Quinnipiac: Clinton leads FL, PA, NC, Trump leads OH

[Gustaf]

These numbers are unrealistic when considered together.

Not at all.  It's clear that in this election OH has moved to the right and NC to the left relative to the average.  With that in mind, these results are reasonable assuming a Clinton national lead of 4 or so.  Polls are naturally going to come in on both sides of the mean (if they consistently don't do so, it's suspicious.)  Expecting polls from different states, even by the same pollster in the same time period, to be perfectly in line with each other is misguided.

Exactly, one needs to consider the MoE on multiple polls (about 4% in this case).

[pbrower2a]

These numbers are unrealistic when considered together.
They show that pollsters are unsure about the turnout model.

Since the coming election will be a low turnout election - essentially the war of the bases, the election results will depend mostly on whose voters show up to vote.

Due to this, it will be difficult to poll the electorate accurately and we will not be sure that even the final polls will give us the right picture.

Until now, Quinnipiac has been very friendly to Republican candidates. Without a switch in methodology, polls that don't look so good for Republicans may indicate a change, and not for the better. These polls may show a Trump collapse underway, and polls under that circumstance are unstable because the situation is itself unstable.

If this will be a low-turnout election, then which potential voters will fail to show up at the polls? If the stay-at-homes are Republicans, then Hillary Clinton wins and Republicans take a beating in the Senate races. 

[heatcharger]

Looks like Quinnipiac choked like a dog. Where are the senate numbers?