Thread by @dogemperor

Yes, this is my usual Don't Rely On Polls and Know The Limitations Of Polling post, this time focusing on Emerson (and I'll compare with observations with Marist--a decent poll--and polls with other methodology issues)

A thread:

1/x

https://twitter.com/IanSams/status/1848356720130572384

So again, we're dealing with another small sample pool of 1000 polled, and tries a mix of online-polling (via CINT, which goes through paid-survey portals), landlines (robocalls), and cells (MMS), and is English-only (so misses a chunk of Hispanic vote)

2/x

And RIGHT AWAY when I look in the crosstabs in their Excel spreadsheet....I see the SAME issue that crops up in EVERY poll except Marist, in that there's a likely undersampling of under-30s (at 16.4%) and oversampling of over-50s (51.7%), with nearly 19% 60-69 alone

3/x

Emerson attempted to set up pools based on a roughly equal set based on roughly 10-year separations between cohorts, whereas Marist weights by American Community Survey numbers (and by generational cohorts) Marist has a pool closer to 36% for people under 35

4/x

Looking at crosstabs, as you get above the 50+ percentage, the majority of the people they surveyed in their pool...are registered Republicans (around 19% overall) and more of a skew towards Independents esp. from 40-70 years of age (again around 15-21%)

5/x

As I've discussed before, Fox News viewership is almost *entirely* concentrated in the 55+ demographic

And voter reg in their pools under 40 skews Dem (most strongly with under-30s, but even up to 40-50 years of age).

6/x

One particularly odd feature in this crosstab is that there's high number of *older* people who voted--like, as high as 27.5% of people over 70, 22.4% of people over 65

In comparison, under-60s it's closer to 10-15% "already voted"

7/x

Another thing that seems kind of odd in comparison to other surveys is that almost every group *under* 50 is "meh" re Biden's performance--as in "majority is 'Neutral or no opinion'"--and fairly strongly with under-30s in particular

8/x

Again, as expected, there's a strong skew towards Harris for people under 50, with 59.7% of under-30s going for Harris (with the caveat this is a national poll with a pool of under-30s of exactly 164 people) and 52.2 of 40-49 demographic

9/x

As per usual, the skew towards The Former Guy starts in 50+ up, and is most pronounced in the 60-69 year range (also, the single largest cohort IN this survey by a fourth) at around 53.9% for TFG

And that itself is a bit of an issue in methodology I'll point out

10/x

The single group that's going the hardest for TFG in this survey...is an age cohort that has 189 people in the pool; others have pool sizes of 159 (save for under-30s at 164)

This. Does not match American Community Survey numbers of population members :D

11/x

Marist breaks out by generation, but Boomers start out at 60 (and the high end at 78), that was 27% in their survey; Silent and Greatest Gen (79+) was 2%

Emerson has a cohort of 15.9% who are 70+ (early Boomers/Silent/Greatest)

12/x

(Also as a small kvetch--there are definite generational differences between older and younger Boomers, and there is a *definite* generational divide between "older" GenX (who DO trend conservative) and what's sometimes called "Oregon Trail Generation".)

13/x

("Oregon Trail Generation" aka "Xennials"--basically the "cusp" gen between older GenX and Millenials--tends to have a lot of characteristics closer to that of Millenials, including trending more progressive.

Hence why I do tend to emphasize the 50+ and 55+ line here.)

14/x

There's also an interesting skew in that the "forcing" question for "undecideds" DID split pretty evenly between 30-50, but did skew Trumpy otherwise--which *doesn't* match with similar forcing polls which have tended to go towards "would not vote"

15/x

In regards to the specific question on when people chose their candidate--again, most people under 40 have made up their mind this month (with many under-30s just in the last *week*), 40-49 you still have over 20% making up their mind but similar last-weekers to < 30s

16/x

In comparison, it's specifically the same 50+ contingent that tends to have largely decided over a month ago, and (again) especially strong in that overly large contingent of 60-69s polls who trend particularly Trumpy.

70+ had HIGHEST rates of indecisiveness!

17/x

And on that, I'll note an observation--many experts on coercive religious groups *do* consider Trumpism to be a cult, and there is an *extremely* strong correlation between being a white evangelical who espouses NAR theology *and* being a MAGAt on main.

18/x

At least one popular book on MAGAtism as a coercive religious movement does note the phenomena of "cross-recruitment" (something I've written about for close to 2 decades in the context of NAR and MLM downlines recruiting for each other)

19/x

And the New Apostolic Reformation is widely considered to be a coercive religious movement; the single highest predictor of a Trump voter is if someone is a white evangelical that embraces "Seven Mountains Mandate" theology (a NAR distinctive) on some level.

