While Joe Biden is handily beating Donald Trump in national polls, two days out from election day, a new poll from Iowa on Saturday night showed the president up by seven points.
As Mr Trump enjoyed the same Iowa lead over Hillary Clinton days before he won the 2016 election with narrow victories in key midwest states, the news could well rattle Democrats anxiously awaiting Tuesday’s decision.
In the poll, conducted by Selzer & Company for the Des Moines Register and Mediacom between October 26th and 29th, Mr Trump led by 48 per cent to 41 per cent for Mr Biden. In September, the same poll showed the two men tied at 47 per cent.
Four years ago, Mr Trump won Iowa by 9.4 points. The Selzer poll involved 814 likely voters and had a 3.4 per cent margin of error. J Ann Selzer, the polling company’s president, said men and political independents continued to support Trump.
“The president is holding demographic groups that he won in Iowa four years ago, and that would give someone a certain level of comfort with their standing,” said Selzer. “There’s a consistent story in 2020 to what happened in 2016.”
He added that as “neither candidate hits 50 per cent, there’s still some play here”. But she also said data indicated 94 per cent of “likely voters” had decided which way to vote, including 98 per cent of Biden supporters and 95 per cent of Trump supporters. A mere 4 per cent of likely voters said they were still persuadable.
Among polling experts, reaction to the Iowa poll was mixed. Nate Silver, of FiveThirtyEight.com, wrote that the survey did not portend a wild swing to Trump across the board.
“One thing to keep in mind if you see late polling movement in a state is whether the movement is in line with fundamentals,” he wrote. “In Iowa, for instance, our model thought Trump ‘should’ have been ahead by three points based on polling in similar states, uniform swing, etc. It’s pretty red.
“So the Selzer poll, which brought our average there from Biden +0.1 to Trump +1.8, about as big a shift as you’ll see, brought the race more in line with the fundamentals there. The same would be true if, say, Biden got a couple of rough polls in Texas tomorrow.”
Nate Cohn, of the New York Times, noted that Selzer’s poll was “the president’s best poll result in a very long time – perhaps of the election cycle.
“It is also worth noting, though, that Selzer can be wrong, and has been before. No pollster has been put on a higher pedestal, but in the end everyone in this business is subject to sampling error and so on. If you expect perfection, you won’t get it.
“And this Selzer Iowa poll is off on its own, not just in Iowa but in terms of the overall story. Every national poll has shown Biden way ahead of Clinton among white voters [and] white working-class voters. He’s excelled across the white, northern tier.”
Both candidates are still competing for Iowa. Mr Biden held a drive-in rally in Des Moines on Friday and Mr Trump, who was in the state in October, was set to host a rally in Dubuque on Sunday.
Nationally, Biden was ahead of Trump 51 per cent to 41 per cent in the most recent Reuters/Ipsos poll. New polls on Sunday also put Biden up nationally and in key battleground states, though races were tightening in Pennsylvania and Florida, two of the most fiercely contested prizes.
Are Trump’s rallies spreading coronavirus?
Stanford University economists estimate that Mr Trump’s campaign rallies have resulted in 30,000 additional confirmed cases of Covid-19. Moreover, they also say the ralles likely led to more than 700 deaths overall, according to a paper posted online this weekend.
The research, led by B. Douglas Bernheim, chairman of economics at Stanford University, analysed data following 18 Trump rallies held between June 20th and September 22nd, three of which were indoors. Mr Bernheim said the work relies on statistical methods to infer causation after an event has occurred.
Infectious disease experts have long suspected that the president’s rallies ahead of the November 3rd election might be so-called superspreader events. But so far, scientists have not been able to get a good read on their impact, in part because of a lack of robust contact tracing in many states.
What is the concern?
In recent months, Mr Trump has held several dozen rallies in states such as Pennsylvania, Minnesota and Wisconsin, where coronavirus infection rates were already on the rise.
At each event, several thousand people were estimated to have participated. While most of the rallies were held outdoors, video footage show that participants gathered in close proximity and many were not wearing masks, creating a risk of spreading the virus as they cheered their candidate on.
It’s not a major stretch to say that large unmasked gatherings are likely to spread the virus, said Dr Amesh Adalja, an infectious disease expert at Johns Hopkins Center for Health Security.
Mr Adalja said the Stanford paper was “suggestive” of spread from the events, but not definitive because it was not based on an investigation of actual cases. That would help confirm whether participants were exposed to the virus at the event, rather than other places where transmission is rampant.
What do we know?
Minnesota public health officials have attributed four Covid-19 outbreaks and more than 25 cases to Trump rallies held in the state in September and October.
An additional 11 state health departments contacted by Reuters said they had not been able to trace infections to the rallies, although some, including Michigan and Wisconsin, have determined that individual people who later tested positive for Covid-19 were present at Trump campaign events.
What data are needed?
Disease experts say that rigorous contact tracing from one such large event could help arrive at an accurate prediction of how infectious such rallies can be.
But the United States has fallen behind other developed countries in this regard, due to a lack of funding and coordination for contact tracing by the Trump administration.
“The problem is we’ve not done anything to get real numbers,” said Dr Eric Topol, a genomics expert and director of the Scripps Research Translational Institute in La Jolla, California. Instead, it is subject to conjecture and mathematical models.
For example, scientists can use gene sequencing to trace minute changes in the genetic code of the virus as it passes from one person to another, allowing them to develop a map of where the virus travels. Such work has been used outside the United States, including in Australia and Hong Kong, to trace Covid-19 outbreaks.
“If we even had one rally where there was definitive tracing, then you could extrapolate. But we’ve had none. Our country has performed as if contact tracing doesn’t exist,” said Dr Topol – Guardian/Reuters