There's a new Emerson poll out that shows Biden ahead by 2 points nationally

[octoberlib]

Link to tweet

Link to tweet

/photo/1


538 which was updated 2 minutes ago had Biden at 50.2%, Trump at 43.1%

So don't panic . Last month Emerson had Biden up 4 points so he's dropped 2 points. No way Trump has 19% Black support.


Michigan Pennsylvania Wisconsin Arizona

+6.3 +4.7 + 5.3 +3.1


Trump's approval rating is at 43.1%

[JI7]

1. The black vote in that poll is bs

[USALiberal]

3. Thanks for your detailed response! nt

[regnaD kciN]

2. Not to be picky, but this is the THIRD thread on this poll so far...

[octoberlib]

5. My point is that one poll and a national one at that is meaningless in a country

where Presidents are chosen by electoral college votes. Also , always go by the aggregates. People are freaking out.

[bearsfootball516]

4. There's absolutely something off about that poll.

Hillary won the popular vote by 3. The idea that Biden, a significantly more liked candidate is only +2 against Trump, who is significantly more disliked than he was in 2016, is laughable.

[octoberlib]

7. This pollster is skeptical of Emerson's methods

Link to tweet

Trump gets a statistically insignificant 2-point RNC "bounce" according to a new Emerson College poll (which we don't include because of pretty egregious data quality issues)


I don't care about the 538 rating. Emerson uses two of the most unreliable sampling procedures in the industry, then proclaims that they have fixed all the errors with post-processing, even though we know MTurk errors run deeper than that

[regnaD kciN]

12. For those who don't know it...

...G. Elliot Morris is the data analyst for The Economist, and runs their prediction site.

[calguy]

10. During the last briefing from The Lincoln Project .....

They said to not pay much attention to the polls. There are so many using so many different criteria in their polling that the numbers can go all over the place.

[FBaggins]

6. Hard to believe the two largest minority groups would be that high for him

Wonder what accounts for it. No Democrat could win with only 77% of the black vote nationally.

[Yavin4]

8. Also shows Trump getting 19% of the Democratic vote

That is truly hard to reason.

[Awsi Dooger]

14. The 90/10 categories are often difficult to capture in polling

I have noted that for decades. The numbers are so small that all it takes is an odd sampling to throw things out of whack. If you look at crosstabs sometimes it can literally be 10 people skewing things, like a Hispanic sampling in some state that is only 6% Hispanic and among a sample of 1000 voters the Hispanics are 30-30 instead of a more logical 40-20. Now it's 50% instead of 67% and everybody freaks out wondering how Hispanics can be 50-50.

I always ignore the partisan and Africian-American polling percentage and assign the standard 90/10. Trump received 8% on the black vote according to the 2016 national exit poll, and likewise 8% of the Democratic vote

[oswaldactedalone]

9. We handed Trump the gift of a lifetime

With “Defund The Police.” You knew they would exploit that to create fear in “white world.” That was a huge fuck up that may swing the election to Trump.

Bottom line, we have to mine every vote that we can and hope the turnout is enough. At this point, it’s a tossup, and Trump will steal a close election. Keep working.

[JonLP24]

18. This poll is flawed

And "defund the police" has actually been around before Trump was even elected. It wasn't until the George Floyd killing that the national media paid attention.

I don't give a crap about white "fear". African Americans have bigger reasons to have fear and it is often from the police.

It is funny that John Oliver & Samantha Bee are more liberal than DU.

It is interesting that so many people buy into Trump's narratives but don't counter him on Trump's efforts to defund education and the post office (to rig an election FFS).

[Grins]

19. Yup!! Royally effed that one up. NM

[regnaD kciN]

11. The thing I find unusual about that poll...

...is that, last time, they had Biden 50% - Trump 46%, with 4% either undecided or third-party. Since then, according to their results, Trump has gained 3% and Biden has gained 1% (in other words, it’s not a matter of Biden losing support, or even staying the same), and basically every undecided voter and every voter previously leaning third-party committing to one of the two main nominees. When has that ever happened before, let alone by late August?

[octoberlib]

16. Great point. As far as I know, it's never happened.

[FBaggins]

20. I think they report two different numbers

One with and one without undecideds. Your comparison is between the "with undecideds" in one vs. "without" in the other.

The comparable current figure is 49/47, not 51/49.

[radius777]

13. Emerson has been an outlier this cycle,

and though it was a high rated poll in the past, several on social media have questioned its current methodologies. Highly doubt Trump is getting 19% of the black vote. Biden is stronger amongst blacks (and whites) than Hillary was.

There is a tightening, Trump's 'law and order' message seems to be working - but Biden's response to this message is powerful (Trump is the source of chaos and covid) and Biden has just started to pound it. This will work as voters tend to blame the sitting president for any problems. Our convention was 'feel good' whereas theirs was negative. Now it's our turn to hit them and drive up their negatives, we have not yet begun to fight.

[VMA131Marine]

15. Focusing on individual poll results is a sure way to drive yourself insane

Even a well-done poll will get a result where the true state of the race is outside the MOE at least 32% of the time. The important thing to watch is the trend in the polling average; this has narrowed from Biden +9 to +7 over the past few days. There is even a margin of error on the polling average so it is not clear whether this change is random noise or a statistically significant trend. Time will tell.
Note that the USC-Dornsife poll which is a much bigger sample, still has Biden +14 as it has done for a couple of weeks.
Note also that pollsters are in it for the money. They write headlines around their polls even when those headlines aren’t warranted. They report a widening or tightening of the polling margin as if it were big news, even when the change is not statistically significant. Scientists have to achieve at least 5-sigma (5 standard deviations) of certainty to claim a statistically significant result. Pollsters never come close to this standard.

[octoberlib]

17. We all just need to vote and stop stressing out over every poll. It's hard , though.

What I hate the most is that Trump and his supporters are going to be crowing over this. Informative post, thank you.

[bamagal62]

21. I can't believe there's a single person

That’s undecided.

[DFW]

22. In 2012, Obama's top election numbers guru told a closed group of us

David Simas said the numbers always tighten around Labor Day. He said don't panic, that he and his team would "tear out their hair so we didn't have to," and then Obama would pull away again so far that Romney never could catch him.

And so it was.

[octoberlib]

23. Thanks for this post! Good to hear.

[DFW]

24. This was in JULY, mind you

He knew exactly what he was talking about.

[Baitball Blogger]

26. Thanks. It's great to hear someone who has seen it from the inside.

I don't think any of us will be the same after 2016.

[Awsi Dooger]

25. Obama won by 3.9% in 2012

We can't afford for it to be that close this time, given how the electoral map has shifted