Steve Kornacki is having a sad about his polls

[deminks]

Steve Kornpucky, as I like to call him, the great and powerful MSNBC man in khackis who analyzes voting data, is having a sad about his Wall Street Journal poll the MEdia has been quoting for 2 days showing Biden and the orange anus 'neck and neck'. It leaves most anchors stunned, but nevertheless quoting it = it must be true because Wall Street Journal.

As pointed out by Mueller, She Wrote yesterday, the poll was partially conducted by The orange anus' campaign pollster, who has been paid $600,000 by the Make America Great Again PAC THIS YEAR.

Kornpucky is indignant that someone might question his integrity:

Link to tweet

My mentions have been flooded with this the tweet below - and accusations that I am pushing partisan garbage cooked up by a pro-Trump group - after showing the new WSJ poll on the air. I think it's worth a response because what is contained the tweet is incomplete & misleading...

Link to tweet

Again, this account is inciting outrage with polling misinformation. In this case, the claim being made is that a national WSJ poll showing a 46-46 Biden/Trump tie can't be accurate because...a poll in California - one of the bluest states - has Biden ahead by 20 points.

Mueller, She Wrote response:

Link to tweet

NOTE: I should say for accuracy that Fabrizio Lee is *partially* responsible for this poll. The other pollster is GBAO, and while they are funded mostly by dems, they were never questioned by law enforcement for giving polling data to indicted agents for the Kremlin.

End snips.

Toto is pulling the curtain back exposing the man in khackis flipping the levers.

Pre-election polls are garbage. Pre-election polls paid for by election campaign PACS even partially are garbage. GOTV. When we vote, we win.

[LakeArenal]

1. Modern polls suck.

Ext week there will be a different poll with different numbers. No way that people think Biden is great this week and equal to Trump the next.

First precious 538 guy is out because of bs he spouted now maybe Kornblow too.

He doth protest too much.

[uponit7771]

42. It's not the polls It's the media companies that need clickbait that headline them

... knowing their accuracy is sus.

polls like this with RV voting this far out. Know that they are worthless.

All media companies know this, they are thirsty for views

[hlthe2b]

2. Kornacki is as addicted to "conventional wisdom" as the rest of them. Maybe if they are put back

on their heels a bit by those exposing their complete credulity "because WSJ!" they might experience sufficient embarrassment to at least do some original source checking. It should not have fallen to an online twitter account (albeit from an award-winning podcaster) to question and identify the competing interests of the pollsters.

[deminks]

22. Amen. Like the conventional wisdom for a red wave in 2022

that didn't happen

[LiberalFighter]

35. Kornacki doesn't have a clue about conventional wisdom unless paid.

[Goddessartist]

3. Yep. Never liked him.

[JohnSJ]

4. That is the problem with National Polls for the Presidential Election.

[uponit7771]

44. +1, National polls are worthless in and of themselves it's battleground states that matter

[JohnSJ]

47. Yes, along with turnout, and independent/undecided

[yardwork]

5. Garbage in, garbage out. The polls are mostly meaningless right now.

Most people aren't paying a lot of attention to the 2024 election right now. Lots of people no longer have landlines. Many people don't answer their phones if the caller is unknown.

Pollsters are going to have to come up with better ways to gather data. I suspect that individual politicians' high-priced pollsters are using focus groups. Polls conducted with random phone calls are not likely to be reliable.

Just my opinion.

[Sympthsical]

9. The only people who need care are campaigns

Because the polls are snapshots of the moment. It could highlight states or issues or demographics that need more attention. It could point out, perhaps, some blind spots. So they're super useful for the people whose job it is to understand voter sentiment and put together strategies to change or influence it.

They're just not useful to us, lol. It's just horse race fodder, which can be fun. Except this is nothing like the track by me that has $1 hot dog and beer Sundays.

If cable news were handing out $1 hot dogs, I might actually watch.

[yardwork]

34. Exactly. And I'm certain those politicians' pollsters use complex sampling methodologies.

But they're not sharing their results with us, or anyone else. We get the horse-racey click-baity ooh looky there stories.

I can't imagine the MSM ever saying "Gee, looks like a landslide for X - nothing to talk about here."

