PA Monmouth - Biden +4

[MT Treasurer]

This isn't good for Biden.

PA was three points more Republican than the "nation as a whole" in 2016. It makes sense that a slight R trend (likely) would result in it being 4 points more R this time, so +4 in PA is absolutely in line with a +8 Biden lead nationally. If this is a bad poll for Biden, I don’t want to know what a bad poll for Trump looks like.

[kph14]

Only 400 people polled?  In a state like Pennsylvania?

Small sample sizes causes quirky results. Anyone still remembers the NYT live polls? With those there was still a lot movement between a sample of 400 and 500

[American2020]

https://twitter.com/NateSilver538/status/1301174612378759169

[Gass3268]

The two LV turnout models have the race even closer at Biden+3 and Biden+1  

LV turnout is always bad for Democrats in the beginning and it almost always narrows by election day.

[Skill and Chance]

This isn't good for Biden.

PA was three points more Republican than the "nation as a whole" in 2016. It makes sense that a slight R trend (likely) would result in it being 4 points more R this time, so +4 in PA is absolutely in line with a +8 Biden lead nationally. If this is a bad poll for Biden, I don’t want to know what a bad poll for Trump looks like.

Yes, but do you think Biden can win the election without PA?

[Woody]

https://twitter.com/NateSilver538/status/1301174612378759169
https://twitter.com/NateSilver538/status/1301177661927153664

[Xing]

Honestly, Biden +13 was never realistic. This is not an unrealistic result, and it still suggests Biden is favored. If his poll numbers continue to drop, then he'll be in trouble, but I don't think the doom and gloom is yet warranted.

[wbrocks67]

I can't take this poll seriously when it has Republicans winning the House swing counties by 10%.

[Fmr. Gov. NickG]

Usual Monmouth caveat of a very small sample and high margin of error.

I urge this board not to overreact - Biden 49/45 and Biden 53/40 are essentially the same result when your margin of error is 5%.

Excellent methodology (which Monmouth has) doesn't defy this.

They really aren't the same result at all.  Just because the confidence intervals of those two margins overlap doesn't mean the intervals are the same.

If Biden is ahead 53/40 in a random sample, it means there is almost zero probability that Trump is actually ahead among the population.  If Biden is only ahead 49/45 in the sample, there is a decent probability Trump is ahead among the population. And this seems like an enormous difference to me.

[Dr. Arch]

Contextualize this poll in that it's right after the RNC, and this is not bad. Let's see what happens after that fluctuation dies off after Labor Day.

[Pollster]

https://twitter.com/gelliottmorris/status/1301176806054260737?s=20

[MT Treasurer]

This isn't good for Biden.

PA was three points more Republican than the "nation as a whole" in 2016. It makes sense that a slight R trend (likely) would result in it being 4 points more R this time, so +4 in PA is absolutely in line with a +8 Biden lead nationally. If this is a bad poll for Biden, I don’t want to know what a bad poll for Trump looks like.

Yes, but do you think Biden can win the election without PA?

It’s not likely, but by no means impossible if he gets AZ and NE-02 (likely) and then some combination of MI/WI/FL/GA (or just TX, of course). MI and WI would be very iffy, but the Sun Belt route has always been an underrated path for Biden/Harris.

(That said, I don’t think the national race will tighten further, so this is probably as close to Trump's best-case scenario as you can get in PA, unless the state trends even more R than I’m expecting it to, which is certainly not out of the question.)

[Virginiá]

I don't really get why everyone is freaking out so much. Aside from conventions creating some movement, it's not uncommon for the race to start breaking one way in October, when most people start tuning in. Polls right after the conventions, which have long produced some benefits for each candidate, should not drive people insane.

Look at Obama's numbers in August/September 2008:

https://www.realclearpolitics.com/epolls/2008/president/pa/pennsylvania_mccain_vs_obama-244.html

Note that the recession didn't just start in the fall either, it was ongoing well before that.

It's been a very stable race for many months, and the fundamentals are definitely not on Trump's side. Maybe if Biden is still teetering on the edge in mid-October, it would be time to panic, but now? I mean... no.

[Eraserhead]

Biden doesn't need PA to win. He will be OK.

I would say he probably does.

[wbrocks67]

Comparing their July/September polls:

Indies
July: Biden +21
September: Tied

Males
July: Trump +2
September: Trump +19

Females
July: Biden +26
September: Biden +24

18-49
July: Biden +31
September: Biden +9

50-64:
July: Trump +13
September: Trump +9

65+:
July: Biden +10
September: Biden +11

Swing Counties
July: Biden +19
September: Trump +2

Whites
July: Biden +2
September: Trump +7

Nonwhites
July: Biden +60
September: Biden +57

So it’s funny how Females stayed totally consistent month to month, and Biden actually even gained among 50-64 year olds and 65+ year olds (with the RNC/Kenosha you would expect otherwise), while he simultaneously collapsed with Men, Indies, and 18-49 year olds.

Considering a few demographics stayed the same while some shifted A LOT, seems like a lot of noise to me, especially the Indies and 18-49 #s. That change is too huge when the demographic you'd expect to be swayed here wasn't at all (old people)

[ηєω ƒяσηтιєя]

Not the best poll for Biden. However, he's still leading by an OK margin in a state Trump won in 2016 and in a poll taken right after the RNC. I'll take it.

[afleitch]

'In 12 congressional districts where the winning margin for either party was greater than 15 points in 2018, the Democrats lead by 8 points (51% to 43%). In the six most competitive districts, though, the Republicans lead by 10 points (50% to 40%). In July, the Democrats held small leads in both the safe seats (49% to 46%) and the competitive ones (48% to 43%).'

