NYT/Ipsos: Biden +3

[kwabbit]

It is quite funny though how NYT commissioned this poll but basically ignored it, b/c it didn't comport with the narrative they were trying to push.

They would rather lead with completely unrealistic statewide numbers that almost 100% wont pan out than a national number that makes more sense, just for the clicks.

Given the choice between a n=407 national poll, of lower quality data with no direct h2h, versus 6 n=700 polls of swing states, with their favored question order/wording and higher quality data, of course they went with the latter. It wasn’t to push a narrative, this was just an experiment where NYT didn’t even intend to report the top lines. You have to go to the Ipsos website to get those.

[wbrocks67]

It is quite funny though how NYT commissioned this poll but basically ignored it, b/c it didn't comport with the narrative they were trying to push.

They would rather lead with completely unrealistic statewide numbers that almost 100% wont pan out than a national number that makes more sense, just for the clicks.

Given the choice between a n=407 national poll, of lower quality data with no direct h2h, versus 6 n=700 polls of swing states, with their favored question order/wording and higher quality data, of course they went with the latter. It wasn’t to push a narrative, this was just an experiment where NYT didn’t even intend to report the top lines. You have to go to the Ipsos website to get those.

I disagree - I think if the results showed Trump up in these, they would've been featured much more prominently. Since Biden was up, they kind of just kicked it to the side and ignored it. I'm not saying this had to be the #1 story on their site, but NYT still paid for and commissioned this poll. The fact that it was relegated to a single story that people had to *find* I think says volumes.

[AncestralDemocrat.]

There cannot be this huge of variance between polling. At least not realistically. This election is going to look like 2020, period.

Literally all evidence to the contrary, but OK.. 

Biden’s favorables have collapsed (and his overall approval are at 30 year lows going into a presidential election) since 2020 and Trump is polling substantially better.

Biden's favorables are worse than Trumps here and yet Biden is leading Trump. That in itself continues to prove that favorables/approval are not everything this election and do not tell the whole story - as most ofus have been saying.

This tracks though with the ABC Ipsos poll as well

Its not the entire story.. but given the deluge of horrific state polling for him.

It obviously is a factor.

[Spectator]

Basically pick your own adventure with what polls you want to believe. 2022 was the only national election we had since 2020 and that was basically 2020 on steroids.

[Amenhotep Bakari-Sellers]

It's the same 303 map as it was in 22 and it will be 26/28

[President Punxsutawney Phil]

It is quite funny though how NYT commissioned this poll but basically ignored it, b/c it didn't comport with the narrative they were trying to push.

They would rather lead with completely unrealistic statewide numbers that almost 100% wont pan out than a national number that makes more sense, just for the clicks.

That tracks.

[Sir Mohamed]

So NYT wants to tell us Biden is up nationally by three while he's trailing in all battleground, even by robust margins in the Sunbelt Stack? Polling is a joke.

It's Iposes here while the state polls are Siena. But it's true that both can't be true at the same time. Polling has been a hot mess this entire cycle. As of today, the race is pretty much a tossup.

[kwabbit]

It is quite funny though how NYT commissioned this poll but basically ignored it, b/c it didn't comport with the narrative they were trying to push.

They would rather lead with completely unrealistic statewide numbers that almost 100% wont pan out than a national number that makes more sense, just for the clicks.

Given the choice between a n=407 national poll, of lower quality data with no direct h2h, versus 6 n=700 polls of swing states, with their favored question order/wording and higher quality data, of course they went with the latter. It wasn’t to push a narrative, this was just an experiment where NYT didn’t even intend to report the top lines. You have to go to the Ipsos website to get those.

I disagree - I think if the results showed Trump up in these, they would've been featured much more prominently. Since Biden was up, they kind of just kicked it to the side and ignored it. I'm not saying this had to be the #1 story on their site, but NYT still paid for and commissioned this poll. The fact that it was relegated to a single story that people had to *find* I think says volumes.

It would've had a tiny headline at best. They didn't even poll a head to head, it obviously wasn't supposed to be given much press, people just fished for headline numbers to point out NYT hypocrisy. Crap like this is going to kill the NYT polling outfit. They conducted a pretty interesting experiment to gauge the effect of question order on third-party support, but basically no one responds to that and only accuses them of trying to push a pro-Trump narrative. Monmouth has already bowed out of real toplines, ABC canceled their high-quality partnership with WAPO because it wasn't worth it fund a high-quality poll if they only got criticism, NYT could reasonably do the same if this stuff continues.

[Oregon Eagle Politics]

So NYT wants to tell us Biden is up nationally by three while he's trailing in all battleground, even by robust margins in the Sunbelt Stack? Polling is a joke.

NYT is predicting the tipping point state to be 4% to the right of the nation, which is realistic as it happened in 2020.

[BlueSwan]

There cannot be this huge of variance between polling. At least not realistically. This election is going to look like 2020, period.

Sure there can. MoE is much bigger than people think and that doesn't even take outliers and bad methodology into account.

If MoE in a poll is 3 points, that means that a Biden +3 result is actually that the result is within Trump +3 to Biden +9 with 95% confidence. Again, that still leaves those remaining 5% and whatever you attribute to bad methodology.

So if this poll actually reflects a Trump +3 electorate then that is perfectly in line with the state polling (which also has MoE).