= new reply since forum marked as read
Is that the sample size is in the 10's and 100's of thousands. People really can't "DU" that poll... as all of them that I saw only permitted one vote. It's easily done by altering the cookie stored by the site on your computer or else (as in the case of Facebook) requesting you log in or else noting your IP address.
In the end, a self-selected poll still does say exactly what it means to, even if you don't at first sight realize exactly what it's telling you.
For instance... the CNN/Facebook poll will tell you the opinions of anyone who watched the debate and saw the referral over to the website to vote, and was motivated to, and had a Facebook or at least knew how to find the link. Obviously, groups such as the blind for obvious instance would be vastly under represented in an online poll. The thing is though, after a certain number of different polls from different sources, with different types of audiences ALL come to about the same conclusion... that Bernie beat Hillary approx. 80 to 17 (with the other three splitting up the rest), the group of polls does become increasingly representative of the general population.
In fact it is impossible to actually know the true representativeness of a poll in regards to the general population without actually having the entire general population tell you how they feel... You can get closer to a representative group... the more you add people to it randomly (as groups like PPP do, though in reverse, they randomly remove people until the poll is significantly "random"
.... But you can never reach exactly the correct representation. (Hence the existence of Margin of Error). However, after a while it's a moot point.
You don't for instance need to know that it's 36 miles and 22 feet to a specific location. The simple 36 miles will do.
With the multitude of polls ALL showing uniformly the same thing and at the same time drawing from different populations of voters, it is fairly certain that the polls are representative of the population. Inversely it is also fairly certain that the absence of ANY poll showing Hillary as the winner is representative of the fact that she wasn't.
