Ask most party strategists, pollsters or American pundits and they will tell you that the 2024 presidential election could be the closest in decades, if not a century.
Given that Joe Biden’s 2020 victory was decided by fewer than 45,000 votes in just three battleground states, that’s quite a statement.
Yet this is what the polls suggest. The competition between Kamala Harris And Donald Trump is extremely tight, both nationally and in key states.
So how much can we trust polls?
It’s a question many are raising after getting it wrong four years ago.
The American Association for Public Opinion Research called it the biggest polling failure in 40 years, showing Mr. Biden’s lead over Mr. Trump in the final two weeks of the campaign twice as large greater than it was when the votes were counted.
In 2012, pollsters largely underestimated Barack Obama’s lead over Mitt Romney. But while Mr. Trump’s victory over Hillary Clinton in 2016 shocked many people, the error in the national polls was relatively small in comparison.
Of course, all polls have some uncertainty, some inherent error, and they are only a snapshot of the current situation, not a prediction of the outcome on Election Day.
But more importantly, there is evidence that they also become a more reliable indicator of the outcome as we get closer to the election. This may seem obvious, but one year before the election, the polls differ on average by seven points from the final tally of each candidate. During the last week of the campaign, this figure fell to less than three points.
Every day, the polls tell us a little more about the result.
So how should we read them?
A key takeaway is the uncertainty surrounding estimated support for candidates.
Uncertainty: “margin of error”
Although pollsters publish a percentage for each, they also report a “margin of error” to indicate the amount of support which can vary.
For example, if a poll gives Mr. Trump a score of 46% with a margin of error of three percentage points, that means his support among voters should be between 43% and 49%. If the same poll puts Ms. Harris at 49%, then her support should be between 46% and 52%.
All that tells us is that the competition is close and either candidate could come out on top.
Similar caution is required when examining trackers using survey averages.
One might think that averaging the surveys would reduce uncertainty, since random errors should cancel out. But some pollsters are systematically more accurate than others, while others may be systematically wrong in one direction. Adding them all together can reinforce these biases.
Learn more about the elections:
What exactly happens on US election night?
How does Donald Trump manage to maintain such strong support?
Trump in Wisconsin earlier this month. Photo: AP
Predicting participation: context matters
Who votes in an election is also crucial to the outcome, but predicting that is a difficult task for pollsters. About a third of eligible Americans don’t vote in presidential elections, and they’re not always the same people.
Context matters. This can make people more or less inclined to vote. For example, potential changes to abortion laws appear to have mobilized many Democrats. in the 2022 midterm elections.
A candidate’s policies and performance can also change the likelihood that more partisan voters will participate.
The Electoral College: Why state contests can be crucial
The same considerations are necessary when examining state polls and are arguably more crucial in determining which candidate is more likely to win the election.
The outcome of a presidential election is decided state by state, by the Electoral College, and the difference between it and the national vote has widened.
Learn more: What is the Electoral College?
Harris discusses abortion rights in Arizona in June. Photo: Reuters/Rebecca Noble
In both 2000 and 2016, the candidate with the most votes nationally lost the election because he or she did not win a combination of states with the highest Electoral College scores.
That’s why polls in battleground states get so much attention and are factored into statisticians’ models that try to predict the outcome.
Of course, they carry the same uncertainty and potential flaws as national polls. And the bad news is that their recent performances haven’t been great.
In 2016, they suggested that Mrs. Clinton would sweep key states needed for a comfortable victory in the Electoral College. In 2020, their plight is even worse.
Some states have proven particularly difficult to survey accurately. In 2016 and 2020, the biggest failures were in Wisconsin, Michigan, North Carolina and Pennsylvania.
These, particularly Pennsylvania, which of the group wins the most votes in the Electoral College, are all potentially decisive in 2024.
Keep up to date with all the latest news from the UK and around the world by following Sky News
So what can polls tell us?
The answer is plentiful if you want to know what people think of a certain candidate or policy.
But if you’re trying to figure out who’s ahead in the presidential race, then all you can say for sure is that the race is very close and could come down to a small number of votes in a few states.
