Dude, Redfield is basically for humor. But even if you decided to alternate reality and believe it, she's incredibly unlikely to win the battleground states at national plus two.

According to Rasmussen's head pollster the Morning Consult is a D+10 sample

What appears to be happening is that the race is settling down to where it has always been (right before the Trump-Biden debate), a slight Trump advantage.

Most of the ebbs and flows in the polling since the Trump-Biden debate are largely response-bias blips, created by a combination of media narratives and temporary bumps in enthusiasm for one side or the other.

As it turns out, the swapping of Biden for Harris hasn't changed the race very much (from where it was prior to Biden's debate collapse), which actually makes sense as Democratic voters aren't really voting for a candidate anyway... they're voting for an avatar.

Quo Vadis? said:

---

According to Rasmussen's head pollster the Morning Consult is a D+10 sample

---

LOL, which is only 13 points off of the current D/R split according to Gallup (which is R +3), and I would bet their Republican samples are heavily weighted with over-educated and over-credentialed voters, which are the easiest/cheapest to poll, but the least "Trumpy" Republican voters.

1836er said:

---

Quo Vadis? said:

---

According to Rasmussen's head pollster the Morning Consult is a D+10 sample

---

LOL, which is only 13 points off of the current D/R split according to Gallup (which is R +3), and I would bet their Republican samples are heavily weighted with over-educated and over-credentialed voters, which are the easiest/cheapest to poll, but the least "Trumpy" Republican voters.

---

A 6 point Kamala lead with a moe of 1% is just contrived. They're not even hiding it.

Since "keeping receipts" is a hot cliche recently, I hope people are doing just that with some of these polls trying to steer a narrative versus reporting reality.

Quo Vadis? said:

---

According to Rasmussen's head pollster the Morning Consult is a D+10 sample

---

been saying this morning consult is trash keep getting taken for gospel around here

📊 TEXAS GE: FAU/@MainStUSApolls

🟥 Trump: 50%
🟦 Harris: 45%
🟪 Other: 2%
—
Senate
🟥 Cruz (inc): 47%
🟦 Allred: 44%

#86 (2.0/3.0) | October 2-6| 775 LVhttps://t.co/rdP7Rdw2tQ pic.twitter.com/PFpP5solrn

— InteractivePolls (@IAPolls2022) October 8, 2024

Based on the polling right now, does it look like Trump wins the house and senate also?

Captn_Ag05 said:

---

📊 TEXAS GE: FAU/@MainStUSApolls

🟥 Trump: 50%
🟦 Harris: 45%
🟪 Other: 2%
—
Senate
🟥 Cruz (inc): 47%
🟦 Allred: 44%

#86 (2.0/3.0) | October 2-6| 775 LVhttps://t.co/rdP7Rdw2tQ pic.twitter.com/PFpP5solrn

— InteractivePolls (@IAPolls2022) October 8, 2024

---

I don't know how accurate that is, but being Texas that is not very comfortable to me.

All of the polls on RCP have been pretty consistent for the past couple of months when it comes to Texas polling.

Trump won Texas by 5.8% over Biden in 2020 just for reference.



I mentioned this when it happened, but I did get to see the Cruz campaign internals from the second week of September. Trump was up around 8 and Cruz around 3-4 at that time. I think that is likely where we still are.

#NEW PENNSYLVANIA poll

🔵 Harris: 50% (+4)
🔴 Trump: 46%

Bullfinch Group | Sept. 26-29 | N=800RV

— Eric Daugherty (@EricLDaugh) October 8, 2024

#NEW PENNSYLVANIA SENATE poll

🔵 Casey: 52% (+10)
🔴 McCormick: 42%

Bullfinch Group | Sept. 26-29 | N=800RV

— Eric Daugherty (@EricLDaugh) October 8, 2024

This was done for a Republican group, so not thrilled with these results.

#NEW PENNSYLVANIA poll

🔵 Harris: 50% (+1)
🔴 Trump: 49%

Research Co. | Oct. 5-7 | N=450LV

— Eric Daugherty (@EricLDaugh) October 8, 2024

#NEW PENNSYLVANIA SENATE poll

🔵 Casey: 51% (+3)
🔴 McCormick: 48%

Research Co. | Oct. 5-7 | N=450LV

— Eric Daugherty (@EricLDaugh) October 8, 2024

📊 ARIZONA poll by Fabrizio Ward (R) & Impact Research (D) for AARP

2-WAY
🟥 Trump: 50%
🟦 Harris: 48%
---
FULL FIELD
🟥 Trump: 49%
🟦 Harris: 47%
🟩 Stein: 1%
——
AZ Senate
🟦 Gallego: 51%
🟥 Lake: 44%
——
#140 (1.7/3.0) | 600 LV | 9/24-10/1 | ±4%https://t.co/UcCz81PKlp pic.twitter.com/G4IALqbmIL

— InteractivePolls (@IAPolls2022) October 8, 2024

Bullfinch was the group that put Harris down 1 in FL a couple weeks ago.

