What I take away from that article is that it’s the close nature of the nature of the elections rather than the polls which is the problem. If the polls predict 60/40 and we end up with 58/42 or 62/38 then no one will care. But if the polls say 50.5/49.5 and we get 49.5/50.5, then we question why the polls missed, even though that outcome is within their margin of error.
This is exactly correct. The poll analysts are quite close in their predictions. They’re much closer now than they would have been in 1984, but in 1984 a five-point miss would have been irrelevant.
I did not mean to imply 1 in 6.
With weighing by recalled vote, the Selzer error goes from 16 points down to 7 points. The biggest factor looks to be Iowa Trump supporters disproportionately being unwilling to answer a call from the Iowa Poll.
Trump now says he wants her investigated for election fraud. I wonder if this is totally from this one poll, or if there already was some feeling that the famous Ann Selzer was a closet Democrat. She did hang out with never-Trumpers before the election.
Of course, if being a Democrat is now a crime, we are in for it.
If he’s operating on the theory that polls showing Harris ahead hurt him (which is dubious – it seems just as likely that polls encourage turnout from the side that thinks they’re behind and depress it from the side that thinks they’re winning), shouldn’t he have been encouraging his supporters to participate in polls? I mean, either the bandwagon effect is real, or it isn’t. If it’s real, it makes sense to try to manipulate it for your own advantage; if it doesn’t, Selzer hasn’t done his campaign any harm, and may even have helped it.
… Never mind, I’m looking for internally consistent logic from Trump. Forget it.
Trump probably never heard of her prior to the poll.
Trump is, let’s bear in mind, an evil, vindictive, stupid man, so of course he’d say something like that, the asshole.
That’s for another thread tho.
I think people are not understanding the purpose of polling. Polling conveys a ton of information beyond the winner. Polling told Democrats that they had a Latino problem and a young men problem. Polling told us that the economy and immigration were very important issues to voters and they trusted Republicans more than Democrats on them. Polling told us what the battleground states were and they told us Biden was a non-viable candidate. If Democrats listened and reacted to the polling more, perhaps the outcome would have been different.
That seems to imply that she retired because she was wrong, which is not true.
In an op-ed for the Register, Selzer said that her decision to end her polling work predated the erroneous results.
“Over a year ago I advised the Register I would not renew when my 2024 contract expired with the latest election poll as I transition to other ventures and opportunities,” Selzer wrote.
People who make claims like this are in such a political bubble they have no idea what the world of an undecided voter looks like. These are the people searching for “did biden drop out” on the day of the election, their consumption of political news is haphazard and idiosyncratic, even if they heard news of this poll, they lack the sufficient context to understand why it’s even important. Anyone who is politically savvy enough to even be following polls (an advanced move even in political junkie circles) already decided their vote long before the election started.
Put another way, we just lived through an election of incredibly turbulent news reflected by an almost stubbornly immovable electorate. Trump gets shot, maybe 2 points? Incredible DNC performance by Kamala, a point and a half? Trump getting almost shot again? Was there even a 0.1 point swing? Puerto Rico being called an island of garbage? Maybe a 2 point swing in Latinos. They’re eating the cats and dogs, maybe a point. And now you’re trying to convince me that some random poll that made the news could result in a 10 point swing? There’s simply no plausible mechanism.
I’m not trying to convince anybody — just saying what I think. And I did not claim ten points due to that one factor.
Also, in Iowa, it is not some random poll, but the one and only famous one.
That’s perfect way of putting it. Though Trump won 312-226 EV I think the polls were fairly accurate in saying the contest was incredibly close…and it was close though ultimately all the swing states did go Trump’s way.
The one big miss was the Ann Selzer poll in Iowa…blunder of a Himalayan level. Whatever justification she may put out now it was indeed a huge miss of epic proportions…something people don’t expect from her.
I can’t fathom why she couldn’t get the so-called shy Trump voter in 2024 when she could do the same in 2016 and 2020.
Sad to see the black mark on her credibility at the fag end of her career.
One thing that’s been infuriating me is the same people that were saying 2020 was incredibly close and Biden barely won are now calling Trump’s win a landslide, when there really wasn’t that much difference.
Both were very close imo regardless of the big difference in electoral votes between the winner and loser.
I read somewhere that basically every other pollster looks at past polling accuracy and tries to do a bunch of sample correcting to try to cancel out the misses, whereas Selzer’s approach was to really hone in on the relatively small group of people she was polling and mitigate the misses by really understanding how to build trust and get accurate answers for polling questions.
I wonder if what’s actually going on is the polarization and media distrust is growing each election, and for the analytical pollsters it means that they correct for the previous miss, but, there’s a combination of little things you aren’t able to correct for and a new widening miss. But for Selzer it eventually reached the point where people stopped trusting her polling and suddenly her methods don’t work any more. Not sure why that would’ve happened all at once though.
Democrats just barely win. Republicans get mandates. That’s been the media narrative at least since 2000.
THANK YOU. So many people do not get this.
Agree…these were the top 2 issues to plurality of voters and republicans were viewed by more voters than democrats to handle these issues better than democrats.
Perhaps in the eyes of quite a few voters Kamala had a credibility problem…on the one hand touting her record as a tough on crime prosecutor in california and otoh her support of the so-called “defund the police” issue in the 2020 campaign sounded contradictory…she has a right to change positions I agree…but imo many voters considered her “new” position on these issues just a lip service till the
election and doubted whether she would actually follow through in implementation if elected.
Wrt the economy she was saddled by the increased inflation (till 2023)…yes cost of eggs, meat, groceries and gas was indeed important to lots of voters in the swing states and they voted on kitchen table issues…no amount of ads saying economy actually improved worked when people saw the grocery prices each time they shopped. Inflation hits all people whereas creation of jobs impacts lesser numbers…unfortunate but true.
And lastly…the negative ad “Kamala is for they/them, President Trump is for you” did indeed have an impact.
Kamala I feel did the right thing by not focussing too much on the trans issue in the campaign but unfortunately the republicans did manage to push it as a central theme of their campaign and it worked exactly how they intended.
My apology for the multiple quotes …can not edit and remove
I think I fixed the multiple quotes in my above post. My apology again…I am fairly new to sdmb as a poster
I didn’t know mandatory voting was a thing until a recent Daily Show segment on Australia. They have a $20 fine if you don’t vote. I think it’s a great idea. Had there been a fine in 2016 I would have voted instead of skipping because I didn’t want either.
Great for Republicans, as they have more low propensity voters.
Let’s wait until MAGA is clearly behind us.
P.S. I wonder what happens to polling when people are fined for not voting. I think that, to match the voting situation, pollsters would have to pay for participation. Otherwise, there would be a tremendous pollster bias in a Democratic Party direction.