Inshallah.
This seems to me the only plausible way in which polling is somewhat broken, and may not have gotten enough traction in this thread:
The Shy Respondent and Propensity to Participate in Surveys: A Proof-of-Concept Study
Reading the above, the problem seems unfixable, and that we should subtract, say, four points from Harris’s numbers. Depressing for a Democrat!
However, the pollsters read studies like the one I linked above, and find weighting factors that help – without, of course, just adding points to Trump, which would be unethical. This is explained here:
Why election polls were so wrong in 2016 and 2020 — and what’s changing to fix that
My best guess: It still is the case that surveys overestimate general election support for a populist like Donald Trump, but not as much as in the previous two cycles.
P.S. to those who will not read my entire links, which I know are long: Populists are low propensity voters, explaining why you do not see the so-called shy Trumper effect as much, or at all, in off-year and special elections.
That was an interesting read. The study may be dated, however, because it could well be that the current post-Dobbs polling/election universe is just THAT different to what was happening four-to-eight years ago.
One thing your link touched on is that 2016/2020 polling often under-represented Trump support. This is purely anecdotal, but: on two other boards I frequent, there were Trump supporters bragging about how they lied to pollsters and represented themselves as voting for Clinton/Biden. Further, they said it was a common thing among Trump supporters and that via social media there was an ethic to urge other like-minded folks to do the same.
Completely within the margin of error, and totally a dead-heat race in the polls right now, but:
Kamara Harris now edges out Donald Trump in the RealClear Polling average. This is the first time Trump has been behind in this index since September 11, 2023.
I’ve memorialized the image below in this thread because the RealClear Polling link above will change with their next update.
Same RealClear Polling link as the above post:
The 7/22 NPR/PBS/Marist poll had Trump at +1 over Harris. Today’s NPR/PBS/Marist poll has Harris at +3, with a bigger sample to boot.
Today’s NPR poll got added to the RealClear Polling average. Still close and all … but Harris is now +0.5 in the aggregate – a 2.5 point move in seven days.
EDIT: Today’s NPR poll is also the first major poll where Harris has gotten over 50% of the respondents.
Largest lead in the RealClear Polling aggregate for the Democratic candidate since August 28 of last year.
RCP rolled off a few aged Trump-plus polls and added the Morning Consult’s huge sample-size survey.
However, only polls within battleground states matter, really.
RCP’s current averages in the big three are:
Michigan - Harris +2.4
Wisconsin - Harris +0.8 (effectively, a tie)
Pennsylvania - Trump +0.8 (effectively, a tie)
True, the trend has been in Harris’ direction in all three, but that momentum has to stall eventually. It may be stalling right now; if so, this will show up in the RCP averages in about ten days. Let’s hope all three states are at Harris +4.0 or better, by then. Join any of several well-organized get-out-the-vote letter-writing campaigns (or the campaign effort of your choice) for whichever of the three DOESN’T reach that level (at this point, Pennsylvania is looking most worrisome – but a convention bounce might be helpful there, and elsewhere).
I’ve deleted RCP as my polling bookmark. They rate a two star (95th rank) poll as reliable. It make Trump win PA in their average. Pollster Ratings | FiveThirtyEight
Going back to 538 after all.
Its my pet theory that the folks who answer polls these days are already pretty politically inclined. The relationship between people and polling is completely skewed. Polls are supposed to measure how people feel about a topic, and people are more or less performing for the pollster. Pundit-minded people, or people who think or fancy themselves to be are supposed to be measured by the poll influences themselves by the poll.
Anywhoot, polls are the thermometer, not the thermostat. People who think they know better screw the relationship up.
It’s mostly the biased Trafalgar Group that makes PA look small for Harris.Pennsylvania : President: general election : 2024 Polls | FiveThirtyEight
I don’t think your site says what you are implying. Just taking a crude average of the recent polls listed by 538 and excluding Trafalgar you get Harris up by less than half a point. Penn appears close because it is incredibly close. If you only look at highly rated posted you get almost the exact same numbers.
Yeah we’ve done that fact checking the claim before, including in the polls thread. The facts be damned apparently. It’s like they’re eating the dogs! People just keep saying it anyway.
We don’t see all the polls, we only see what we happen to find. What you see depends a lot on where you look. If you go to the Microsoft start page using Edge, you’re going to see a bunch of headlines- if it’s from Newsweek (now a right wing cesspit) or the Washington Examiner or the NY Post you’re going to think it’s all gloom and doom for the Democrats. These outlets consistently focus on the polls that show Republicans in the best light and imply that we’re on the verge of a red tsunami. Nate Silver is in my opinion a shadow of his former self. I think he’s more interested in keeping the race appear 50/50 so he can claim to be accurate in either event. 538 has gone down the tubes too, they seem to give weight to quite biased pollsters these days. We don’t see the parties’ internal polls, that’s probably the best picture available.
What do you base that assessment of 538 on? Hearing it repeated, or on some actual analysis of their numbers? In fact they down weight biased polling dramatically such that the top line number is virtually the same with or without them.
Why do you believe that internal polling has greater knowledge of what the actual voting population will look like than that of publicly available polling?
An hour ago according to Quinnipiac, TFG leads by 2 in both Michigan and Wisconsin, and that’s being reflected (at least over at Predictit) with the two of them basically in a dead heat now*.
- ETA : Well, actually, now he’s pulled ahead by a hair over at Predictit.
https://www.predictit.org/markets/detail/6867/Which-party-will-win-the-2024-US-presidential-election
I had no huge confidence based on polling, and I have no panic either.
Last week news: the polls call this close. This week’s the same.
Yeah, I forget which one of y’all it was that said in one of these parallel threads not too long ago, “don’t live and die by each daily poll fluctuation, it’s going to be a coin-flip from here on out regardless”, but I definitely need to repeat that to myself a few times.
Yes, this election is the closest one ever and America has never been this polarized before. Just like it was in 2020. And 2016. And 2012. And every other election since I’ve been old enough to vote.
Just a comment on Simon Rosenberg’s Hopium Chronicles … but I bought a jug of what this guy is pouring:
Travis:
Y’all : She is going to win MI. My entire family lives there - some urban, some rural. Vibes are good. MI GOP is a trainwreck. EV (= early voting - b) data is very good so far. AND the polling of MI has been generally, like most places, missing to the right. Whitmer was ahead by 4.5 in the AVG and won by over 10. ALSO…QPac polls always swing crazy bc their methodology basically guarantees it. They don’t weight for partisanship AT ALL, and use randomly generated phone numbers to get their pool…I don’t want to get into the methodology weeds because I’m actually not a data person, but they have insane swings every cycle, but particularly this one. I also strongly recommend looking at [Tom Bonier’s] takes on EV voting more than polling over the next few weeks - I know its still too early to draw conclusions but so far its been very good for us.
FWIW I remain very optimistic. Polls just only so useful. They aren’t broken but working as well as they can they aren’t able to be accurate enough either.
Almost much all the not polling indicators are positive though in my read (voter registration trends exception noted). Proxies for enthusiasm. The fact that she is a stronger candidate than Biden 2020 was and Trump today a weaker one than Trump 2020 was, and Biden won. The ground game. Funding. So on.
Given that polling is not able to tell us much right now, I’m going with those facts as a basis for belief.