Trump Is Expected To Be The Next President
... according the the most respected and usually accurate polling, out today.
The most current, just released, and final New York Times-Sienna poll has just come out and, alarmingly, Harris and Trump are tied at the national level. Normally, if a Democrat is tied at the national level, they will lose the electoral college. That basic math has been true for many cycles, and is almost certainly true today. Polls are not predictions of the future, but rather, samples of very recent thinking and behavior of a sample of voters. But we use them to predict the near term future, and based on this, we will be electing Donald Trump for a second term as President, and we as a nation will choose to end our democracy. Good people will be imprisoned or killed, bad people will rise to global domination, we will all suffer, and we will all ask ourselves, “why did we bring children in to this world.”
Or the polls could be wrong. But before you think that or say it out loud, let me warn you: Almost everything I hear people say about “polls being wrong” is stupid. “They use land lines.” No they don’t. “Pollster don’t sample correctly,” lets see your PhD in statistics. “They use the wrong weighting,” what weighting do they use, do you know? No. And what should they use, do you know? No. And so on. I hereby scold you for thinking that you know more than professional pollsters about how they do their jobs. Get real.
Having said that, polling might be wrong. I have extensive comments on that below. Way down below, in the weeds. I don’t offer an alternative interpretation of the polls there, but I give you enough wishywashyness for you to be able to talk yourself off the ledge between now and when the results of the election are actually known. After that, we go to Plan B. Or Plan A.
By the way, my current and probably final prediction for the outcome of this race is that Harris/Walz will garner 277 electoral votes, the Democrats will take back the House, and the Democrats will retain the Senate. That is based on a state by state analysis, and also uses polls. Am I saying the polling is wrong? No. the polling says either candidate can win, and either party could take the House or the Senate. The whole damn election is 50-50. So the polls will be correct no matter what happens. But, the polls don’t matter once the election is over. What we do after that election will matter.
Plan A and Plan B
If MAGA takes over this country, will have to fight them in any way we can. If MAGA does not take over this country, we need to immediately organize and grow in strength in order to meet the following modest objective: Ensure that the next election, and maybe the next one or two after that, are not the most consequential election of our lifetime, yet again. We need to end that constant threat. My local political groups already have meetings planned in November and December to advance either of those objectives. Do your groups have these plans in place? It isn’t hard. Book a room in your local library, or wherever you usually meet, now.
See you on the other side.
Appendix: Is polling broken?
Here are a few words about the quality of polls, based in large part on some opinion writing by Nate Cohen in the New York times. I wont’ give in-text links because they will be annoying to non-subscribers, but I’ll put references to those pieces at the end of this essay, and you can find them how you might, if you are interested.
One theory is that a goodly number of Trump-friendly voters are reticent with respect to polling, so they don’t show up in polls. So, for example, the current NYT-Sienna poll, which shows an even split between Trump and Harris nationally, really means that Trump will win the popular vote nationally this election.
This theory works with the fact that midterm election polling is more accurate than presidential election polling. The MAGAs who are poll-reticent are not engaged voters, so they show up in midterm polls as murky unknowns, then they don’t vote in those elections, so their existence is irrelevant. But during a presidential election in which Trump is a candidate, they do vote and can decide the outcome.
There is strong evidence for this theory. This theory, according to Nate Cohen, unifies the explanation for polling limitations in both 2016 and 2020.
One problem with the unified theory is that 2020 – both the election itself and the campaign season writ large – happened in a Covid year. Having one theory marry two very different cycles is a little uncomfortable. This alternative explanation is weedier and more complicated, and it is thus more likely to be true. (No, Occam’s razor does not apply to most things – the simplest explanation is the one most likely to be wrong in history and politics. Duh.)
In one version of this theory, 2016 polling had reached a nadir in quality for a range of reasons, but polling recovered methodologically by 2020. Therefore, the two years can’t be compared. The mistaken polling in 2016 was because poling sucked. In my view, this is an insufficient explanation, because I think the higher level polls in 2016 all agreed with each other, and effectively predicted the outcome: The chance of each candidate winning was about even, and in fact, one of the two candidates did win.
Cohen has an opposite and probably more correct view, suggesting that “’gold standard’ national polls were pretty good in 2016, while they were terrible in 2020.” The difficulties in the polls, according to him, had to do with people entering the undecided category because they were traditionally Republican voters who had a hard time with Trump. We might have to trust that this could have been true, but I feel like most Republicans have been waiting for the Messiah since Goldwater, and Trump is that Messiah. That doesn’t wash with any buyer’s remorse theory. I’d rather look at the “I’m a true no partisan saint of a voter” class for these difficulties. Either way, same idea. The polling methodology isn’t the problem so much as the subjects of polling being reactionary in the context of being polled.
Cohen points out other differences between 2016 and 2020. In polls, the information that comes from a specific individual (who is asked questions in the polls) is assumed to be biased because of inadequacy of the sampling, and these biases are adjusted for. This is the correct way to do polling, and any knee-jerk reaction against that is likely based on ignorance of how polling and sampling is done. But the adjustments matter a great deal and can be wrong (or right). Cohen points out that there was a difference between these two years in how education level was adjusted. Polls also may have accounted for late deciders differently and there is some evidence that late deciders were more common in the 2016 cycle.
Even though Trump lost in 2020, the polls had suggested that he was going to be trounced, but he wasn’t. He almost won in some of the swing states where he was expected to lose badly. Keep this in mind. Polls doing better or worse is not a matter of whether they predicted which of the possible outcomes happened, but rather, what spread of votes they predicted, In this way, the 2016 polls were “worse” in predicting who won, but the 2020 polls were worse in predicting the distribution of votes. And, I restate: Polls are not predictive tools, but samplings of recent thinking or behavior.
The Pandemic effect, according to Cohen, included the fact that response rates went way up in 2020. I guess people sitting in isolation were happy to answer the phone and talk to somebody! But if one political party is in stupidly in denial of a pandemic (that would be the Republicans) and the other is responding intelligently to the pandemic (that would be the Democrats), that could be a problem for polling. Polling agencies are not naive. They would be aware of this. But exactly how to you make that adjudgment, mathematically, in a way that adjusts the raw data correctly? Give me a dozen pandemics and I’ll figure that out. One pandemic, in the middle of it? There is no really good way to do that.
Today, for the 2024 election, know this: Pollsters have overhauled their methods based on knowledge and experience gained over the last two cycles. That is good. But the problem is, the last two cycles were very different from each other, and it may be difficult to know if a change in how adjustments are made is better than the old way. Polls now use a wider range of data collection, use new methods of weighting results, there is a new generation of clever pollsters taking over for the old guard, and there is more data than ever to use as the basis of adjusting raw data. Read Cohen’s opinion items in the NYT for details.
By Nate Cohen:
Two Theories for Why the Polls Failed in 2020, and What It Means for 2024. 22 October 2024.
How Polls Have Changed to Try to Avoid a 2020 Repeat. 23 October 2024.
Leave a comment, so they can find you later.
Sad to know there are so many people willing to sacrifice rights for women, minorities, and others, just to be able to say to their racist friends “No, I didn’t vote for the uppity n&*&*r either.”
I've heard from several sources that Republicans are publishing dozens of skewed polls to make it seem as if Trump and Harris are neck-and-neck. These sources include Marc Elias of Democracy Docket and, most recently, Rick Wilson of The Lincoln Project.
The theory is that this would encourage Trump supporters to vote, and if Trump loses (which I expect) would provide more pretext for claiming that the election was stolen from him and filing legal challenges on that basis.