Supporters of U.S. President Donald Trump wave flags during the 2020 U.S. presidential election, outside Versailles Restaurant at Little Havana neighbourhood in Miami
Credit: Reuters
US Election Article Bar
As pollsters licked their wounds after the 2016 US election, they promised they wouldn’t make the same mistakes again.
Hillary Clinton, the former Democratic presidential candidate, had consistently led the Republican in state and national polls four years ago before Donald Trump pulled off a shock victory to win the White House.
According to the latest polling averages, Joe Biden’s lead over the incumbent has remained solid, despite a slight decline in the wake of the US presidential debates and the US president’s diagnosis with coronavirus.
Mr Biden’s polling average has remained above 50 per cent since October 4, and the Democratic nominee has consistently polled in the lead since the race began.
Opinion poll experts say there are good reasons to trust this year’s polls more than those of 2016.
Trump vs. Biden 2020 polling live
Yet there were early signs on Wednesday that the 2020 race would be far tighter than the polls had suggested.
"It’s happening," Jason Miller, a senior Trump campaign advisor, tweeted as results trickled in.
In Florida, opinion polls had shown Mr Biden edging even or ahead, offering the Democrat the tantalising prospect of knocking his opponent out at the start of counting.
And yet, with about 94% of the estimated total votes counted in the key state, Mr Trump was leading Mr Biden 51.2% to 47.7%.
Part of Mr Trump’s strength in Florida came from an improved performance relative to 2016 in the state’s counties with large Latino populations.
Mr Biden was leading in counties where Hispanics make up more than 20% of the population, but Mr Trump’s share of the vote in those counties was larger than it was in the 2016 election. More than four in 10 Hispanic voters in Florida said they cast a ballot for Mr Trump, according to an Edison Research exit poll.
Betting market odds flip
The odds also showed the race tightening considerably.
After shortening following the final head-to-head debate, Mr Trump’s US election odds had stagnated, with Mr Biden still the odds-on favourite. But as races were called, the odds began shortening rapidly.
"The swing is now complete. Trump now rated a strong favourite in the betting," Betfair Exchange tweeted.
1⃣6⃣ The swing is now complete. Trump now rated a strong favourite in the betting. pic.twitter.com/yB0y7UIv80
— Betfair Exchange (@BetfairExchange) November 4, 2020
UK-based Smarkets Exchange was giving Mr Trump 65 per cent win-odds, up from 39 per cent when polls opened. Mr Biden’s win-odds have fallen to 35 per cent from 61 per cent earlier.
"This has mainly been driven by Trump seemingly holding the key swing state of Florida as a result of a very strong showing for him in the Miami-Dade county with a large Cuban population," said Patrick Flynn, political analyst at Smarkets.
However, early results in several key battleground states were in flux as election officials processed a historically large number of mail-in votes.
Democrats typically outperform Republicans in mail voting, while the GOP looks to make up ground in Election Day turnout. That means the early margins between the candidates could be influenced by which type of votes — early or Election Day — were being reported by the states.
Why pollsters got it wrong in 2016
Many pollsters tweaked their methods to better account for a demographic group that tends to avoid polls but largely favours Mr Trump: whites without a college degree.
It is not entirely clear why non-college whites are less likely than others to participate in polls. But when evaluating their survey samples, pollsters routinely saw that those Americans were not well represented.
That was a problem in political polls, since non-college whites tend to favour Republicans. They also make up a substantial proportion of the US electorate: Pew Research estimated that 44% of voters in 2016 were whites without a college degree.
Trump voter base 2016 v Clinton
As a fix, many polls, including Reuters/Ipsos, installed an education “weight” to calibrate their survey data so that it better reflects the education levels of all Americans.
Another factor four years ago was the indecision of voters.
A lot of Americans had trouble making a decision in 2016. Neither Mr Trump nor Mrs Clinton was popular, and opinion polls routinely showed that 20 per cent of likely voters remained undecided until the final weeks of the campaign.
That added a considerable amount of uncertainty to an already volatile year in American politics. Mrs Clinton’s once-sizable lead in October nearly vanished just before Election Day in 2016 as more and more voters made up their minds.
This year, indecisive voters were much less of a concern. The latest Reuters/Ipsos poll showed that less than 7% of likely voters have yet to back a major-party candidate, less than half of what it was four years ago.
That prompted political analysts to feel more confident about this year’s polls because the public appeared to be much more certain about their vote.
Свежие комментарии