Huffpost has the final prediction as a 0.8% advantage for Obama, so obviously their model was wrong. Obama won the popular vote by 4%.
The ongoing 538 projections were more accurate, figuring Obama for about a 3-point margin; Sam Wang was quite a bit more off, giving Obama only a two-point advantage and incorrectly calling Florida for Romney.
Although the fact that it is Marist and that Florida pretty much unchanged from their last poll (others not polled by Marist before) does reassure …
Is it just me or do polls seem to have more variation about the mean this time? Quinnipiac and Marist are both reputable polling houses (538 rated as A- and A respectively and both with mild R leans) yet they are running 10 to 11 points different in multiple states. Both RV polls. Why?
More than last time same stage I don’t know. But as for reason the partisan article linked above calling the Q polls ‘rubbish’ was factual in one sense. Quinnipiac is making a determination that white turnout will be higher than 2012. The article called that ‘rigging the poll’ and not giving the ‘real numbers’ but I assume Q has a reason for believing that, which might be valid, and their polls do not have a record of being skewed significantly to Republicans in the past.
There aren’t ‘real numbers’ for turnout by group in elections which haven’t happened yet. It’s varied in the past, and some of ‘trend to higher minority turnout’ could be simply Obama.
Today’s new national polls: CBS/NYT (538 rating: A- D+0.06) tied; Rasmussen (C+ R+2) Trump +7, LV others are RV, it is valid to say Rasmussen’s polls have been R friendly but wow; YouGov (B, D+1.6) Clinton plus 2. Anybody can speculate as they like and say no f’ing way this or that future outcome, but based on facts as available it’s a close race.
Rasmussen’s polls are not only R-friendly but they seem to have a habit of having huge +R outliers early in the cycle and then contracting the closer you get to election day so the final results aren’t as mind-bogglingly off as the initial polls.
Had not actually read the article but I guess I should not be surprised. So RV polls are still guessing likely voting demographics in how they weight their populations, not just weighted to RV demographics … huh.
It has of course been the big issue in my mind re polling results: the question of who actually turns out this time. Does Trump actually get non-college educated Whites, especially males, to get off the butts and vote more than they have in the past? Or do they stay home again? Black voter turnout though we can state with some confidence was NOT just “simply Obama.” See here. It was pretty much rising linearly from 1996 to 2008 (from 53 to 65%) and actually flattened in 2012. Will Hispanic voters turnout more given Trump? So on.
This is a petty ridiculous thing, I know. But is anyone else annoyed by how long Trump wears his tie? I just saw a picture of him with Pence, and the end of his tie had to be hanging three or four inches below his belt. I know, I know. It’s a dumb thing to nitpick and ties are dumb. But still. The point is supposed to be right at your belt line, dammit!
Normally, a candidate gets a bounce in the polls after announcing a running mate.
It seems to me that Trump is unlikely to get much of one, since his announcement is massively overshadowed by the tragedy in Nice, and also because his running mate is pure white-bread milquetoast, unlikely to appeal to any but a specific set of white, evangelical “Christians.”
Interestingly, the letter sent identified a number of companies that the RNC is saying backed out, and some of those company’s are fighting their being included in this list.
