Kamala Harris Will Win This Election
Throughout this election, I have been much more sanguine on the prospects for Democrats than most. The prevailing wisdom has this as a tight-race. The WaPo has Harris a very slight favorite (as of Sunday afternoon); ABC News-538 has her a very slight underdog (same time frame). So, naturally, this is a perfect time to stick my neck out and declare a winner: Kamala Harris.
Based On What?
That’s a fair question. After all, the polling aggregators cited above are using the data as best they see it. The WaPo doesn’t include partisan polling operations; ABC does but with caveats based on how it measures poll accuracy. Even the biggest critics of the pundits’ handling of polling (Carl Allen) is only about 60% certain that Harris will win (that translates from Statistical to English as “not certain at all”).
So why am I?
Besides the argument about partisan polling, even the nonpartisan outfits have been trying to correct for their apparent 2016 and 2020 errors (Carl would vehemently disagree about using that term, but they don’t). From what I’ve seen, more than a few of them have in fact been overcorrecting, especially with far too many rural voters in their weighted samples. I’ve taken particular aim at the New York Times/Siena polls over on my podcast’s site, but that’s not the only pollster. I would argue that fear of “getting it wrong” has led to the biggest mistake pollsters can make: herding. and have more on this (NBC).
But the fact that so many polls are reporting the exact same margins and results raises a troubling possibility: that some pollsters are making adjustments in such similar ways that those choices are causing the results to bunch together, creating a potential illusion of certainty — or that some pollsters are even looking to others’ results to guide their own (i.e., “herding”). If so, the artificial similarity of polls may be creating a false impression that may not play out on Election Day. We could well be in for a very close election. But there’s also a significant chance one candidate or the other could sweep every swing state and win the presidency somewhat comfortably, at least compared to the evenly balanced picture in the polls.
While all this was going on, one pollster stood apart: Ann Selzer. Her poll, conducted for the Des Moines Register, was the canary in the coal mine for Hillary Clinton in 2016. She was down seven; she ended up losing by nine on her way to losing the election. Given that Iowa was, up to this point, a leaning-Democratic state, Selzer’s poll was a surprise – and one the Clinton campaign dismissed – to their detriment. Four years later, as Joe Biden carried large leads in swing state polls to Election Day, Selzer once again gave a warning sign. Trump was up seven; he won the state by eight as he nearly repeated his razor-thin 2016 win in the “Blue Wall” states.
This time? Selzer has Harris up three.
So One Good Poll for Harris is the Best One? Seriously?
I get that reaction. I can already hear Chris Saxman laughing, thank you very much. I still think that poll points to a Harris victory.
For one, Selzer’s poll got 2016 exactly right in her state. The weekend before the election, Selzer had Clinton down seven points. On election day itself, exit pollsters found that Iowans who decided after the poll was taken went for Trump by 22 points; they were 10% of the voters. That’s a 2.2% overall shift to Trump; subtract that from the actual 2016 result puts the weekend-before-the-election population at … Trump +7.2%. Four years later, Trump was again up seven points in the Selzer poll. This time the late deciders were too small sample to poll meaningfully, but Trump won by 8.2%. So, yeah, the Selzer poll is a pretty good indicator of what Iowa is doing.
Of course, Iowa is not America. It is, however, part of the Midwest. So I took a look at the rest of the region. I examined the change in outcomes between 2016 and 2020, compared to the Selzer poll (which had no change). Here’s what I found.
| 2016 | Dem. Margin | 2020 | Dem. Margin | Delta | Dem. Margin | ||
| Selzer Poll | -7% | Selzer Poll | -7% | Selzer Poll | 0.0% | ||
| Illinois | 17% | Illinois | 17% | Illinois | 0.1% | ||
| Indiana | -19% | Indiana | -16% | Indiana | 3.0% | ||
| Iowa | -9% | Iowa | -8% | Iowa | 1.2% | ||
| Michigan | 0% | Michigan | 3% | Michigan | 3.0% | ||
| Minnesota | 2% | Minnesota | 7% | Minnesota | 5.6% | ||
| Missouri | -19% | Missouri | -15% | Missouri | 3.1% | ||
| Ohio | -8% | Ohio | -8% | Ohio | 0.1% | ||
| Pennsylvania | -1% | Pennsylvania | 1% | Pennsylvania | 1.9% | ||
| Wisconsin | -1% | Wisconsin | 1% | Wisconsin | 1.4% |
As you can see, only one state moved more than 3 points: Minnesota. A majority moved less than two points. This is a pretty good sign that the trajectory of Selzer’s samples correlates pretty well with the movement of voters in the region.
