IA (DMR/Selzer): Trump +15

[Tekken_Guy]

“Her early polls aren’t accurate because they don’t predict the final result, unlike her final polls,” is something you might say if you have no idea what a poll is.

I think the takeaway should be “Biden has a lot of work to do” rather than “it’s over, Trump is #47”.

[Devils30]

Biden needs the number to be below 10 in Iowa to have a shot in WI, PA, MI. Selzer is top notch and you can't dismiss her polls.

This does lend support to the idea Trump is at or close to his final number while Biden can still get some support back to around 2020 level.

What's more striking is that Trump gets 88 % of Republicans in this Poll.

Conclusion: The notion that 1 in 5 Republicans would not vote for Trump in November is completely baloney + malpractice & debunked. Trump will get 90 % or above of Republican Support come November.

Yeah of course each side should win 90+ of respective partisans. Biden has way more room to get this back to 2020's 8 point margin than Trump has to get to +20.

[Birdish]

“Her early polls aren’t accurate because they don’t predict the final result, unlike her final polls,” is something you might say if you have no idea what a poll is.

I am very well aware a poll is a snapshot. Is it wrong to point out that Selzer's early cycle "snapshots" haven't been all that predictive of the final result?

Edit: Of course I'm not trying to be dismissive of the results, just some resistance to the idea that this poll means Minnesota is about to flip.

[Ljube]

It probably means Wisconsin is Trump+5 and Michigan Trump+2.
Minnesota could be between a tie and Biden+2.

[Respect and Compassion]

Birdish, earlier in this thread, mentioned a possible anti-incumbent bias. The problem is that there is no demonstrable anti-incumbent bias in American polling. Multiple 2018 Democratic incumbents in the Senate did NOT outperform polling - they actually underperformed polling! Ron Johnson was an incumbent in 2022 and did not outperform polling. Also, while Hochul was an incumbent in 2022 and outperformed some polling (e.g. Trafalgar) she did not demonstrably outperform polling by a lot. Schumer also didn't outperform polling in 2022. Speaking of Iowa, there was no demonstrable anti-incumbency bias with Reynolds and Grassley: https://en.wikipedia.org/wiki/2022_United_States_Senate_election_in_Iowa#Polling_2

and also, in the cases where Reynolds and Grassley outperformed polling, it can more easily be explained by the same reason Trump overperformed Iowa polling in 2016, as I'll outline below

The incumbent overperformances in 2022 are mainly due to Democrats being much better at GOTV ops than Republicans that year (e.g. Nevada, Arizona) and pollsters underestimating Republican support with midwestern whites when there wasn't a well funded Democrat running against them (e.g. Iowa). Whitmer outperformed some R leaning pollsters, and Hassan managed to outperform polling in general, but in the case of MI/PA/NH I'm confident it's because pollsters expected more noncollege white turnout than what 2022 actually ended up having.

Here's my theory of how pollsters can end up missing the mark in American elections: when noncollege turnout is higher than expected (e.g. 2016 and 2020) R's overperform, when noncollege turnout is not higher than expected or actually lower than expected, D's overperform. Also, when Democrats put more effort into efficient GOTV operations with mail-ins and the like, D's overperform --> this explains why Trump overperformed in multiple midwestern states in 2020 (but not particularly in New Hampshire), as noncollege white turnout was higher than pollsters expected, it also explains why Mastriano and Dixon underperformed. Also explains the Nevada and Arizona underperformances with R's. I also argue it explains 2018 polling in certain states like Indiana where Donnelly underperformed polling.

I think my theory is much more effective at explaining things than "anti-incumbency bias." I suspect the 'anti-incumbency bias' theory is thrown around because people want the psychological comfort of expecting that Biden outperforms polling this time around.

[The Mikado]

In the last two elections Trump got 51 and 53 in Iowa. This poll is totally conaistent with Trump ending up at 53 or so again or maybe a little higher.

I fully expect most of the third party campaigns to lose a ton of steam over the course of the year as they always do. Something like Trump 53-42 or so in Iowa is believable to me.

We'll see. Having double digit support for Someone Else never pans out.

[Birdish]

Birdish, earlier in this thread, mentioned a possible anti-incumbent bias. The problem is that there is no demonstrable anti-incumbent bias in American polling. Multiple 2018 Democratic incumbents in the Senate did NOT outperform polling - they actually underperformed polling! Ron Johnson was an incumbent in 2022 and did not outperform polling. Also, while Hochul was an incumbent in 2022 and outperformed some polling (e.g. Trafalgar) she did not demonstrably outperform polling by a lot. Schumer also didn't outperform polling in 2022. Speaking of Iowa, there was no demonstrable anti-incumbency bias with Reynolds and Grassley: https://en.wikipedia.org/wiki/2022_United_States_Senate_election_in_Iowa#Polling_2

and also, in the cases where Reynolds and Grassley outperformed polling, it can more easily be explained by the same reason Trump overperformed Iowa polling in 2016, as I'll outline below

The incumbent overperformances in 2022 are mainly due to Democrats being much better at GOTV ops than Republicans that year (e.g. Nevada, Arizona) and pollsters underestimating Republican support with midwestern whites when there wasn't a well funded Democrat running against them (e.g. Iowa). Whitmer outperformed some R leaning pollsters, and Hassan managed to outperform polling in general, but in the case of MI/PA/NH I'm confident it's because pollsters expected more noncollege white turnout than what 2022 actually ended up having.

