The closest election of the century is getting closer and closer. Two weeks ago, 538’s presidential election forecast gave Vice President Kamala Harris just a 58-in-100 chance of defeating former President Donald Trump on November 5th. Now – just 18 days before Election Day – our forecast gives Trump the bare minimum lead in the race with a 52 in 100 chance of winning.*

You might be tempted to make a big deal about our prognosis “flipping” on Trump, but it’s important to remember that a 52 in 100 chance for Trump isn’t all that different from a 58 chance 100 for Harris – both are hardly better than a coin toss for the leading candidate. While Trump has undeniably gained some ground in recent weeks, a few good polls for Harris could easily put her back in the “lead” tomorrow. Our general characterization of the race – that it is back and forth – remains unchanged.

The reason our forecast is close is because the polls are close – well into the range where even a small poll error could be crucial. According to our polling averages, the gap between Trump and Harris is 2 percentage points or less in all seven major swing states (Arizona, Georgia, Michigan, Nevada, North Carolina, Pennsylvania and Wisconsin). And in our average of national polls, Harris leads Trump by just 2.0 points. That’s closer than the gap between Trump and President Joe Biden on any day of the 2020 campaign and closer than the gap between Trump and former Secretary of State Hillary Clinton on any day between Oct. 1 and Election Day 2016. (Because of favoritism of Republicans by the Electoral College, our model currently estimates that Harris will likely need to win the statewide popular vote by 2.0 points to win a majority of the electoral vote.)

Recent high-profile polls have mostly just confirmed that the race is close. An ABC News/Ipsos poll from Oct. 4-8 showed Harris ahead by 2 points nationally. A poll conducted Oct. 1-10 by Marquette University Law School, one of the most accurate and transparent pollsters in America, showed her up 3 points. The list goes on: YouGov/CBS News recently gave Harris a 3-point lead; Marist College gave her a 5-point lead; Fairleigh Dickinson University with a 3 point lead. Fewer pollsters gave Trump a lead nationally, 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 increased it by 1.

In national polls, the race was not only close, but also stable. In fact, Harris’ current 2.0-point lead in national polls is pretty similar to what it was two months ago, on August 18th. The crosstabs have hardly changed either. Among black and Hispanic voters, for example, Harris’ support has risen or fallen by less than two points since she joined the race. Of course, these two points could be crucial, but it is remarkable how little has changed over the course of the campaign.

State polls, however, have shown comparatively more movement – and most of them against Harris. Compared to her numbers earlier in the month, Harris has lost ground in all seven key swing states. For this reason, their chances of winning have decreased in our forecast. In American presidential elections, you don’t get points for winning the national popular vote.

Do partisan pollsters distort our averages?

One question we sometimes get is whether polling averages like the 538 are trending in favor of Trump due to the influx of polls conducted by Republican-aligned companies. In the last 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 how accurate state polls are, the impact of partisan polls on our averages is not a major concern of mine. That’s because we work hard to remove potential statistical biases from each survey before incorporating them into our averages. As you can see in our poll averages methodology, we adjust partisan polls to account for the fact that these polls are usually a little too good for the sponsoring party.

And even for nonpartisan polls, we apply what we call a “house effects” adjustment, which takes into account how much more Democratic or Republican a pollster leans than their colleagues (whether due to the partisan leanings of their clients or simply due to methodological choices). typically produce more liberal or conservative samples). For example, if a pollster’s polls were consistently 2 points better for Trump than the poll average, after accounting for factors such as a poll’s population (likely voters versus registered voters or all adults) and mode (e.g. live telephone, online Panel, SMS, etc.), we adjust these polls by 2 points towards Harris.

Finally, we give less weight to polls from pollsters without a 538 rating and from pollsters that release a series of polls in a short period of time. This ensures that pollsters “flooding” the zone with surveys do not have undue influence on our averages.

One test of whether these adjustments are working correctly is to see what the 538 averages would look like if we did not include surveys from the companies that are supposedly trying to influence the averages. The table below shows the results of this comparison. The second column shows the averages of 538 as of October 18, 4:00 p.m. Eastern time. In the third column you can see what our averages would have been at this point if we had removed these polls, but otherwise ran the same algorithm we normally use:

As the table shows, this does not significantly change our averages. In most places, the pollsters in question are actually more pro-Trump than other pollsters. However, this has only a small impact on our averages, shifting them towards Trump by just 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 bipartisan average gives her a 0.9 point lead.) That’s not a significant difference in a world where The average polling error in the presidential election is 4.3 points, small enough that it could easily be attributed to sampling error or a methodological factor other than partisan bias. For comparison, our averages regularly move between 0.1 and 0.3 points daily, and we do not recommend that anyone interpret these changes.

Admittedly, that says nothing about the companies’ motivations (we like to stick to the data here at 538) or what goes on behind the scenes (we don’t really know). What we can say is that poll averages that are more comprehensive have tended to perform better in the long run than averages that use a more limited group of pollsters or an inconsistent methodology for weighting and making adjustments. One of the reasons we have all the rules and methods we use is to have confidence in our work even when we have doubts; This means sticking to our rules even when we may be tempted to make an exception.

Importantly, however, this does not mean we should read all polls without criticism or take a closer look at how pollsters generate their data. In this case, our empirical look at how our averages work gives us confidence that things are working as intended.

What if the election took place today?

In summary, the presidential race is close no matter how you look at it. This is both because the polls are currently extremely close and because there is still uncertainty about how the final two and a half weeks of the election campaign will unfold. As we have written, the race can change quickly in recent weeks. For example, in the final weeks of the 2016 and 2020 campaigns, polls trended toward Trump by about 4 and 2 points, respectively.

The added value of our forecast over our polling averages is that it attempts to quantify how likely such a shift is – and how inaccurate polls might be even on Election Day itself. But this also means that as the time left for people to change their minds decreases, the uncertainty of the race forecast decreases – and very quickly.

But because the polls are so tense right now, if they don’t change, the forecast won’t change. If you run our presidential election forecast in “nowcast” mode – where we pretend the election is today and remove any volatility from the number of days left until November 5 – Trump still has a 52 in 100 chance of winning .

Finally, I must remind you that a close election in terms of the odds does not necessarily mean that the final result will be close. If the polls fail to reach a historically normal result, any candidate with 300 Electoral College votes could easily win. 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 history as one of the closest in terms of vote totals.

Mary Radcliffe contributed to the research.

Footnotes

*All figures in this article as of October 18, 4 p.m. Eastern Time.

**Includes all surveys added to our database from October 5th to October 18th at 4:00 pm Eastern Time. Partisan polls are polls conducted by a pollster or sponsor that are identified as partisan in our database, including internal polls conducted for a campaign.