20/x

So yes, in this particular survey we're actually seeing evidence of how coercive political movements and coercive religious groups are *likely* setting MAGAts' ideas in stone (and yes, I'd count a LOT of Fox News programming in that regard, too).

Continuing:

21/x

One GOOD sign I am seeing among younger voters IS that they're knowing the importance of "voting blue all the way down"--57.92% of under-30s will vote D for Congress, 52% from 30-50, and again the shift to R for 50+ (and again most pronounced in that 60-69 cohort)

22/x

62.8% of under 30's have a very or somewhat favorable view of Harris, 57.8% very or somewhat favorable for 30-49, and 56.6% very or somewhat favorable for 40-49 (with a HEAVY shift towards "Very Favorable" in 40-49 in particular)

23/x

Again, only start to get the shifts towards unfavorables in 50+ cohorts, and again *particularly* in that 60-69 year cohort, which makes me wonder if they accidentally hit a particularly Trumpy retirement community in some of their surveying

24/x

And a possible sign we may have a more-conservative-than-normal pool is the Abortion Question. In almost every poll that doesn't show symptoms of having gotten a *very* conservative polling poll, abortion access tends to be THE major issue for under-45s overall

25/x

In THIS particular poll, everyone's claiming it's "the economy"--including 18-29s, which in almost EVERY other poll have STRONGLY skewed towards abortion as THE issue (often upwards of 60% or more)

Abortion was a primary issue for *14.6%* in THIS one.

Something's *fucky*.

26/x

Marist did *NOT* ask the Abortion Question, but even polls that have strongly skewed towards older respondents that DO ask the Abortion Question (even NYT/Siena and THAT poll skews old) STILL have under-45s noting "it's the reproductive access, stupid" in SPADES

27/x

Another sign something is decidedly Amiss compared to other polls: the fact that "Immigration" is ranking as highly as abortion--this is something that typically pops mostly in right-wing survey groups, and (again) shows STRONGLY in that 60-69 contingent

28/x

When a similar 14.6% of under 30s are going on about Immigrants in this survey being a major factor in their vote--when it is *negligible* in other surveys--that to me *screams* that this poll trends fairly hard conservative

(And there's OTHER bits very Off.)

29/x

Interestingly, this is the first poll I've seen that asks a specific question regarding Project 2025--and THIS and some of the followup questions make me think this particular poll was VERY skewed R AND they had a very conservative polling pool

30/x

The Project 2025 question asks "Does Project 2025 make you more or less likely to support Trump or does it make no difference?"

There are multiple things with the answers that are extraordinarily Sus. *EXTRAORDINARILY* SUS.

31/x

When Project 2025 has been mentioned especially by youth influencers and musicians popular among youth, this has resulted in measurable spikes of Google searches, and even on this hijacked hellsite counter-Project-2025 stuff trends fairly frequently

32/x

In this poll, 31.7% of under-30s claimed it would "make no difference", while 17.6% claimed it would make them *more* likely to vote for The Former Guy, and 7.9% claimed to have never heard of Project 2025 (with 42.6% less likely to vote for TFG because of Project 2025)

33/x

This is sus as FUCK on multiple levels.

One: The 17.6% "oh yes this sounds wonderful" makes me think they either got a bunch of incel trolls and/or TPUSA members to bomb the survey or their survey poll must have consisted of Liberty University and Patrick Henry College

34/x

And I CAN say that because while this is the first *Presidential preference poll* that has mentioned Project 2025, specific polls on the subject of Project 2025 itself have found it to be EXTRAORDINARILY unpopular and increasingly so over time.

35/x

Even in the NBC/Hart polls which have skewed towards very right-wing pools of respondents, Project 2025 had a sum favorability of *4 percent* (with 23% unaware of what it was, and 57% negative with 51% very negative)



36/xs3.documentcloud.org/documents/2517…

According to UMass Amherst poll conducted in late July on YouGov (self-selects conservative), over half of Americans (53%) had heard of Project 2025 and even among a STRONGLY conservative poll that shows evidence of being "TPUSA bombed" STILL showed strong negatives

37/x

shows this (and the 18-29 spread here shows what I mean by "TPUSA bombing", you'd think Andrew Tate's fan club literally bombed the poll, but STILL generally highly negative about Project 2025 even despite the attempt to bomb the poll.)

