September 19, 2024: Someone is still going up ... pic.twitter.com/bZCFrPDlXJ

— Rasmussen Reports (@Rasmussen_Poll) September 20, 2024

📊 National poll by RMG Research

🟦 Harris: 50% [-1]
🟥 Trump: 48% [+1]

[+/- change vs 9/9-12]
——
Trends
Aug. 28 - 🔵 Harris +3
Sept. 5 - 🔵 Harris +2
Sept. 12 - 🔵 Harris +4
Sept. 19 - 🔵 Harris +2
—
#60 (2.3/3.0) | 2,969 LV | 9/16-19 | ±1.8%https://t.co/e1c0T3z1Qt pic.twitter.com/7dL2gqgYyL

— InteractivePolls (@IAPolls2022) September 20, 2024

Hungry Ojos said:

---

Captn_Ag05 said:

---

Twelve national A/B-rated post-debate polls are in:

🔵 ABC/Ipsos: Harris +6
🔵 Reuters/Ipsos: Harris +5
🔵 Yahoo News/YouGov: Harris +4
🔵 TIPP: Harris +4
🔵 Times/YouGov: Harris +4
🔵 Data For Progress: Harris +4
🔵 Angus Reid: Harris +4
🔵 Economist/YouGov:…

— Political Polls (@PpollingNumbers) September 19, 2024

---

Yeah, this is really bad. I don't think we can just hand waive away the fact that ALL major polls have her winning, by a comfortable margin. Wish I had y'all's optimism, but by all accounts, Trump is losing.

---

I can understand your pessimism on an emotional level, but "by all accounts" Trump is not losing. No matter how you choose to interpret the polls Trump is doing better than 2016, and considerably better than 2020. This is essentially an even race nationally, which advantages Trump slightly in the electoral college.

... and speaking of Pennsylvania,

PSA: here's the track record of pollsters currently in the RCP Average in PA. Over the last 2 prez cycles, the "high-quality pollsters" cited below had an average error of 6-8% in favor of Dems. Maybe they'll get it right this year. We'll know on Nov 6.https://t.co/fI20xUIEfS pic.twitter.com/NgZQlW739c

— Tom Bevan (@TomBevanRCP) September 20, 2024

Legal Custodian said:

---

Legal Custodian said:

---

According to RealClearPolitics average, Kamala's post-debate bump in polls is at 1.5% right now.

Two weeks leading up to the debate they had Kamala with a 1.1% lead
The ten polls on RCP post-debate she has a 2.6% lead

So that's a 1.5% bump. For reference, in the two weeks following Trump/Biden first debate 4 years ago where Biden was considered the winner, he got a 3.1% point bump and he carried that bump for about 3 weeks until the next debate. After that last debate the polls got a tad bit tighter but still had Biden with a 7% lead going into Election Day.

The first 5 polls right after the debate, Kamala had a 4.2% lead according to the average.

The most recent 5 polls , Kamala has a 1% lead.

So one can assume the initial wave that Kamala got from the debate was very short lived.

In my opinion, the populace watched the debate and were turned off again by Trump being his normal self. But after a week that wore off and people went back to the thinking of "Am I better off now than when I was under a Trump Presidency". Again, that's just my opinion trying to analyze the data.

---

I guess something changed in the past hour from when I posted the post-debate numbers. I just checked again cause I'm a nerd and now there are 11 post-debate polls but I can't tell which one was added.

Now the numbers are:

Pre-debate: Kamala +1.1%
Post-debate: Kamala +2.2%

First week after debate: Kamala +2.6%
Most recent 5 polls: Kamala 1.6%

Looks like one was added recently but not sure which one.

---

the high quality polls showed. O bounce. The manipulated polls using online panels and internet querying adjusted their numbers (urban/rural, race, wealth, region) to achieve the desired results for a meager post debate bounce. It was to create a false narrative.

agsalaska said:

---

The best forecasters in the business are calling this a statistical tie at this point.

That has nothing to do with optimism or slant or anything else. It is a tie ball game late in the 3rd quarter.

---

The best polling operations are actually saying it's a high probability that Trump wins. Trafalgar, Rasmussen, AtlasIntel, Nate Silver.

Low quality polls - based upon their accuracy in the past national elections - Fox, WP, Quinnipiac, Emerson, etc. these polls are pushing narratives.

That's not what I saw when looking at the raw data, but I'll take your word for it.

WestAustinAg said:

---

agsalaska said:

---

The best forecasters in the business are calling this a statistical tie at this point.

That has nothing to do with optimism or slant or anything else. It is a tie ball game late in the 3rd quarter.

---

The best polling operations are actually saying it's a high probability that Trump wins. Trafalgar, Rasmussen, AtlasIntel, Nate Silver.

Low quality polls - based upon their accuracy in the past national elections - Fox, WP, Quinnipiac, Emerson, etc. these polls are pushing narratives.

