But 16 days before the national vote, Ms Clinton had a 95 per cent chance of winning according to a Reuters/Ipsos electoral-vote analysis, also released on Saturday. Another new poll had the Democrat soaring to a double-digit lead over Mr Trump.
Analysts have said the results of the IBD/TIPP poll are surprising. Statistics on how 11 national polls performed in the last three elections rank the survey top, but the likelihood of so many other polls all being wrong is extremely low.
The IBD/TIPP predictions suggested Mr Trump could secure 42.1 per cent of the vote, while Ms Clinton would only get 40.8 per cent.
The poll surveyed 783 likely voters. It reflected the rolling average of six days’ worth of polling, with a margin of error of 3.6 percentage points.
Respondents were asked how they would vote in a four way race, with Libertarian Party candidate Gary Johnson and Green Party nominee Jill Stein included. They got seven per cent and three per cent support respectively.
On Sunday morning Mr Trump tweeted a link to the poll results. Critics said he was "cherry-picking" polls to claim a lead he does not really have.
Less surprising were the demographic trends shown by the new poll.
According to IBD/TIPP data Mr Trump does better than Ms Clinton among men (48 per cent to 34 per cent), lower income voters (48 per cent to 35 per cent), rural voters (60 per cent to 22 per cent), and those who identify as religious.
Ms Clinton beats Mr Trump with women (46 per cent to 38 per cent), voters aged 18 to 44 (41 per cent to 34 per cent), the wealthy (47 per cent to 38 per cent), and investors (48 per cent to 38 per cent).
Ms Clinton has been described by US media as acting as if the presidential race is all but won, while Mr Trump himself admitted it was his “last chance”.
On Friday The New York Times described Ms Clinton as “working to ensure a victory that is decisive enough to earn a mandate for her presidency.”
The Washington Post reported: “Emboldened by polls predicting an electoral-college landslide in the presidential race, Clinton is shifting her strategy to lift up other Democrats coast to coast.”
Although the new data appears to suggest Ms Clinton has assumed victory prematurely, polling experts have said it is extremely unlikely the IBD/TIPP poll will prove to be correct. It is more likely polls showing Trump leading are outliers, either by statistical chance or by an error in their design.
Many things can go wrong in polling and even the stated margin of error only means 95 percent of the time the true value will be within that range. Five per cent of the time polls are expected to be way off.
Two other polls have shown Mr Trump leading in recent weeks: Rasmussen and USC Dornsife/Los Angeles Times.
The USC/LA Times poll was expected to produce different results to most because of its unique methodology, the Huffington Post reported.
Researchers asked: “What is the percent chance that if you were to vote, you will vote for Clinton, Trump, or someone else?”, instead of the usual “Who would you vote for?” question.
Meanwhile, Rasmussen polls often lean more Republican than other surveys. Some critics speculate this is intentional, but it could also result from the surveys landline-heavy automated polling methodology.
The IBD/TIPP poll has particularly surprised analysts because it is highly regarded and has previously performed extremely well.
However, experts have cautioned that even the best polls can suffer errors.
Join our new commenting forum
Join thought-provoking conversations, follow other Independent readers and see their replies