38/xumass.edu/political-scie…

A Navigator Research poll conducted in early August showed almost all support for Project 2025 was from MAGAts, and much of this survey (which did include a "MAGA Identifying GOP" question) noted abortion rights ARE a big part of the concern

39/xnavigatorresearch.org/americans-incr…

This particular panel (which, again, was done through "paid survey" portals with followups) also skewed unusually conservative in aspects, and only had a 19% favorable rating for Project 2025 among *registered Republicans*

40/xnavigatorresearch.org/wp-content/upl…

A different YouGov poll (and again: Keep in mind, paid survey portal, self-selects con, THIS particular pool selecting STRONGLY hard R) only had a 14% favorable response (even with a pool where LITERALLY 58% want to militarize the US border)


41/xd3nkl3psvxxpe9.cloudfront.net/documents/Proj…

YET ANOTHER YouGov poll in early August (again: self-selecting conservative pools at a paid-survey portal!) shows a MUCH more typical spread on the abortion issue (55% consider "very important" under 30, 54% 30-44)

42/xd3nkl3psvxxpe9.cloudfront.net/documents/econ…

And this poll STILL only shows a 16% favorable rating for Project 2025, with only 4% "very favorable", and STRONG unfavorables (53% overall) among under-30s in a polling pool that has a history of trending *STRONGLY* conservative at points

43/x

And very likely what we are seeing in certain cases is pollbombing, which is something I keep noting is a risk with online polls in particular, but the point is that *even despite pollbombing* Project 2025 is PROFOUNDLY unpopular except among Christian Nationalists.

44/x

Another sign we're dealing with a conservative skewed poll is the responses to hurricane response (which skew more towards "Poor" starting at the 50+ cohort, again with the peak of 60-69 which must have been the faculty of a Bible college).

45/x

Another thing that makes me *extremely* sus of pollbombing being involved: 52% of 18-29 year olds claiming that "the number of migrants seeking sanctuary in the US is a crisis".

This NEVER shows up as a major issue in this cohort EXCEPT with extremely far-right pools

46/x

The one area you DO see numbers like that? Fox News viewers, older conservative voters. This survey shows 57.1% of voters 60-69 (that same Trumpy contingent!) also consider The Migrants And Refugees A Crisis, and THAT actually maps with other surveys out there

47/x

At this point I am *legitimately* questioning whether their polling pool consisted entirely of people in Bible colleges and Libertarians, and there's also some other signs of some strangeness afoot in the survey itself

Like. The basic *demographics* of the survey

48/x

Apparently the vast majority of people of other races (of literally ANY rate not "white") were HEAVILY distributed towards those under 40, and that even includes Hispanics and Latinos

Usually there's a MUCH BETTER spread across ages and racial/ethnic demographics

49/x

AND looking deep in, we see a reason for the skew.

This is one of the few national surveys that actually specifically tracks which regions of the country are being interviewed.

This particular survey skews HEAVILY towards respondents in the South.

Uh-oh.

50/x

So y'all who have followed me on my deconstruction of political surveys so far have seen me mention a thing called the Christian Nationalist Belt.

And this is a MAJOR factor that can badly, BADLY skew a national survey--especially if most of your respondents are in it!

51/x

So on average, *nationwide* around 30% of Americans support SOME form of Christian Nationalism, with about 10-15% of those being "hard" Christian Nationalists (the full-on Seven Mountains sorts, what most people *think of* as a Christian Nationalist)

52/x

But there is an identifiable belt of states--largely in the Midwest and especially in the South up through Appalachia, tracking through the "Bible Belt"--where Christian Nationalist sentiment (including "hard" Christian Nationalism) is FAR higher

53/x

If we narrow it down to *WHITE* Christian Nationalists, these actually form a de facto majority among whites in certain areas--including Mississippi

There are specific counties in this survey in Tennessee that have individual counties of over 70% Christian Nationalists

54/x

And in red states, there are (unsurprisingly) more Christian Nationalists, including "hard" Christian Nationalists (15% on average are "hard" Christian Nationalists in red states)

And there is a DEFINITE association between MAGAts and "hard" Christian Nationalism

55/x

Something like 21% of people who are registered R are "hard" Christian Nationalists, likewise, around 21% who openly identify as MAGAts or see TFG as favorable are "hard" Christian Nationalists

And the questions re immigrants and hurricanes are a Tell for another reason

56/x

Christian Nationalist groups have really had an "alt-media" platform for a LONG time (really since the days of "The PTL Club" and CBN) which nowadays not just includes televangelists but more secular-seeming "news with views" like OANN and NewsMax and RSBN

