it's just a misconception I see a lot.
http://www.surveysystem.com/sscalc.htm