Here's my theory of how pollsters can end up missing the mark in American elections: when noncollege turnout is higher than expected (e.g. 2016 and 2020) R's overperform, when noncollege turnout is not higher than expected or actually lower than expected, D's overperform. Also, when Democrats put more effort into efficient GOTV operations with mail-ins and the like, D's overperform --> this explains why Trump overperformed in multiple midwestern states in 2020 (but not particularly in New Hampshire), as noncollege white turnout was higher than pollsters expected, it also explains why Mastriano and Dixon underperformed. Also explains the Nevada and Arizona underperformances with R's. I also argue it explains 2018 polling in certain states like Indiana where Donnelly underperformed polling.

I think my theory is much more effective at explaining things than "anti-incumbency bias." I suspect the 'anti-incumbency bias' theory is thrown around because people want the psychological comfort of expecting that Biden outperforms polling this time around.

I'll stop you right there.

I mentioned Selzer's early cycle polling potentially having an anti-incumbent bias. Not polling in general.

[Respect and Compassion]

@Birdish: it's a stretch to assume that, it's more likely the proper explanation is that R-leaning voters are less likely to settle on a definitive response early on.

Note that Biden's favorability in this poll is rather atrocious, which means that the Someone else + Not sure respondants are overwhelmingly respondents who have an unfavorable view of Biden, which is why even with these already high margins, it's more likely that Trump ends up doing even better than +15 in Iowa.

[Redban]

ITT - Biden stooges argue that no pollster may be trusted unless that pollster has never ever ever released a poll, at any point of a race, that didn't match the actual result in exactitude.

[Birdish]

ITT - Biden stooges argue that no pollster may be trusted unless that pollster has never ever ever released a poll, at any point of a race, that didn't match the actual result in exactitude.

You are so hostile. Holy cow.

It's a cycle. Selzer early snapshots aren't always predicative. Grassley in 2022, Trump in 2020 and 2016, Obama in 2012.

It's okay to point it out.

[Birdish]

@Birdish: it's a stretch to assume that, it's more likely the proper explanation is that R-leaning voters are less likely to settle on a definitive response early on.

Note that Biden's favorability in this poll is rather atrocious, which means that the Someone else + Not sure respondants are overwhelmingly respondents who have an unfavorable view of Biden, which is why even with these already high margins, it's more likely that Trump ends up doing even better than +15 in Iowa.

Unless you have me on ignore like @Senator Incitatus appears to have done, you can just quote me. I can't see if you've responded to me If you don't.

Anyway, It's quite possible. I'm just not entirely sure Iowa has hit the point where Biden is going to lose it by 15%.

[President Johnson]

Safe Republican, nothing to see here. It's Colorado in reverse. Iowa might go the route of Missouri and I could see in a 2028 matchup like Harris vs. Vance, the latter breaking 60%.

However, I wonder what exactly has driven all the Obama two time voters away from the Democrats? Cultural issues? It's not like many where just Democratic voters in 2008 and 2012, Iowa voted for Gore and still to the left of the nation in 2004 while barely going for Dubya.

[Tekken_Guy]

Safe Republican, nothing to see here. It's Colorado in reverse.

However, I wonder what exactly has driven all the Obama two time voters away from the Democrats? Cultural issues? It's not like many where just Democratic voters in 2008 and 2012, Iowa voted for Gore and still to the left of the nation in 2004 while barely going for Dubya.

Who are the Clinton/Biden voters who won’t vote for him again?

[Respect and Compassion]

@Birdish: the problem is that Trump was the guy who early Selzer polls underestimated, and he's also the guy running against Biden yet again

Obama was more popular with the same noncollege whites who voted in droves for Trump. Noncollege whites in the midwest are demonstrably difficult to poll.

Is it possible that Trump overperforms a +15 poll in Iowa? While it sounds a bit out there if you think 2016 results & 2020 results comprise this powerful 'gravity well' from which Trump cannot reasonably outperform by any more than a miniscule amount, it does seem possible when you consider the fact that Biden is much more unpopular now than in 2020. Note that Trump2020 already outperformed 2016 results with Hispanics substantially, so it's not far-fetched for Trump2024 to outperform Trump2020 substantially.

I've seen multiple people online treat the 2020 results as the "realistic" or "reasonable" results. That is to say -- when a pollster publishes results close to the 2020 margins&percentages, some people chime in to say "finally, a realistic poll!" However...given how much things have changed since 2020, it is not at all reasonable to expect results to be in line with the 2020 results!

