On Jan. 28, 2025, 538 launched our average of polls of President Donald Trump’s approval rating. For Trump’s second term, we are debuting a brand-spanking-new methodology for our presidential approval tracker, building upon refinements we made throughout former President Joe Biden’s term and during the 2024 presidential election. (For now, though, we are not changing the methodology of our existing averages, such as our other approval averages, our favorability averages and our election averages.) Read on for a full description of how we’re calculating Trump’s approval rating.
Which polls we include
538’s philosophy is to collect as many polls as possible for every topic or race we’re actively tracking — so long as they are publicly available and meet the basic criteria for inclusion listed in our polls policy. After determining that a poll meets our standards, we have to answer a few more questions about it before sending it off to the various computer programs that power our models.
- Which version should we use? Whenever a pollster releases multiple versions of a survey — say, an estimate of Trump’s approval rating among both all adults and just registered voters — we choose the survey that best matches either the breakdown of polls in our historical database or the preferred target population for that type of poll. For presidential approval ratings, we prefer polls of all adults to polls of registered voters and polls of registered voters to polls of likely voters.
- Is this a tracking poll? Some pollsters release daily results of surveys that may overlap with each other. We account for this potential overlap in these “tracking” polls by running through our database every day and dynamically removing polls that have field dates that overlap with each other until none are overlapping and we have retained the greatest number of polls possible for that series and firm, paying special attention to include the most recent poll.
- Is it an especially large survey? When polls are fed into the model, we decrease the effective sample sizes of large surveys. Leaving these large numbers as they are would give those polls too much weight in our average. As a default, we cap sample sizes at 5,000. Then, we use a method called winsorizing to limit extreme values.
- Do we know the sample size? Some pollsters do not report sample sizes with their surveys. While we can usually obtain this number for recent surveys by calling up the firm, sometimes we have to make informed guesses. First we assume that a missing sample size is equal to the median sample size of other polls from that same pollster on the same topic (e.g., presidential approval). If there are no other polls conducted by that firm in our database, we use the median sample size of all other polls for that poll type.
How we weight polls
After all this data is in our database, we compute three weights for each survey that control how much influence it has in our average, based on the following factors:
- Sample size. We weight polls using a function that involves the square root of its sample size. We want to account for the fact that additional interviews have diminishing returns after a certain point. The statistical formula for a poll’s margin of error — a number that pollsters (usually) release that tells us how much their poll could be off due to random sampling error alone — uses a square-root function, so our weighting does, too.
- Multiple polls in a short window. We want to avoid a situation where a single pollster floods an average with its data, overwhelming the signal from other pollsters. To do that, we decrease the weight of individual surveys from pollsters that release multiple polls in a short time period. If a pollster releases multiple polls within a 14-day window, those polls each receive a reduced weight equal to the square root of 1 over the number of polls inside that window. That means if a pollster releases two polls in two weeks, each would receive a weight of 0.71 (the square root of 0.5). If it releases three polls, each would receive a weight of 0.57 (the square root of 0.33).
- Pollster rating. Starting in 2025, we now weight polls that go into our presidential approval averages by their firm’s 538 pollster rating — a score we give pollsters based on their historical accuracy in predicting election outcomes and their current methodological transparency. To calculate the weight we give to a poll based on its pollster’s 538 rating, we first divide its rating by 3 (the maximum possible score) and take the square root of the resulting number. For example, a pollster that we give 0.5 stars out of 3 gets 41 percent of the weight that a 3-star poll gets.
These three weights are all multiplied together in our final model.
How we average polls together
Once we have all our polls and weights, it is time to average them together. But which methodology for aggregation should we choose? Broadly speaking, the most commonly used polling averages for U.S. public opinion have followed one of three approaches:
- Take a simple moving average of polls released over some number of previous days.
- Calculate a trendline through the polls using various statistical techniques, such as a polynomial trendline or Kalman filter.
- Combine these approaches, putting a certain amount of weight on the moving average and the rest on the fancier trend.
There are a lot of benefits to this third option, and it has historically been the solution used by 538. The average-of-averages approach allows you to use predictions from the best parts of a slow-moving exponentially weighted moving average and a fast-moving polynomial trendline; it is computationally efficient to do so; and it’s easy to explain this model to the public. It’s also relatively trivial to tweak the model if we find something is working incorrectly.
However, this model has some shortcomings too. Our poll-averaging model for favorability ratings, primary elections and the generic congressional ballot is really a set of many different models that work together iteratively: First, we use models to reduce the weight on polls that are very old or have small sample sizes; then we use models to average polls and detect outliers; then we run new averaging models to detect house effects; and so on and so on, for nearly a dozen individual steps.
If a modeler isn’t careful, this can introduce some problems — some of them practical and others statistical. First, it’s hard to account for uncertainty in the average, especially when using ad hoc weights for sample size and other factors. That’s because we potentially generate statistical error every time we move from one model to the next, and we have to run the program thousands of times every time we want to update! It’s also a little more sensitive to noise than we’d like it to be, even when designed to accurately predict support in future polls given everything that came before.