#NEW MICHIGAN poll

🔵 Harris: 51% (+3)
🔴 Trump: 48%

Research Co. | Oct. 5-7 | N=450LV

— Eric Daugherty (@EricLDaugh) October 8, 2024

#NEW MICHIGAN SENATE poll

🔵 Slotkin: 52% (+5)
🔴 Rogers: 47%

Research Co. | Oct. 5-7 | N=450LV

— Eric Daugherty (@EricLDaugh) October 8, 2024

#NEW WISCONSIN poll

🔵 Harris: 50% (+2)
🔴 Trump: 48%

Research Co. | Oct. 5-7 | N=450LV

— Eric Daugherty (@EricLDaugh) October 8, 2024

#NEW WISCONSIN SENATE poll

🔵 Baldwin: 52% (+5)
🔴 Hovde: 47%

Research Co. | Oct. 5-7 | N=450LV

— Eric Daugherty (@EricLDaugh) October 8, 2024

Something to keep in mind about Wisconsin is that it is typically the hardest of the rust belt states to poll... and it typically polls 5 or more points higher for Democratic presidential candidates than actual results.

On this day (October 8) in 2016 and 2020, for example, the RCP average polling for Wisconsin was Biden +5.5% in 2020 and Clinton +5.7% in 2016.

Compared to the actual results, Biden "won" Wisconsin in 2020 by +0.7%, and Trump won Wisconsin in 2016 by +0.7%, which means the polls in Wisconsin at this point in the election understated Trump's support compared to the actual election results by 4.8% in 2020 and by 6.4% in 2016.

If we split the difference between their 2016 and 2020 polling 'misses' in Wisconsin, and apply that number to where Trump is in the RCP today (October 8, 2024, Harris +0.8%), he is actually up by 4.8% in Wisconsin, assuming a comparable polling 'miss.'

If the current polling in Wisconsin is somehow totally accurate (which I see no reason to believe it is) we've got an essentially tied race. If the current polling is Wisconsin ends up being consistent with previous recent polling in Wisconsin in terms of the degree to which it understates Trump's support, Trump is up by around 5%.

i have been to Wisconsin 3 times since July. Even in MKE - trump signs are freaking everywhere!


Madison can only do so much if MKE is pulling even 1/3 for trump.

Wisconsin will (yet again) be the furthest right of the three rust belt battlegrounds. Easiest path for Trump is GA+NC+AZ+WI

Updated overnight results: pic.twitter.com/612i9J6NIL

— Rasmussen Reports (@Rasmussen_Poll) October 8, 2024

#New General election poll - Pennsylvania

🔴 Trump 49% (+2)
🔵 Harris 47%

Last poll - 🔴 Trump +1

Insider advantage #B - 800 LV - 10/8

— Political Polls (@PpollingNumbers) October 8, 2024

Captn_Ag05 said:

---

Wisconsin will (yet again) be the furthest right of the three rust belt battlegrounds. Easiest path for Trump is GA+NC+AZ+WI

---

I concur.

Are we still under the impression that the polls are still overestimating Harris as they did in 2016 and 2020? Or have to pollsters reconfigured and are polling better?

will25u said:

---

Are we still under the impression that the polls are still overestimating Harris as they did in 2016 and 2020? Or have to pollsters reconfigured and are polling better?

---

It's the major question of this election cycle.

If it's anywhere like the overestimating of 2020, then Trump is going to win in a landslide. They were off by 5-6% in both 2016 and 2020 polling for Wisconsin. Currently the RCP Average is Harris +.8%.

I have to think they've tried to reconfigure their models, but not in a trustworthy way. In the sense they probably just reworked their models by throwing whatever stuck to the wall to recreate the 2020 results and went with it without any scientific method.

So no, I don't trust really any of these polls. To the point I don't even know if they swung hard to the other side either.

But with the leaks that internals for both candidacies are mirroring one another I have to trust theirs more than these public ones.

2024 Wisconsin GE:

Trump 49% (+2)
Harris 47%

.@InsiderPolling, 800 LV, 10/7-8https://t.co/vy90cflztS

— Political Polls (@Politics_Polls) October 8, 2024

will25u said:

---

Are we still under the impression that the polls are still overestimating Harris as they did in 2016 and 2020? Or have to pollsters reconfigured and are polling better?

---

There's no way to know for sure, but I would be shocked if the answer isn't yes. I think the more relevant question is by how much.

Here's just one reason why.

Assuming that the pollster is actually trying to be accurate, within a reasonable budget, rather than just cranking out Democrat propaganda disguised as polls, it is getting harder rather than easier, it is getting more expensive rather than less expensive, to reach the lower propensity, working class voters that increasingly make up the Trump/MAGA/New GOP coalition.

Anecdotal information that I've heard suggests that a college educated, highly credentialed voter (generally speaking not very Trumpy), with a history of voting in both presidential and non-presidential elections, is a whopping 50 times more likely to respond to a pollster than a working class MAGA voter who shows up to vote for Trump in a presidential election, but rarely or never votes in non-presidential years.

