The closest election of the century is getting closer and closer. Two weeks ago, 538’s presidential election predictions gave Vice President Kamala Harris just a 58 in 100 chance of defeating former President Donald Trump on November 5. Now, with just 18 days until Election Day, our forecast gives Trump the bare minimum. advantage in the race with 52 chances out of 100 to win.*
You might be tempted to make a big deal about our forecasts “swinging” toward Trump, but it’s important to remember that a 52/100 chance for Trump isn’t that different from a 58/58 chance. out of 100 for Harris – both are little better than a toss-up for the leading candidate. Although Trump has undeniably gained ground over the past two weeks, a few strong polls in Harris’ favor could easily put her back in the lead tomorrow. Our overall characterization of the race – that it is a toss-up – remains unchanged.
The reason our forecasts are tight is because the polls are close – well within a range where even a small polling error could be decisive. According to our polling averages, the margin between Trump and Harris is 2 percentage points or less in the seven key swing states (Arizona, Georgia, Michigan, Nevada, North Carolina, Pennsylvania and Wisconsin). And in our national polling average, Harris leads Trump by just 2.0 points. That’s narrower than the margin between Trump and President Joe Biden on any day of the 2020 campaign, and narrower than the margin between Trump and former Secretary of State Hillary Clinton on any day between Oct. 1 and today (Due to the Electoral College’s bias toward Republicans, our model currently estimates that Harris will likely need to win the national popular vote by 2.0 points to be favored and win a majority of the electoral votes. )
For the most part, recent high-quality polls have only confirmed the close nature of the race. An ABC News/Ipsos poll conducted October 4-8 found Harris up 2 points nationally. A survey conducted Oct. 1-10 by Marquette University Law School, one of the most accurate and transparent pollsters in America, found it up 3 points. The list goes on: YouGov/CBS News recently gave Harris a 3-point lead; Marist College gave him a 5-point lead; Fairleigh Dickinson University has a 3-point lead. Fewer pollsters gave Trump a national lead, but there were a few: Beacon Research/Shaw & Co. Research/Fox News found Trump up 2 points; Hart Research Associates/Public Opinion Strategies/NBC News gave him a lead of 1.
In national polls, the race was not only close, but also stable. In fact, Harris’ current margin of 2.0 points in national polls is quite similar to that of two months ago, on August 18. The crosstabs have also barely changed. For example, among black and Hispanic voters, support for Harris has increased or decreased by less than 2 points since she joined the race. Sure, those 2 points could be decisive, but it’s remarkable how little has changed over the course of the campaign.
State polls, however, showed comparatively more movement — and most of it against Harris. Compared to her numbers from earlier this month, Harris lost ground in all seven key states. This is why his chances of winning according to our predictions have decreased. In US presidential elections, you don’t get points for winning the national popular vote.
Do partisan pollsters distort our averages?
One question sometimes asked is whether polling averages, like 538, are skewed toward Trump because of the influx of polls conducted by Republican-aligned companies. Over the past two weeks, 23 of the 121 polls released in the seven key swing states came from a Republican pollster or sponsor.** Only four came from Democratic organizations, and the remaining 93 were nonpartisan.
While there is always uncertainty about the accuracy of state polls, partisan polling affecting our averages is not one of my biggest concerns. Indeed, we work hard to subtract potential statistical biases from each survey before integrating them into our averages. As you can see in our poll averaging methodology, we adjust partisan polls to account for the fact that these polls are usually a little too good for the party sponsoring them.
And even for nonpartisan polls, we apply what’s called a “house effects” adjustment that accounts for whether a pollster leans more Democratic or Republican relative to their peers (whether due to partisan leanings of its directors or simply methodological choices which generally produce more liberal or conservative samples). For example, if a pollster’s polls have consistently been 2 points better for Trump than the polling average, after controlling for factors such as poll population (likely voters versus registered voters or all adults) and mode (e.g. live phone, online panel, SMS, etc.), we adjust these polls by 2 points in favor of Harris.
Finally, we give less weight to polls from pollsters without a 538 rating and to pollsters that release a large number of polls in a short period of time. This ensures that pollsters who “flood the area” with polls do not have an outsized influence on our averages.
One test of whether these adjustments are working properly is to see what the 538 averages would look like if we didn’t include surveys from companies that are supposed to be trying to influence the averages. The table below presents the results of this comparison. In the second column you will find the averages from 538 as of 4 p.m. Eastern Time on October 18. The third column shows what our averages would have been at that time if we had removed these polls but otherwise ran the same algorithm we usually use:
As the table shows, this does not significantly change our averages. In most places, the pollsters in question are indeed more pro-Trump than other pollsters. However, this has only a slight effect on our averages, bringing them closer to Trump by only 0.3 points on average. (The biggest difference is in Pennsylvania, where our published average gives Harris a 0.1 point lead over Trump, but the nonpartisan average gives her a 0.9 point advantage.) That’s not a difference significant in a world where the average poll error in presidential elections. elections is 4.3 points, and that’s small enough that it could easily be attributed to sampling error or a methodological factor other than partisan bias. For comparison, our averages regularly change by 0.1 to 0.3 points on a daily basis, and we do not recommend anyone to take these changes into account.
Granted, this says nothing about the companies’ motivations (we like to stick to the data here at 538) or what’s going on behind the scenes (we don’t really know). What we can say is that over the long term, averages from more inclusive polls have generally outperformed averages that use a smaller set of pollsters or an inconsistent weighting and adjustment methodology. One of the reasons we have all these rules and methods is to have confidence in our work even when we have doubts; it means sticking to our rules even though we may be tempted to make an exception.
But importantly, this doesn’t mean we should review all polls uncritically or take a closer look at how pollsters generate their data. In this case, our empirical examination of how our averages work reassures us that things are working as expected.
What if the elections took place today?
In conclusion, the presidential race is close no matter how you divide it. This is both because the polls are currently extremely close and because there remains uncertainty over how the final two and a half weeks of the campaign will play out. As we have written, the race can change quickly in the final weeks. For example, in the final weeks of the 2016 and 2020 campaigns, polls rose by about 4 and 2 points in Trump’s favor, respectively.
The added value of our forecasts over our polling averages is that they attempt to quantify the likelihood of such a change – and how inaccurate the polls might be, even on Election Day itself. But it also means that as the time left for people to change their minds decreases, forecast uncertainty about the race also decreases — and very quickly.
But since the polls are very close right now, if they don’t move, the forecast won’t change either. If you run our presidential election forecast in “nowcast” mode – in which we pretend the election is today and remove any volatility related to the number of days remaining until November 5 – Trump still has a 52 chance out of 100 to win. .
Finally, I must remind you that a close election in terms of odds does not necessarily mean that the final result will be close. If the polls fail by a historically normal amount, either candidate could easily win with 300 electoral college votes. While we can say that this is one of the closest elections in modern history in the polls, we cannot guarantee that it will go down in the history books as one of the closest tight in terms of votes.
Mary Radcliffe contributed to the research.
Footnotes
*All issues of this article are as of October 18 at 4 p.m. Eastern Time.
**Includes all surveys added to our database from October 5 to October 18 at 4 p.m. Eastern Time. Partisan polls are defined as polls from a pollster or sponsor marked as partisan in our database, including internal polls conducted for a campaign.