Wouldn't that include Trump blowout counties too? I don't understand this part.

[wbrocks67]

What's odd to me is that the GCB somehow stayed the same? And the fact that Biden's favorability went up, even when this wasn't a very Dem sample.

[Skill and Chance]

Comparing their July/September polls:

Indies
July: Biden +21
September: Tied

Males
July: Trump +2
September: Trump +19

Females
July: Biden +26
September: Biden +24

18-49
July: Biden +31
September: Biden +9

50-64:
July: Trump +13
September: Trump +9

65+:
July: Biden +10
September: Biden +11

Swing Counties
July: Biden +19
September: Trump +2

Whites
July: Biden +2
September: Trump +7

Nonwhites
July: Biden +60
September: Biden +57

So it’s funny how Females stayed totally consistent month to month, and Biden actually even gained among 50-64 year olds and 65+ year olds (with the RNC/Kenosha you would expect otherwise), while he simultaneously collapsed with Men, Indies, and 18-49 year olds.

Considering a few demographics stayed the same while some shifted A LOT, seems like a lot of noise to me, especially the Indies and 18-49 #s. That change is too huge when the demographic you'd expect to be swayed here wasn't at all (old people)

This is not that surprising to me.  The Trump/GOP message is more aimed at families in their peak earning years (~age 35-55) this time than at retirees.  Biden clearly sees a Trump weakness with retirees he is trying to exploit, while he seems to be pricing in a weaker performance (turnout and margins) with 18-29 year-olds than Obama got.

[Pollster]

Usual Monmouth caveat of a very small sample and high margin of error.

I urge this board not to overreact - Biden 49/45 and Biden 53/40 are essentially the same result when your margin of error is 5%.

Excellent methodology (which Monmouth has) doesn't defy this.

They really aren't the same result at all.  Just because the confidence intervals of those two margins overlap doesn't mean the intervals are the same.

If Biden is ahead 53/40 in a random sample, it means there is almost zero probability that Trump is actually ahead among the population.  If Biden is only ahead 49/45 in the sample, there is a decent probability Trump is ahead among the population. And this seems like an enormous difference to me.

This is true, but what you are talking about is much more critical in more fluid races, usually primaries with multiple candidates where there are far more variables likely impacting vote choice, as we saw during the Iowa caucuses and early primaries this year. Polls are much better at measuring levels of support rather than margins (some pollsters argue that polls can't even be predictive of margins, though this is a debated premise) and MOE is much more important to the former than the latter, especially in 1-on-1 races.

[wbrocks67]

Comparing their July/September polls:

Indies
July: Biden +21
September: Tied

Males
July: Trump +2
September: Trump +19

Females
July: Biden +26
September: Biden +24

18-49
July: Biden +31
September: Biden +9

50-64:
July: Trump +13
September: Trump +9

65+:
July: Biden +10
September: Biden +11

Swing Counties
July: Biden +19
September: Trump +2

Whites
July: Biden +2
September: Trump +7

Nonwhites
July: Biden +60
September: Biden +57

So it’s funny how Females stayed totally consistent month to month, and Biden actually even gained among 50-64 year olds and 65+ year olds (with the RNC/Kenosha you would expect otherwise), while he simultaneously collapsed with Men, Indies, and 18-49 year olds.

Considering a few demographics stayed the same while some shifted A LOT, seems like a lot of noise to me, especially the Indies and 18-49 #s. That change is too huge when the demographic you'd expect to be swayed here wasn't at all (old people)

This is not that surprising to me.  The Trump/GOP message is more aimed at families in their peak earning years (~age 35-55) this time than at retirees.  Biden clearly sees a Trump weakness with retirees he is trying to exploit, while he seems to be pricing in a weaker performance (turnout and margins) with 18-29 year-olds than Obama got.

Clinton won that age demo by 10% and Casey 2018 won it by 21%. So I'd be... very shocked that Biden would be doing worse than both of them, especially Clinton. IDK if Biden +31 was correct either, but I'd imagine it'd be closer to somewhere in the middle.

[This user has not been convicted of 34 felonies]

Biden doesn't need PA to win. He will be OK.

I would say he probably does.

He absolutely does. Without PA, he needs to win WI, MI, and Arizona *to tie* which is unlikely. Losing PA almost guarantees losing at least WI, so then he would need to win FL to make up for that. It's much easier to win PA, WI, and MI than it is to put all your eggs in the Florida basket--we all know how that will turn out.

[MT Treasurer]

This is not that surprising to me.  The Trump/GOP message is more aimed at families in their peak earning years (~age 35-55) this time than at retirees.  Biden clearly sees a Trump weakness with retirees he is trying to exploit, while he seems to be pricing in a weaker performance (turnout and margins) with 18-29 year-olds than Obama got.

I really think FL will be unusually important this year, more so than in 2008/2012 and (ironically) even 2016. I know that the conventional wisdom is that it’s a must-win for Trump (I don’t necessarily disagree), but I also think all bets are off if Trump takes it. Given how fast-counting a state it is, we’ll get a pretty good idea of the election relatively early in the night with the FL results + the margins in rural IN/KY (in addition to places like Hamilton IN, of course).

[BigSerg]

Biden +1
https://twitter.com/Politics_Polls/status/1301178432274006022?s=19

[Skill and Chance]

Hmmm... is EV by CD just a statute that could potentially be repealed by the one party governments in NE and ME, or is it in the state constitutions?  SCOTUS historically has a rule against altering election procedures too close to an election.  Would they strike down an attempt by either state to repeal it in October?  Could it even be repealed in late November/December and apply retroactively to this election?  Asking for John Roberts.