[Sympthsical]

39. To be fair

The country is divided pretty closely on the national level. And hey, for a world of 24 hour news, that's the dream. Got nothing else to talk about? Poll time.

"Two people had a conversation, agreed to disagree, went on with rest of their day. Least interesting story ever. Those reasonable assholes. I bet one of them said something slightly untoward on TikTok. Let's get a producer on this. Sammy? Can we find out if one of them said something about the Barbie movie? No? What about immigration? Well, dig damnit. Twitter doesn't write itself, and we have a panel in studio with nothin goin on today."

And now I feel like I'm just writing about my job.

[Silent3]

25. "Doesn't predict the election" and "meaningless" are not the same thing

It’s amazing how many people can’t grasp this simple fact.

[yardwork]

33. If the data is skewed, the results are unreliable, i.e., meaningless.

I'm not a statistician but i know that much. Unreliable data = meaningless results.

[Silent3]

37. A lot of people complain, however because they confuse "don't like what I'm hearing"...

...with "must be bullshit".

As a read of the general current mood and standing of the electorate, not as something so super-accurate that it tells you where a tight election will land, a poll has to be really bad to not giving you a rough read of where things are.

And if lots of polls are telling you pretty close to the same thing, people shouldn't be so fast to assume there's a vast conspiracy to trick us, or a HUGE error baked into nearly every poll, like they're all equally stupid about cell phones or likely voters.

While I believe Biden will most likely beat Trump, people are sticking their heads in the sand if they don't realize, unless circumstances change greatly, we're heading for a yet another close race. Circumstances could change greatly both in and against our favor.

If people are saying "polls are meaningless this early", again, that's treating polls as if the only purpose in the world for them is predicting an electoral outcome. If you're saying "polls are meaningless this early" because "people aren't paying attention this far from the election", well, you're still learning something useful. Sad, but useful: that a lot of voters are fucking clueless.

If you need November 2024 to be a month away before you pick up on the fact that Biden is actually doing a great job, and Trump is far too corrupt and incompetent to run the corner 7/11, then you're either stupid, ignorant, hateful, or a wonderful mix of all of those things. The early polls are telling us that a lot of these awful people are out there, and the polls aren't lying to us about that.

[uponit7771]

45. National polls are bullshit seeing we don't elect by popular vote and battleground

... state polling is a way better gauge

[Silent3]

46. That's still sticking your head in the sand about the reality of public sentiment

Whenever a candidate has won the electoral vote, but lost the popular vote, the race has been a tight race.

These polls are showing a tight race at this point. They aren't all hugely off base, and they aren't all fabricated lies.

[emulatorloo]

28. Absolutely. A lot of people won't get serious about the election until after Labor Day 2024

[Silent3]

38. If "people won't get serious about the election until after Labor Day 2024" that doesn't mean the poll is bad

That means we're surrounded by a lot of people too fucking stupid to see how clearly superior Biden is to Trump. Anyone with half a brain shouldn't need to election to be a month or two away to see that.

These polls, many of which show Biden and Trump very close, aren't all lying to you or making major statistical errors. They're letting you know, not with great predictive value, but still with rough accuracy, that we've got a lot of stupid, ignorant, or just plain nasty (they know what Trump is, and they like that) people out there.

[Walleye]

6. If they don't spend their time talking about polls, they might have to report real stories.Hard work

[erronis]

19. You nailed it. Doing real reporting requires getting out of the chair and gathering data.

Not pushing buttons on some gee-whiz graphing package loaded with very selective data.

[Walleye]

30. Reliance on polls has just gotten way out of hand

[Sympthsical]

7. People are so pressed about this man

Like watching a mailman deliver bills and seeing the recipient scream, "I never liked that blue-vested motherfucker!"

I dunno. It's funny to me.

[MyNameIsJonas]

8. It's not unusual for polls to have both a GOP & Dem pollster.

It is unusual, however, if that GOP pollster involved also works directly for the campaign they're polling on.

That seems sketchy.

[dchill]

13. Kornacki IS sketchy. Polls are sketchy...

I think data is being farmed. Push the narrative. Catapult the propaganda.