---

So will leave some of that alone,

But Nate Silver has it as a tie. May want to check the rest.

agsalaska said:

---

WestAustinAg said:

---

agsalaska said:

---

The best forecasters in the business are calling this a statistical tie at this point.

That has nothing to do with optimism or slant or anything else. It is a tie ball game late in the 3rd quarter.

---

The best polling operations are actually saying it's a high probability that Trump wins. Trafalgar, Rasmussen, AtlasIntel, Nate Silver.

Low quality polls - based upon their accuracy in the past national elections - Fox, WP, Quinnipiac, Emerson, etc. these polls are pushing narratives.

---

So will leave some of that alone,

But Nate Silver has it as a tie. May want to check the rest.

---

It's complicated but Nate Silver is saying that though the GE polls show a tie race nationally, it really means Trump wins because it isn't a national election. It's a state by state election. And Harris needs to win the national election popular vote by 5-7 points to have a chance at the electoral college election.

#NEW Nate Silver 2024 election model update

WIN ODDS:
🔴 Trump: 51% (+2.2)
🔵 Harris: 48.8%

Silver Bulletin | Sept. 18 pic.twitter.com/NCjUrHWQ8V

— Eric Daugherty (@EricLDaugh) September 19, 2024

HIs update today has Harris the favorite. SIlver's "model" is not a poll. He feeds national and state polls into his model and weighs them based on the quality he assigns to them. He also puts in other metrics like unemployment rate, stock market, etc. but has not really explained fully what factors he uses or how they are factored in.

You're correct. It isn't polling. I thought it was. It's a model that he sells to candidates and parties. And evidently he uses some less than respectable polls in his evaluative modeling like Morning Consult. And it just flips by about 10 points from last week which is pretty much an impossibility when it comes to a National election.

WestAustinAg said:

---

agsalaska said:

---

WestAustinAg said:

---

agsalaska said:

---

The best forecasters in the business are calling this a statistical tie at this point.

That has nothing to do with optimism or slant or anything else. It is a tie ball game late in the 3rd quarter.

---

The best polling operations are actually saying it's a high probability that Trump wins. Trafalgar, Rasmussen, AtlasIntel, Nate Silver.

Low quality polls - based upon their accuracy in the past national elections - Fox, WP, Quinnipiac, Emerson, etc. these polls are pushing narratives.

---

So will leave some of that alone,

But Nate Silver has it as a tie. May want to check the rest.

---

It's complicated but Nate Silver is saying that though the GE polls show a tie race nationally, it really means Trump wins because it isn't a national election. It's a state by state election. And Harris needs to win the national election popular vote by 5-7 points to have a chance at the electoral college election.

#NEW Nate Silver 2024 election model update

WIN ODDS:
🔴 Trump: 51% (+2.2)
🔵 Harris: 48.8%

Silver Bulletin | Sept. 18 pic.twitter.com/NCjUrHWQ8V

— Eric Daugherty (@EricLDaugh) September 19, 2024

---

Ok. That's actually not what he is saying. But ok.

Never mind. I'll move on to the next topic.

WestAustinAg said:

---

You're correct. It isn't polling. I thought it was. It's a model that he sells to candidates and parties. And evidently he uses some less than respectable polls in his evaluative modeling like Morning Consult. And it just flips by about 10 points from last week which is pretty much an impossibility when it comes to a National election.

---

the wild variation in silver's model is mostly because it has PA as the deciding state, and over the last week or two the polls feeding the model have tipped from trump to harris

Old McDonald said:

---

WestAustinAg said:

---

You're correct. It isn't polling. I thought it was. It's a model that he sells to candidates and parties. And evidently he uses some less than respectable polls in his evaluative modeling like Morning Consult. And it just flips by about 10 points from last week which is pretty much an impossibility when it comes to a National election.

---

the wild variation in silver's model is mostly because it has PA as the deciding state, and over the last week or two the polls feeding the model have tipped from trump to harris

---

I mean I guess

Old McDonald said:

---

WestAustinAg said:

---

You're correct. It isn't polling. I thought it was. It's a model that he sells to candidates and parties. And evidently he uses some less than respectable polls in his evaluative modeling like Morning Consult. And it just flips by about 10 points from last week which is pretty much an impossibility when it comes to a National election.

---

the wild variation in silver's model is mostly because it has PA as the deciding state, and over the last week or two the polls feeding the model have tipped from trump to harris

---

There was a post that clarified this a minute ago. Silver's modeling goes beyond polling and includes many other variables including enthusiasm, mail in ballots, registrations, economic data, etc. The guess is that the change had something to do with economic news as much as anything else. I would guess that to be true and more of a factor than just PA.

The SilverBulletin is the most respect aggregator and predictor on this thread.

