I'll tell you how I read polls. As polling points themselves. That is to say, I look at them from afar as if each of them is a member of the electorate from which they attempt to extrapolate.
What this means is as I'm indulging in my daily two cups of coffee and dozens of news sites, I take note of the polling headlines and sort of keep track of the intimations. Granted, in web publishing today the headline or lede is often crafted as "eye bait" so I'm fairly careful to ignore the histrionics/wishful thinking and just take the general directions into account.
So, naturally, when there is consistently one outlier, consistently referenced by conventional thinking ouliers (climate deniers, science deniers, reality deniers, austerity fetishists, etc), it really makes me suspicious.
I'm not naive, and I fully understand how polls are used in politics. They are completely reversible. They are better than a Land's End vest in a Vermont Hunting season. If they support whatever it is you support that day, polls are great, polls are telling, and polls are completely reliable. If not, shit. It's just a poll. They probably just called their [lib/con] friends and asked leading questons or tricked them into giving specific answers with questions scientifically designed to manipulate the simple folk.
Given all that, I'm a little wary of polls. And I'm absolutely dismissive of Internet trolls that constantly dismiss every consistent poll to point to the perpetual outlier, which as we all know is Rasmussen.
Oh, yeah, and my favorite stratagem of the Rasumussen apologist: The creeping margin of error.
By the time this election is over, Barack Obama's lead of 28% nationally will dismissed as being entirely within the standard "margin of error" for polls. Which would be true, again, of course, for Rasmussen where the margin of error is only surpassed by the margin of partisanship which hovers at or near 100%.
So, if you're Newsvine/Internet Rasmussen apologist or weekly dial-in poll respondent, shut up already.