[everyonematters]

10. All of the polls are saying the same tihing, a dead heat betwen Trump and Biden.

In 2016 the polls substantially underestimated the Republican vote. In 2020 it was not as bad, but still undercounted the GOP vote. So they may still be underestimating the Republican vote.

[Sky Jewels]

21. In the midterms they overestimated the Republican vote.

In the wake of Dobbs, a lot of the polls were oddly weighted for older and male respondents.

[ALBliberal]

11. Any poll that shows Biden tied with Trump IMO

is biased in some way. And yet it can only energize us to get out and vote. That’s how I have to look at these things going forward. I do agree that this poll was influenced by a Trump “insider” if you will and I’m certainly not trying to argue that point.

[CrispyQ]

12. I love this part!!

The other pollster is GBAO, and while they are funded mostly by dems, they were never questioned by law enforcement for giving polling data to indicted agents for the Kremlin.

[senseandsensibility]

36. Lol

[Fiendish Thingy]

14. So, what was Kornacki's response?

In the first tweet he mentions a response is warranted; the second tweet doesn’t appear to be that response, or isn’t the full response…for the sake of accuracy and balance, shouldn’t Kornacki’s full response to Meuller She Wrote be included?

[cilla4progress]

15. "AG," the Allison Gill of

Mueller She Wrote (MSW network) is smart, connected, well-informed and does good research.

Interesting woman and story.

She does the Jack podcast with Andy McCabe.

She does the Cleanup on Aisle 45 with Pete Strzok.

These are 2 venerable legit dudes.

She also does The Daily Beans with Dana Goldberg.

I'm a patron.

[MontanaMama]

24. THIS. All of it.

Allison Gill is also a regular guest on Stephanie Miller. She is the real deal and an excellent researcher. That said, Kornacki needs to find a new shtick. He's annoying as hell and almost never right.

[cilla4progress]

29. ...

[GreenEyedLefty]

48. Thank you for saying this.

I am 1000% better informed for just a few bucks a month (Patreon). MSW is worth every penny.

And who doesn't love news with swearing??

[FakeNoose]

16. If this is how Steve Kornacki rolls ... not checking his sources before broadcasting sketchy polls ...

... then he just lost me. I've been sticking up for this guy for the last 5 years here on DU.

Not any more. I'm done with him.

[IbogaProject]

17. 600 is too low

beyond that that was 600 registered republicans. Statisticians worked out presidential polling in the 1950s with that lower population the best number to poll was a little over 3,000. By the late 1980s the models had be refined and just above 2,000 actually did better mathematically. No way 600 is enough. I'm guessing they did more than one and went with one that fit their narrative, make it a horse-race with donations needed with some likely to win hype to keep the faithful's wallets open.

[Recycle_Guru]

20. not too low

600 sample plan is not too low--if structured correctly you could end ip with 2 to 3% margin of error +-; however this poll group sample is garbage

[IbogaProject]

26. + - 3 % is meaning less with how narrow the electorate is

While the national vote difference was huge, the last two presidential elections were decided by less than 200,000 votes out of over 100,000,000.

[Recycle_Guru]

31. so you agree national polls are inexact at best. State level polls are more helpful

but are less accurate so far out from the election and are usually conducted less frequently than national polls

[Recycle_Guru]

18. aww!! horse race media whore got caught and don't like it!

[BlueIdaho]

23. Kornacki is full of shit

Time to bench this mentally lazy slob.

[Paladin]

27. Call us even, Kornacki.

A little retribution for all the years you've been gleefully delivering depressing polling news to Democrats.

Whine a little louder for the folks in the balcony seats---they're not getting the full effect (or the full enjoyment).

[Hortensis]

32. After decades of learning how misleading at best polling is, but always early,

I don't follow polls before the GE, if then. Saves a lot of totally wasted and unnecessary upset.

I strongly recommend avoidance for that reason -- except for whose interest is in how they're used to understand and affect elections and manipulate populations. A fascinating topic for those who study it, of course, but they discuss elsewhere, mostly in closed groups.

[Kid Berwyn]

40. NBC has to talk about a horse race.

Think of the ratings.

Wonder what they'll cover when Trump is in prison?

[senseandsensibility]

43. Probably "his base will still vote for him"

Over and over again. Kind of like now.

[uponit7771]

41. Good, rv polling right now is mostly click bait