WestAustinAg said:

---

agsalaska said:

---

The best forecasters in the business are calling this a statistical tie at this point.

That has nothing to do with optimism or slant or anything else. It is a tie ball game late in the 3rd quarter.

---

The best polling operations are actually saying it's a high probability that Trump wins. Trafalgar, Rasmussen, AtlasIntel, Nate Silver.

Low quality polls - based upon their accuracy in the past national elections - Fox, WP, Quinnipiac, Emerson, etc. these polls are pushing narratives.

---

This, except that Nate Silver doesn't actual poll.

Quote:

---

For most of the rest of the media/university polls you will see that their structural Democrat bias is pretty comparable as well, including the ones that heavily utilize the Ipsos panels.

---

I've been part of Ipsos surveys for a few years now. I only get a couple a week but I let them observe my laptop for extra money, it's my old work laptop and pretty much only is on texags (but they don't know that lol).

It earns me around $100 a month so it's a nice little way to earn extra income and it's not time consuming at all since I only get a couple of surveys a week.

Anyway, I used to get quite a few surveys on who I would vote for, especially for the 2022 cycle and I even got some earlier this year when it was still a Biden-Trump race. Since Kamala was announced I've received exactly ZERO surveys asking me who I would vote for in the upcoming election.

They know I've always listed myself as very conservative and Republican. I don't even list myself as a "strong" Republican since I feel my values are leaning more towards Libertarianism so you'd think they would want to poll me, especially since they've done it a lot in the past.

It's almost as if they aren't polling survey members who wouldn't say they are voting for Kamala.

Prosperdick said:

---

Quote:

---

For most of the rest of the media/university polls you will see that their structural Democrat bias is pretty comparable as well, including the ones that heavily utilize the Ipsos panels.

---

I've been part of Ipsos surveys for a few years now. I only get a couple a week but I let them observe my laptop for extra money, it's my old work laptop and pretty much only is on texags (but they don't know that lol).

It earns me around $100 a month so it's a nice little way to earn extra income and it's not time consuming at all since I only get a couple of surveys a week.

Anyway, I used to get quite a few surveys on who I would vote for, especially for the 2022 cycle and I even got some earlier this year when it was still a Biden-Trump race. Since Kamala was announced I've received exactly ZERO surveys asking me who I would vote for in the upcoming election.

They know I've always listed myself as very conservative and Republican. I don't even list myself as a "strong" Republican since I feel my values are leaning more towards Libertarianism so you'd think they would want to poll me, especially since they've done it a lot in the past.

It's almost as if they aren't polling survey members who wouldn't say they are voting for Kamala.

---

Thanks for volunteering this info, as it highlights one of the big problems with most of the media polls.

If the poll is too dependent, much less primarily dependent, upon opt-in panels (that are intentionally comprised of overeducated/over-credentialed respondents compared to the electorate), you're essentially polling the same exact people over and over again and you are going to get intentionally unrepresentative results.

Old McDonald said:

---

WestAustinAg said:

---

You're correct. It isn't polling. I thought it was. It's a model that he sells to candidates and parties. And evidently he uses some less than respectable polls in his evaluative modeling like Morning Consult. And it just flips by about 10 points from last week which is pretty much an impossibility when it comes to a National election.

---

the wild variation in silver's model is mostly because it has PA as the deciding state, and over the last week or two the polls feeding the model have tipped from trump to harris

---

He added a very biased poll by morning consult all if a sudden. This polling company has really bad history of inaccuracies.

That’s only bc you added morning consult which is the worst pollster of 2020. This will be back and forth. I’ve never seen Trump at toss up so all of this is amazing to us lol

— Brittany Rae (@legitbrittFLA) September 20, 2024

this comment is funny because i see libs complaining constantly about nate's inclusion of "low quality republican polls" in his model.

silver's model does not weight all polls evenly. each one is adjusted for pollster partisan bias and quality. its an imperfect model like all the others.

Turnout was high across the board in VA … but particularly high in HEAVY R counties 👀 👀

All over 200%! https://t.co/5Tw9AzkBHV

— Eric Daugherty (@EricLDaugh) September 20, 2024

sometimes it cuts both ways, several of the R-lean pollsters that were accurate in 2020 over-predicted the 2022 red wave. in silver's model, pollster quality has to do with statistical and sampling methodologies as well as how accurate they were in the past.

Nm

🚨 In Georgia, there are currently 85% less mail-in ballot requests than at this same point 2020. pic.twitter.com/sTlAjOVNXf

— Eric Daugherty (@EricLDaugh) September 20, 2024

🚨 #NEW North Carolina 2024 election mail-in requests show a clear difference from 2020...