I've also seen some people point out that in certain polls, Trump is at his 2020 election percentages, which means that the remaining undecideds are much more likely to break for Biden. This argument is also likely to be proven false in November, given that in 2020, Trump's percentages in polling was nowhere near his inevitable 2020 election percentages, because multiple undecideds broke for Trump. In 2024, the undecideds in polling are *overwhelmingly* Biden disapprovers, so it's quite likely Trump wins more of the undecideds and thereby adds to his already high margins.

[kwabbit]

ITT - Biden stooges argue that no pollster may be trusted unless that pollster has never ever ever released a poll, at any point of a race, that didn't match the actual result in exactitude.

You are so hostile. Holy cow.

It's a cycle. Selzer early snapshots aren't always predicative. Grassley in 2022, Trump in 2020 and 2016, Obama in 2012.

It's okay to point it out.

Selzer in March isn't predictive of the final outcome but she has a good enough track record where we can accept it as a good marker of where the race currently stands. Polls are a snapshot, not a prediction, and her final polls are just a snapshot of the race very close to the election.

This poll supports the Trump +3 nationally and winning 312 EV prevailing view of the race as it currently stands. That being more likely to be the truth is concerning to Biden since it's obviously better for him to be losing by less than more, even if March isn't hugely predictive of November.

[Birdish]

ITT - Biden stooges argue that no pollster may be trusted unless that pollster has never ever ever released a poll, at any point of a race, that didn't match the actual result in exactitude.

You are so hostile. Holy cow.

It's a cycle. Selzer early snapshots aren't always predicative. Grassley in 2022, Trump in 2020 and 2016, Obama in 2012.

It's okay to point it out.

Selzer in March isn't predictive of the final outcome but she has a good enough track record where we can accept it as a good marker of where the race currently stands. Polls are a snapshot, not a prediction, and her final polls are just a snapshot of the race very close to the election.

This poll supports the Trump +3 nationally and winning 312 EV prevailing view of the race as it currently stands. That being more likely to be the truth is concerning to Biden since it's obviously better for him to be losing by less than more, even if March isn't hugely predictive of November.

I truly wish you guys would stop. I'm aware...

[Amenhotep Bakari-Sellers]

Biden needs the number to be below 10 in Iowa to have a shot in WI, PA, MI. Selzer is top notch and you can't dismiss her polls.

This does lend support to the idea Trump is at or close to his final number while Biden can still get some support back to around 2020 level.

What's more striking is that Trump gets 88 % of Republicans in this Poll.

Conclusion: The notion that 1 in 5 Republicans would not vote for Trump in November is completely baloney + malpractice & debunked. Trump will get 90 % or above of Republican Support come November.

Biden is leading in PA with this poll, Grassley won by 12 and Evers won in 22 so that's wrong

[Amenhotep Bakari-Sellers]

Safe Republican, nothing to see here. It's Colorado in reverse. Iowa might go the route of Missouri and I could see in a 2028 matchup like Harris vs. Vance, the latter breaking 60%.

However, I wonder what exactly has driven all the Obama two time voters away from the Democrats? Cultural issues? It's not like many where just Democratic voters in 2008 and 2012, Iowa voted for Gore and still to the left of the nation in 2004 while barely going for Dubya.

There's no NAACP chapters in Iowa, Barber said triage Sinema, Manchin and Iowa as forgone, even FL has an NAACP chapters, IA is a white state

[Arizona Iced Tea]

Why do people get so riled up about a single poll 8 months before election day. Just aggregate it, interpret it if you want, and move on.

[Ljube]

Why do people get so riled up about a single poll 8 months before election day. Just aggregate it, interpret it if you want, and move on.

Because this is the best poll. All other polls could be wrong.

[Brittain33]

Yup, IA is gone. A 15 pts. margin is the very least I can see for 2024.

IA is basically the R-version of CO.

A 7 points swing to the right is the minimum swing you foresee? That’s a bold move, Cotton.

[Brittain33]

I don’t see what’s wrong with this poll. It’s believable. The most likely scenario is that Iowa is a solid R state from the get go, Trump has consolidated his support, Biden has not, and there’s scope for that to change markedly as we get closer to the election.

[Spectator]

Not a good poll for Biden obviously, but Iowa is obviously a red state now. Probably should ‘t be using it to predict WI/MI/PA.

[Progressive Pessimist]

Biden will get more than 33% here by election day.

Needless to say, Iowa is gone for Democrats.

I would like to remind posters here about discerning other Midwestern states' margins from Iowa though. Remember in 2020 all the panic over Selzer's last poll and how it was a bad onen for Biden in Minnesota and Wisconsin?

[Brittain33]

Biden will get more than 33% here by election day.

Needless to say, Iowa is gone for Democrats.

I would like to remind posters here about discerning other Midwestern states' margins from Iowa though. Remember in 2020 all the panic over Selzer's last poll and how it was a bad onen for Biden in Minnesota and Wisconsin?

Wasn’t it accurately a bad omen for Wisconsin though? It finished so close.