Ignore the Monmouth Poll

[Rule of Claw]

"This includes 320 contacted by a live interviewer on a landline telephone and 489 contacted by a live interviewer on a cell phone."

40 percent landline? This survey says it is more like 6.5

https://nationalinterest.org/blog/buzz/study-only-65-percent-us-households-have-landline-135212#:~:text=Study%3A%20Only%206.5%20Percent%20of,a%20Landline%20%7C%20The%20National%20Interest

There is usually a 15-20 pt swing in support landline to cell.

This poll should be plus 11 Joe, nationally. In fact it probably is.

https://pbs.twimg.com/media/EjBjAxxVoAAUOz6?format=jpg&name=medium

Plus 65 percent white? If the electorate is only 65 percent white Joe would win 55 percent of the vote at 42 percent white support.

And swing counties? ABC just did those: plus 12 Joe.

This poll is not remotely logical.

[liskddksil]

1. A+ pollster, ignore at your own peril. That said all polls are snapshots in time

and no one poll should be treated as gospel.

[Rule of Claw]

3. Poll is illogical.

42 times .65 percent white plus 66 times .15 Hispanic is 37.2. At 90 percent with AAs and 60 percent with Asian voters (both low) Joe would be at 53.2.

(Joe would have to be at 60 percent with AAs and 55 percent with Asian voters for these numbers to work.)

53.2 is in line with most polls. None of these internals suggest anything but accelerating momentum, particularly with likely voter demographic makeup.

[jorgevlorgan]

2. It brought down the lead from 7.3 to 7. The others still had him between 7-10

Lead. And monmouth also had a registered voters poll that was averaged in. It had Joe leading by 6

[BannonsLiver]

4. Horse Race Nate says it is A rated but concedes it has a pro Trump bias.

Last edited Mon Sep 28, 2020, 04:32 PM - Edit history (1)

[Rule of Claw]

5. Ya think?

Any poll that shows 57.5 Biden percent support with the Asian/AA community might.

[brooklynite]

26. Actually, 538 says Monmouth skews D+1.3%

https://projects.fivethirtyeight.com/pollster-ratings/

[BannonsLiver]

27. Hmm I wonder why he said the opposite on Twitter today

Maybe he’s having a rough day. Thanks BKL!

Link to tweet

?s=21

[Statistical]

6. Not everything has to be a conspiracy. Manmouth is a good poller.

If you have a dozen polls one will be the outlier high and one will be the outlier low.

The poll has Biden up 4. The Margin of error is 4.9% so even if the poll reflects reality perfectly Biden could be up as much as 8.9%.

[Fiendish Thingy]

15. You are misstating the MOE-538 lists the results as Biden 50/Trump 45, so with a +/-4.9%MOE...

It could be as extreme as Biden 54.9/Trump 40.1, OR Trump 49.9/Biden 45.1

The MOE is for the responses for each individual candidate, so both candidates’ results can move up or down. If you add 4.9 to one candidate, you have to subtract from the other.

So, in the case of this poll, the race is statistically tied within the MOE.

[JI7]

7. What's your agenda with these type of posts ?

[Rule of Claw]

8. Inappropriate phrasing.

I ask you to please address me respectfully.

[Peacetrain]

18. ROFL!!

[ms liberty]

14. Good question. n/t

[David__77]

9. It's just a data point among many.

I suggest that people consider that the race is winnable and by no means won, regardless of what the polls say.

[Rule of Claw]

11. Of course.

But we have to watch that narrative like a hound dog. Murray used this data to say the race was tightening.

Yes Patrick, if only 60 percent of AA's vote for Joe the race is tightening.

[David__77]

13. "Tightening?" So what? That's been said numerous times to date.

I don't see much problem with it. I certainly don't want people to think it's a done deal - that was a big part of the problem four years ago, I think. Many people bought into a sort of disbelief that Trump could win. He could and he can, of course.

Put another way, I would prefer that the polls show a "Trump bias" than show a "Biden bias."

[brush]

20. That's the point. Any poll that says only 60% of African Americans will vote Biden...

is not to be taken seriously. It is pretty well known that African American support for Biden is around 90%—with AA women it's 95% or higher and a little lower than 90% for AA men.