🔵 Democratic: 37.7% [-11.3]
🔴 Republican: 23.2% [+5.2]

[+/- from Sept. 2020]

Data: @WinWithJMC & @MichaelPruser

— Eric Daugherty (@EricLDaugh) September 21, 2024

Nate Silver has made a nice living making predictions to the tenths place based on underlying data that is +/- 3%

Honest question - weren't ballot requests hugely in favor of Rs in the midterms?

sam callahan said:

---

Honest question - weren't ballot requests hugely in favor of Rs in the midterms?

---

Yes, and they were +3%.

will25u said:

---

🚨 In Georgia, there are currently 85% less mail-in ballot requests than at this same point 2020. pic.twitter.com/sTlAjOVNXf

— Eric Daugherty (@EricLDaugh) September 20, 2024

---

Separate thread for turnout/early vote discussion:
https://texags.com/forums/16/topics/3488894

I know this was reported yesterday but the comparison to 2020 in the belly of the beast/DC area is quite amazing.

🚨 VIRGINIA poll comparison, September 2020 vs 2024 - University of Mary Washington

🔵 SEPT. 2020: Biden+17
🟡 SEPT. 2024: Harris+1

👀👀

— Eric Daugherty (@EricLDaugh) September 20, 2024

I want this to be true, and while this is just a small academic type poll, they are in no way likely biased in favor of over-sampling pro-GOP segments, imho. Also, impact on RCP status;

Virginia officially enters the sub-5 point range in RealClearPolitics… which would turn VA from “likely” to “LEAN” D… https://t.co/BgB1BLnFfU

— Eric Daugherty (@EricLDaugh) September 21, 2024

It's a small poll, BUT, state universities polling their own state are normally pretty accurate.

Agreed. Ohio looks to likely tighten into a dead heat (senate race):

Quote:

---

The latest RealClearPolling Average shows Brown with a 3.6 point lead over Moreno, a millionaire Cleveland businessman campaigning on a by-his-bootstraps immigration story despite his family's wealth and political connections; the Decision Desk HQ and The Hill's polling average places him just 3.2 points ahead, values that fall within the polls' margins of error. The Cook Political Report has labeled the race a "toss up."

The race can be expected to tighten in the weeks ahead, effectively placing the candidates in a "dead heat," according to Jacob Neiheisel, a University at Buffalo professor of political science and Ohio native.

---

Separately, pretty amusing take down of one poll from PA:

Quote:

---

PA is one of the easiest states in America to poll because we register by party and we have a relatively small percentage of non-major party voters.

For years PA elections have roughly been some variation of independents between 9-12% of the vote. They will not under any circumstances be 27% of the vote. I hope to hear from someone at SpotlightPA/MassINC about it.

However, that's not the major issue. They missed the geographic quotas in PA in a HORRENDOUS fashion assuming their naming conventions in the publicly released docs are what I think they are.
If "Philadelphia" above means the city/county of Philly then they should have gotten 9-10% NOT 17%. I think that's really the only way to read it since there is a separate "South East" category that I assume is the rest of the media market.

My best guess is this error alone gives democrats about a 2.5 point boost on the statewide ballot.
I am assuming Northeast / Central is the Wilkes-Barre and Harrisburg media markets which they would be under sampling by 4 points.

---

The trash poll in question:

#New General election poll - Pennsylvania

🔵 Harris 52% (+5)
🔴 Trump 47%

Senate
🔵 Casey 49%
🔴 McCormick 42%

MassINC #A+ - 800 LV - 9/18

— Political Polls (@PpollingNumbers) September 22, 2024

Again, not all pollster rankings are…accurate.

🇺🇲 National poll by CBS/YouGov

🟦 Harris: 52% [+1]
🟥 Trump: 48% [=]
—
Trends
July 18 - 🔴 Trump +3
Aug. 2 - 🔵 Harris +1
Aug. 16 - 🔵 Harris +3
Sep. 20 - 🔵 Harris +4
——
Battlegrounds
🟦 Harris: 51% [+1]
🟥 Trump: 49% [-1]

[+/- change vs 8/4-16]
—
#4 (2.9/3.0) | 3,121 LV |… pic.twitter.com/WOJoLspzEx

— InteractivePolls (@IAPolls2022) September 22, 2024

CBS News Battleground-state estimates

Nevada - 🔵 Harris 51-48%
Wisconsin - 🔵 Harris 50-48%
Michigan - 🔵 Harris 50-48%
Arizona - 🔵 Harris 50-49%
North Carolina - 🔵 Harris 49-48%
Pennsylvania - 🟡 Tie 49-49%
Georgia - 🔴 Trump 50-49%
NE-2 - 🔵 Harris 51-47%

Estimates updated… pic.twitter.com/lpZzwUPhpl

— InteractivePolls (@IAPolls2022) September 22, 2024

Question about methodology: how do the polls determine if someone is a likely voter or not? Is it as simple as asking "are you registered to vote, and how likely are you to vote in the upcoming election?" Do they do any due diligence to check responders' history of participation?