famousringo said:
If it was done properly, it could be quite accurate, actually. Most polls/surveys go for the 1000 person sample size. Google turns up a nice blog post on the issue: http://blog.bzst.com/2007/02/magical-sample-size-in-polls.html Of course, there are lots of mistakes which might have been made in the execution which might give inaccurate results, but a sample size of 1000 isn't one of them. |
True, but that assumes the sample is randomly selected, so as to best represent the populace. Here, we have a self-selected sample, which are almost always off. Dewy Beats Truman, anyone?