FL/OH/PA-Quinnipiac: Trump leads or tied in all three states

[Reginald]

lol the first thing I did was check the PA partisan topline, because:

QU is routinely undersampling Democrats in PA and I don't know why. Registration statistics and the 2012 exit poll show about a ten point margin for Democrats, but this poll has a (weighted!) four point margin, last month's had a five point margin, and the one in April had a Republican plurality.

And lo and behold, Dems have a one-point advantage over Republicans for this (weighted) sample. Baffling.

[Reginald]

In 2012, Mitt Romney and his supporters kept criticizing polls on the basis of oversampling, just as you are doing now. Be careful.

During the primaries, the GOP turnout hit records, surpassing the Democratic primary's turnout; and Trump is getting all of the media attention. Why, then, is it so unreasonable to expect that Republicans will turn out at a much higher rate than they did 4 years ago?

Were they using prior data at all in their criticisms? Didn't keep up with the minutia of their attack of the polls. I just looked at the PA SOS, and the split in party reg as of today is 49-38 in favor of Dems. Maybe I'm crazy, but I don't think a ten-point advantage for Democrats is going to evaporate in the span of an election cycle, enthusiasm gap or not. Wouldn't you say it's a little presumptuous of a pollster to make such a shift in the weights, based off a guess?

[Reginald]

PA is one of the few states that have people that would be considered "Reagan Democrats" where they are technically Democrats, but would self-identify as Republicans here.

I think the crossover in the primaries was over 100,000 in PA.

I do think Trump +6 is overstated here, but I am not shocked that the state is competitive.

Well, leaving aside a discussion of Reagan Democrats, Republicans are also undersampled here -- less so, but still. It's Independents who are extremely overrepresented, but I suppose that's inevitable in a "Do you consider yourself" sort of a question. Don't get me wrong, maybe Republicans will turnout at a higher rate. But there's a built-in 900,000 voter advantage for Democrats they have to work around, and the partisan crosstab here is about equally polarized. It's winnable for sure, but I question how representative the sample is.