FL/OH/PA Quinnipiac: Close races everywhere (Trump +4 in OH, Clinton +1 in FL/PA

[Virginiá]

So do these polls (meaning all Florida polls) take into account the recent influx of Puerto Ricans into the state?

Well, they have a whiter electorate than 2012 everywhere, which is suspect but plausible if Trump supercharges turnout with downscale white men.

I don't really understand how that is possible. Demographic changes are much more well-studied than this poll. They can't know what the final turnout will really look like, but the demographic makeup of the eligible voter electorate is well-understood. Even with higher white voter turnout this year, to be whiter than 2012 would take a BIG increase in white turnout - Something they cannot predict right now.

[Virginiá]

The increase by 2-3 points of the non-white share of voters in Presidential elections since 1992 has been quite consistent.  It's something that has transcended individual elections.  I know Trump is a unique candidate, and perhaps he is energizing racist and blue-collar whites, but he's also energizing Latinos and other minority groups.  I think those factors will be a wash, and we'll see another 2-3 point increase.

Yes. If this poll is predicated on the notion that the 2016 electorate will be more white than 2012, then it is dead wrong. Not only was 2016 widely understood to be 2%~ less white than 2012 (if not more) no matter who was running, the Trump effect would create a situation where higher minority turnout could actually make the electorate even less white than expected.

At the very least, 2016 will definitely reflect ongoing demographic trends, not a rollback to pre-2012. I'm sure we'll see more legitimate polls with close numbers like this, but the demographics have to be right, otherwise it is inaccurate.