The questions are in the link provided above.Cluck U wrote:What were the questions?
I've done polling...questions can be easily skewed to produce desired results.
The error is controlled - hence results providing a "margin of error."Cluck U wrote:And smelly...sampling is often inaccurate.
The poll used ~1,200 people. That's pretty typical for nationwide surveys. It provides a confidence interval of +-3% 19 times out of 20 (95%). (The average sample size of polls Nate Silver used in his 2012 election forecasts was 900).
Yes, you could sample a lot more people - 10,000 - and your margin of error would drop to just under 1% at 99% confidence. But, that's incredibly rare in this kind of polling. 95% confidence is the standard in science.
Again, if you have a better poll that uses something like 10,000 respondents, feel fee to post.