As a result, the Republican and Republican-leaning independent samples in their polls are comprised of too many highly educated and too many over-credentialled voters, which leaves your data not representative of the Republican side of the electorate.

One of the easiest ways to recognize this, is if in the internals of a poll Harris is winning 10% or 12% or 15% crossover Republican votes, or if in the internals it shows Harris winning independents (at all), much less by 10% or 12% or 15% or 18%.

will25u said:

---

Are we still under the impression that the polls are still overestimating Harris as they did in 2016 and 2020? Or have to pollsters reconfigured and are polling better?

---

Even if they are off by 1/3 of what they were in 2016 or 1/2 of 2020 then trump would win based on current polling

Interesting stuff here.

New national NYT/Siena poll:

White non-college: Trump +30
White college grads: Harris +23
Non-White college: Harris +39
Non-White no college: Harris +31

— Ryan Struyk (@ryanstruyk) October 8, 2024

[We're going to remind you this thread is not for your personal opinions on which candidates are better or worse. It is for a discussion on polls. If you want to ad lib on the topic, take it to a different thread. If you continue with the approach here, you'll get a timeout and they'll increase in length every time you do it. We're done warning you -- Staff]

Legal Custodian said:

---

But with the leaks that internals for both candidacies are mirroring one another I have to trust theirs more than these public ones.

---

Please tell me more........leaks that both internals show Trump ahead???

dirtylondrie712 said:

---

Legal Custodian said:

---

But with the leaks that internals for both candidacies are mirroring one another I have to trust theirs more than these public ones.

---

Please tell me more........leaks that both internals show Trump ahead???

---

Yes. In both PA and WI. Maybe MI as well, but I'm less certain on that one.

Waffledynamics said:

---

New national NYT/Siena poll:

White non-college: Trump +30
White college grads: Harris +23
Non-White college: Harris +39
Non-White no college: Harris +31

— Ryan Struyk (@ryanstruyk) October 8, 2024

---

Since they were kind enough not to put their internals behind a paywall... some thoughts on this poll, which on the surface appears to be pretty reasonable in terms of the 'official' result (Harris +2), assuming a close race in the national popular vote.

I'll touch on the composition of the sample first

1) their sample is about 20-22% not white enough
2) their sample is about 8-9% too black
3) their sample is about 12% too Hispanic
4) their sample of high school (or less) educated is about 12% too low
5) amongst the college educated, their sample is about 5% too high for those with post-graduate degrees
6) renters are about 5% too high
7) homeowners are about 8% too low
8) their sample is about 1% too female (not a big problem)
9) their sample is about 4% too low with age 65+ seniors (this one can be tough to analyze the impact of because they don't subdivide their seniors in this poll between the age 65-75 group and the 75+ group - the 65-75 group will likely poll better for Trump and the 75+ group will likely poll better for Harris)
11) their sample is about 5% not Republican enough (this one can be a problem to analyze as well, because it depends upon whether we're talking self-identified Republicans, registered Republicans, or inferred Republicans)

Thoughts on the sample:

1) By itself, the impact of a less than representative sample can be partially mitigated with proper weighting, but only so much. Bad data can only be "fixed" so much with weighting.
2) If the characteristics of the sample go both ways, and by that I mean if some of the characteristics of the sample will likely favor Democrats and some of the characteristics of the sample will likely favor Republicans, then the discrepancies sort of cancel each other out.
3) On the other hand, if the characteristics of the sample are all likely to all favor one side over the other, its going to skew the results in favor of one side now matter how you weight the poll.

Compared to the actual voting population, therefore, let's summarize what their sample looks like, before we decide whether it looks more like option #2 or option #3.

A over-educated, perhaps with a graduate degree, minority voter, likely black or Hispanic, who rents rather than owns a home, which means they are more likely to live in an urban area than a rural area or the suburbs.

If their sample is too minority, too educated, and too urban overall, they probably polled in places and amongst populations, in general, that are overrepresented by Democrats and underrepresented by Republicans.

In addition, if their sample is too minority, too educated, and too urban overall, what kind of Republicans, whom there aren't enough of in their sample to begin with, are they most likely to have polled... more Trumpy or less Trumpy Republicans? Even the Republican portion of their sample, when weighted, is probably going to understate Trump's support just amongst Republicans.

Finally let's look at the weighting.

1) They realize their sample isn't white enough, so they weight it up to 65%, which is still 5% (or more) too low.
2) They realize their sample isn't Republican enough, so they weight it up to about 32%, which compared to their weighted number for Democrats, is at least 3% if not 5% too low.

In other words, the weighting still does not, IMO, accurately reflect the electorate... all in the Democrats favor... and yet, even with a sample that is too Democratic, and weighting that isn't Republican enough... they are still only able to get Harris to +2 nationally, which probably loses her the election anyway.