57/x

And yes, the "Migrants scary" and "Hurricane response bungled/deliberately meant to hurt" are far-right narratives, and Christian Nationalists tend to rely on far-right "news with views" even more than Fox News--30% of the audience as "hard" Christian Nationalist

58/x

And yes, Christian Nationalism is a phenomena that tends to skew older, particularly "hard" Christian Nationalism--there's an ever-growing community of #exvangelicals (myself included!) that are proof of that

And there's another factor or two that plays in

59/x

Something like 30% of self-identified "white evangelical Protestants" are "hard" Christian Nationalists (and a larger percentage, close to the combo of "hard" Christian Nationalists and sympathisers, do espouse at least SOME aspects of NAR theology).

60/x

There is also a subset of something like 22% of Hispanic Protestants that qualify as "hard" Christian Nationalists--almost to a one New Apostolic Reformation (which has *heavily* targeted the Hispanic community for some 50+ years now).

61/x

There is ONE notable divide between white Christian Nationalists who are Protestants and literally *Every* other group which espouses some form of Christian Nationalism--generally, Hispanic NAR types are *not nearly* as virulently opposed to migrants seeking asylum.

62/x

And the theological belief discussions note STRONG associations between Christian Nationalist beliefs and specific distinctives of Manifest Sons of God-lineage groups (including the NAR, "prosperity gospel" groups, etc.) AND likelihood to embrace violence like J6

63/x

And in general Christian Nationalists are STRONGLY anti-LGBTQIA, strongly anti-immigration, and strongly anti-reproductive-health

And the GOP skew is ALMOST ENTIRELY WHITE CHRISTIAN NATIONALISTS with a HANDFUL of Hispanic NAR acolytes

64/x

(In all seriousness: This is a REQUIRED bit of reading if you want to basically grasp why your churchy grandma fell right down the QAnonsense hole. )

And there's some more things that show correlations with Christian Nationalism

65/xprri.org/wp-content/upl…

In general, white Christian Nationalists are more likely to not have gotten a degree, trend 50+, tend to be Republican or "independent", tend to self-identify as "born again" or otherwise evangelical, and in ALL groups it's connected to MSoG distinctives

66/x

And, well, knowing where the Christian Nationalist Belt is can impact directly your survey.

If you hit, say, Alabama--where fully 23% are "hard" Christian Nationalists--in your survey, it's gonna get skewed a LOT, just saying.

67/x

So THIS becomes important in our survey, because...a HEAVY number of those surveyed are in the South.

18-29, nearly 40% are from the South, and 26.8% comparatively from the West (the next largest group, and likely included because of fear-stoking re migrants).

68/x

With our Very Trumpy Cohort in the 60-69 age range, 39.6% are from the South, and the next largest group was 24.8%...from the *MIDWEST*. (In other words, both sides of the Christian Nationalist Belt!)

Same skew for 75+ for South/Midwest, same skew South/West for under-50s

69/x

All in all: 37.4% of respondents entirely in this survey are from the South, 21.7% from the Midwest, and 23.4% from the West (likely including Idaho and retirement communities in Arizona that trend 55+)

Only 17.3% from the Northeast US.

So. There's THAT

70/x

And yes, it sounds like a long missive, because it IS a deconstruction to show you how even the best laid-out polls have some fundamental limitations that you have to recognize, especially in an era where a lot of the group you WANT to poll (under-45s) actively AVOIDS polls

71/x

And what I can say is REASONABLE to glean out of this one is that even with a fairly strongly conservatively-skewed pool, Kamala isn't doing badly (and it tracks with polls that are SIMILARLY skewed hard-right with likely heavy Southern polling pools)

72/x

This, to me, shows 3 actual actionable pieces of info we can work with:

a) A lot of people are seeing TFG's deterioration and making up their minds
b) And the ones not drinking the redcap flavoraid are STRONGLY breaking Harris
c) And these tend to be younger voters

74/x

And basically that's the thing--you have to know the biases (deliberate or otherwise), you have to know limitations on a poll, you have to learn to NOT FUCKING PANIC, and you need to look at what you can reasonably gain from things in the end

75/x

And in the end? I don't care if the polls show her 80% favorable with people who decided last week. We STILL GOTV like she's 20 points behind.

All gas, no brakes.

We got two weeks. Let's get the assignment done, kids.

76/end

@threadreaderapp unroll