I don't trust any of the political polls completely. I've taken quite a few courses in advanced statistics and I know how poll results can vary based on who is sampled, where they live, their demographics, the size of the sample, when they are sampled, and many other factors. Probably the biggest factor effecting the accuracy of political polls is the time between when the poll is taken and when the election is actually held. Lots of people can change their minds. Political polling is done so frequently because it's like shooting at a moving target.

I generally watch polls run by major news organizations or professional market research companies. But I accept none of them individually. Rather, I try to get the "drift" of things. The better ones come up with results reasonably close to one another, but all have a margin or error that would win or lose a close election. (That's lousy polling by the way. It's done that way because it would be way too expensive to design and conduct a poll with a 1/2% or less margin of error.)

A properly designed poll of a proper statistical size should not have a "margin of error" of 3-4%. Political polls measure a moving target, of course...how voters feel on a given day long before the election about a candidate or a question. If you wanted to actually pick a candidate that was the true statistical choice of the majority of the country, or state, you could do it without actually holding an election. Probably with less error than actually occurs in polling places and counting votes throughout the country. Remember the hanging chads? The key variable there would be sample size and conducting the poll at pretty much an exact point in time. But of course, none of the candidates or political parties would accept an election result based on a poll run by even the best, most apolitical statisticians anyone could find. I guess on election day, that's why you vote in polling places.

When a poll like Newsmax comes out with a plurality of 65% for one candidate when all the other polls are showing it to be a close election, there's no sense even considering it. It was either ill-conceived or the polling was purposely done to achieve a certain result. I suspect that was true of the Newsmax poll. In that I don't read Newsmax (nor do I read The Daily Kos), I have no idea how their poll tracks with their editorial agenda. But I can guess.

That's my point, Bucco.