So What Does That Mean for Tuesday?
Well, if Trump has lost 10 points net in Iowa, than most of the rest of the Midwest is moving away from him, too. So we would expect something like this.
| Currently Projected 2024 Status | Dem. Margin |
| Selzer Poll | 3% |
| Illinois | 27% |
| Indiana | -6% |
| Iowa | 2% |
| Michigan | 13% |
| Minnesota | 17% |
| Missouri | -5% |
| Ohio | 2% |
| Pennsylvania | 11% |
| Wisconsin | 11% |
That said, there is the matter of late deciders. After all, the poll was dropped three days before November 5. If 2016 is any indication, there can still be a shift to Trump of 2 points. While the Don’t-Know/Refused part of the poll doesn’t look like a more Republican bunch (as in it’s fairly evenly divided amongst the parties and genders), I’m not going to say it’s automatic that they’ll split 50-50. The only actual data we have is the 2016 shift, so I’d include that in the final projection.
| Projected 2024 Result | Dem. Margin |
| Selzer Poll | 3% |
| Illinois | 25% |
| Indiana | -8% |
| Iowa | <-1% |
| Michigan | 10% |
| Minnesota | 15% |
| Missouri | 8% |
| Ohio | <-1% |
| Pennsylvania | 9% |
| Wisconsin | 8% |
But All of the States Had Some Shift. Didn’t They?
Yes, they did. So simply assuming an even 10 point movement across the region is not credible. It’s far better to put together a range of outcomes. We can use the above table as the central forecast (as in, it’s coin flip between the results being more Republican or more Democratic). Here are the other assumptions I used to create the best and worst cases for Harris.
- Worst case: use the largest 2016-2020 variation (Minnesota, 5.6%) and subtract it from Harris
- Best case: use the largest 2016-2020 variation (Minnesota, 5.6%) and add it to Harris; remove the post-poll movement to Trump (again, this is Harris’ best case).
Those assumptions lead to these scenarios.
| 2024 (Best) | D Margin | 2024 (Central) | D Margin | 2024 (Worst) | D Margin | ||
| IA (Selzer) | 3% | IA (Selzer) | 3% | IA (Selzer) | 3% | ||
| Illinois | 33% | Illinois | 25% | Illinois | 19% | ||
| Indiana | <-1% | Indiana | -8% | Indiana | -14% | ||
| Iowa | 7% | Iowa | <-1% | Iowa | -6% | ||
| Michigan | 18% | Michigan | 10% | Michigan | 5% | ||
| Minnesota | 23% | Minnesota | 15% | Minnesota | 9% | ||
| Missouri | <1% | Missouri | -8% | Missouri | -13% | ||
| Ohio | 8% | Ohio | <-1% | Ohio | -6% | ||
| Pennsylvania | 17% | Pennsylvania | 9% | Pennsylvania | 3% | ||
| Wisconsin | 16% | Wisconsin | 8% | Wisconsin | 3% |
In every scenario, Harris wins the “Blue Wall” (Michigan, Pennsylvania, and Wisconsin) fairly comfortably – and thus, she wins the election.
That’s Still a Lot to Put on One Poll
Yes, it is. I have my reasons, though. For one, Iowa hasn’t been one of the states on which everyone has been fixated. Thus, it’s less susceptible to the partisan polling and herding issues. Secondly, Selzer’s reputation is clearly well-earned. Finally, I’ve seen multiple elections in which herding has led pollsters astray (UK 2015, US 2016, UK 2017, US 2020, etc.), each had a poll that seemed an outlier (Savanta, Selzer, YouGov, Selzer again) … until it wasn’t.
Selzer is the outlier again. This time, I’m going with the outlier. I’m convinced that Selzer has managed to cut through the partisan junk and the herding to get to what’s really happening. I’m convinced: Kamala Harris will win, and she’ll win the tipping point state with a bigger margin than Biden had.