So this year, we are introducing a new type of statistical model for averaging polls of presidential approval. It is similar to the one we used for polls of the 2024 general election, and derivations of it have been used to model approval ratings of government leaders in the United Kingdom and for elections in Australia. Oversimplifying a bit, you can think of our updated presidential approval polling average as one giant model that is trying to predict the results of polls we collect based on (1) the overall state of public opinion on any given day and (2) various factors that could have influenced the result of a particular poll. These are:
- The polling firm responsible for the poll. The specific ways in which a pollster collects and analyzes its data can lead to systematic differences between its polls and other pollsters’ polls. For example, some pollsters, such as the Trafalgar Group, usually underestimate support for Democrats, while other pollsters, like Center Street PAC, overestimate them. We call these differences “house effects,” and we apply a house-effect adjustment to ensure they’re not biasing our averages. House effects in approval ratings are usually less consequential than in election polling, but they can still be large, and it’s important a model takes them into account.
- The mode used to conduct the poll. Different groups of people answer polls in different ways. Young people are likelier to be online, for instance, and phone polls reach older, whiter voters more readily. If pollsters aren’t careful, these biases can creep into all polls conducted using a single mode. So we apply a mode-effects adjustment to correct for those biases before we aggregate those surveys.
- Whether the poll sampled likely voters, registered voters or all adults. Historically, pollsters have used all adults as their survey population for polls of the president’s approval rating. Therefore, we have our model look for any systematic differences between all-adult, registered-voter and likely-voter polls and subtract those differences from the latter two types of poll. This gives us an average that is more closely comparable to historical trends.
- Whether the poll was conducted by a campaign or other partisan organization. We apply a partisanship adjustment to account for this. For example, if a poll sponsored by a Republican-friendly organization gives a result that is more friendly to Trump than another poll, we subtract some of the difference.
Finally, our prediction for a given poll also accounts for the value of the polling average on the day it was conducted. That’s because if overall approval for a president is 50 percent, we should expect polls from that day to reveal higher support than if the president were at, say, 30 percent overall approval. This also means the model implicitly puts less weight on polls that look like huge outliers, after adjusting for all the factors above.
That brings up the question of how exactly the average is being calculated.
We use a random walk to model averages over time. In essence, we tell our computers that support for the president in national polls should start at some point on Day 1 and move by some amount on average each subsequent day. Support for the president might move by 0.1 points on Day 2, -0.2 points on Day 3, 0.4 points on Day 4, 0 points on Day 5, and so on and so on. Every time we run our model, it determines the likeliest values of these daily changes in presidential approval while adjusting polls for all the factors mentioned above. We can extract those daily values and add them to the starting value for the president’s approval at the beginning of his term: That gives us his approval rating.
(Actually, we run three different versions of our average, to account for the chance that daily volatility in public opinion changes over time. For every day of a president’s term, we calculate one estimate of his approval rating by running our model with all polls from the last 365 days; one with all polls from the last 90 days; and one with all polls from the last 30 days. Then, we take the final values for those three models and average them together. This helps our aggregate react to quick changes in opinion while removing noise from periods of stability.)
Finally, we account for any additional detectable error in a poll. This is noise that goes above and beyond the patterns of bias we can account for with the adjustments listed above. The primary source of this noise is sampling error, derived from the number of interviews a pollster does: A larger sample size means less variance due to “stochasticity” (random weirdness) in a poll’s sample.
But there is also non-sampling error in each poll — a blanket term encompassing any additional noise that could be a result of faulty weights, certain groups not picking up the phone, a bad questionnaire design or really anything else that we haven’t added explicit adjustment factors for. Our model decides how much non-sampling error is present across the polls by adding an additional constant to the standard deviation implied by each poll’s sample size via the sum of squares formula (with the model deciding how large that constant should be).
How we measure uncertainty
Those familiar with our previous presidential approval averages will remember that they include error bands — shaded areas on our graphs that are meant to represent our uncertainty about the state of public opinion. Traditionally, this shaded interval has been calculated to show the range in which 95 percent of all polls are supposed to fall. For our average of Trump’s second-term approval rating, we have adjusted this 95 percent interval to show the uncertainty we have about the average itself. You will notice the new interval is much smaller; that’s because uncertainty about the average is much smaller than uncertainty about individual polls.
Our new model also lets us account for uncertainty in the polling average in a very straightforward way. Imagine we are not calculating support for a president one single time, but thousands of times, where each time we see what his support would be if the data and parameters of our model had different values. For example, what if the latest poll from The New York Times/Siena College had Trump’s approval rating 3 points higher, as a poll of 1,000 people would be expected to have about 5 percent of the time? Or what if the population adjustment were smaller?
Our model answers these questions by simulating thousands of different polling averages each time it runs. That, in turn, lets us show uncertainty intervals directly on the average.
Conclusion
And that’s it! 538’s average of Trump’s approval rating will update regularly as new polls are released. If you spot a poll that we haven’t entered after a business day, or have other questions about our methodology, feel free to drop us a line.
Version history
2.0 | Debuted new presidential approval average methodology for the Trump administration. | Jan. 28, 2025
1.3 | Added downballot primary polling averages and clarified that presidential general election averages have a different methodology. | April 25, 2024
1.2 | Added adjustment for partisan polls, updated population and sample-size adjustments. | Nov. 3, 2023
1.1 | Improved outlier detection, more stable averages, better optimization. | Oct. 9, 2023
1.0 | New favorability, approval and horse-race averages debuted. | June 28, 2023
References
Trump approval rating tracker