[JustAnotherGen]

10. What time frame was that?

I'm being polled this year and participated in Monmouth -

Told them I was voting straight GOP. Hunterdon County women of color have committed to 'saying that' - although it's the opposite.


Let them be sucker punched on election night.

[Rule of Claw]

12. 9-24-27

[Drunken Irishman]

16. Your point about landlines means little without other polling context.

Comparing it to the national average is not really helpful since I'm assuming most polls contact people with landlines at a higher rate than the national average of households who own them.

I'm not saying it's a wrong point but without being able to compare it to other polls, it's hard to know if their totals are significantly off.

I checked three recent polls:

ABC/WaPost only mentioned it did a mix of landlines & cell phones but did not break it down.

Echelon was online only.

NYT/Sienna didn't state either way.

[Rule of Claw]

17. One of the recent national

polls I believe was 10 percent landline. Irrelevant now, I suppose given the scandal and the debate if he shows, but if you are a pollster, and you look at the data, and nearly every poll for 60 years shows 80 percent or better with AA's for Dems and you see 57 on yours, you just shrug? "It's just a data point, yade da."

[Drunken Irishman]

19. I think most pollsters adjust for that.

Monmouth could be wrong in their models. I'm not saying otherwise but frankly, unless there's really something weird that stands out, like Trump actually leading, it's impossible to say. Unfortunately, polling in 2020 is much harder than polling in 2000. The dynamics force these pollsters to alter their models and adjust their results so they accurately portray the electorate as best as possible.

Monmouth may be wrong but I suspect they're likely closer than polls that show Biden up double digits.

[Rule of Claw]

21. No. Much more likely Biden leads by 10 or more

than 7 or less. This is just math. Ok quickly the electorate will likely be no more than 70 percent white, probably 68, let's say 69.

Joe is probably not getting to 48 with whites but 42 is good. 69 times .42 plus 31 times .78 (non-white vote) gets him to 53.16

From there the percentages go up, not down. More base consolidation, more strength with college whites, suburbs, so I have him winning pre-tax scandal 54-45, now 55-44. Vs. Trump or Pence. 54 vs. Pence.

[Awsi Dooger]

23. 7 points or less is multiple times more likely than 10+

Not remotely close. There are 9% more conservatives than liberals in this country. That basic reality prevents the type of slant everyone is so comfortable proclaiming. This is a general election, not a primary where opinions change like water in a tank. Those rigid stacks of self-identified conservatives in each state regulate Biden's upside. In a state with 40% conservatives Trump will receive the typical 82-85% of that conservative vote, pushing him to 33-34% already before any liberals or moderates are included at all.

I don't care about polls in relation to ideological logic. In a polarized nation both sides will turn out, and that 9% advantage of conservatives over liberals prevents a huge Biden win. It is the same reason a wave election in 2018 was 8%, not double digits, and the same reason neither Clinton in 1996 or Obama in 2008 managed anything close to 10 points despite a huge national tilt in our favor.

I realize 538 has a relatively high belief in 10+. I believe they are wrong. Hispanics voting for the incumbent is another reason the Biden upside is not as high as it should be. Hispanics always display unusual belief in the incumbent, as I've pointed out for years.

I was arguing the same thing here in 2012 when late adjusters were proclaiming Hillary's 3 or 4 point late poll lead was actually at least 6 or 7.

If Trump's approval were still at 40%, then I'd pay more attention to an assertion of 10+

[Rule of Claw]

24. You are just wrong.

I show math, you argue. You. Argued. With. Math.

Indisputable mathematical computations based on decades of demographic trends, facts, data, I just shake my head.

[Drunken Irishman]

28. lol okay buddy. Please inform The Economist they're wrong too.

...because their forecast has Biden winning 53.5-46.5.

That's two that project Biden around 7 points.

But hey, I'm glad they're wrong and you're right.

[Drunken Irishman]

25. Sorry but at this point I don't see evidence that it's much more likely.

Right now, 538 projects roughly a 53-46 popular vote win. I think that's far more likely than a ten point win.

[greenjar_01]

22. Silly