#Open journalism No news is bad news

Your contributions will help us continue to deliver the stories that are important to you

Support The Journal
Dublin: 8°C Monday 25 October 2021
Advertisement

The pollsters got it VERY wrong in the US this week - how can they learn from their mistakes?

Only one polling organisation was consistently giving Trump the edge in advance of polling day. So what happened?

THE ELECTION OF Donald Trump to the US presidency stunned the world this week.

The majority of political journalists, commentators and pollsters all got it wrong.

Almost as soon as they’d scooped their jaws off the floor, America’s highest profile polling experts began to ask the question:

What happened?

More to the point … what happens next?

The scale of the polling blunder

Of the twenty major polling institutions, including national news networks, prominent newspapers and newswires that conducted more than eighty polls since mid-September, only one organisation – LA Times/USC Tracking – consistently gave Trump the edge.

On election day morning the RealClearPolitics rolling average showed Clinton ahead by 3.3% points nationally.

But after some positive early signs for Clinton, the scale of Trump’s triumph became clear later in the night on Tuesday – as he turned long-standing blue states red to romp past the 270 mark in the Electoral College needed to clinch the presidency.

The respected site FiveThirtyEight.com had forecast Clinton would win the key battleground states Florida, North Carolina, Pennsylvania and Wisconsin.

Trump won all four, and the election.

The polls, it quickly became apparent as election night wore on, had got most of their predictions spectacularly wrong.

What went wrong?

From as soon as it became apparent that Trump and Hillary Clinton would be their respective parties’ nominees earlier this year the former Secretary of State had been consistently in front – although she suffered a dent in her support after a number of campaign controversies, including the Halloween weekend announcement that the FBI was investigating more emails connected to her campaign.

Trump, in contrast, stumbled from controversy to controversy in the wake of the Republican convention.

Based on polling and their experience covering previous campaigns, journalists and commentators reckoned he had little chance of taking over from Barack Obama as the 45th US President.

So what happened?

“If anything, the situation in the United States further reinforces how ineffective established polling houses are,” Dr Kevin Cunningham, a lecturer in statistics and former strategist for the British Labour Party told TheJournal.ie via email.

They overestimated turnout among ethnic and younger groups that do not turnout, and based their estimates of turnout on the people that were willing to respond.

According to Cunningham, who also runs polling and data analysis company Ireland Thinks:
While the pre-poll results pointed to a significant swing towards Trump, the media focused entirely on notional differences in minority groups, which for statistical reasons will always show a higher variance.
The clearly more important Republican-Democrat gap had clearly narrowed sufficiently for Trump where the data was available in Florida and Ohio.

2016 Election Clinton Source: Matt Rourke

University of Virginia political professor Larry Sabato, like most forecasters, predicted a Clinton victory this week.

“The crystal ball has a big crack in it, my friend,” Sabato told AFP.

Many pollsters weight their samples based on the electorate as it was composed in prior election contests, according to Sabato.

That proved their undoing, because polls simply underestimated the number of quiet, poll-avoiding Trump supporters out there.

“White turnout in rural America was through the ever-loving roof,” he said, while African-American and millennial turnout was down.

Rural America

While pollsters anticipated a drop in the black and youth votes compared to 2012 when Barack Obama won, “their likely voter screens simply did not catch the high impending turnout in these white rural areas”.

The Clinton campaign’s own internal tracking polls also badly misread the white working-class vote, according to one election analyst who spoke to AFP.

“They were completely wrong - and they spent a fortune,” the analyst said.

Cunningham, the former British Labour strategist, agreed:

The divide is much clearer along the rural and urban divide and ethnicity.

Campaign 2016 Trump Source: John Minchillo

‘They were coming out of nowhere’

News outlets in the US and across the globe have been putting together ‘Why Trump Won’ pieces over the last few days.

This excerpt from a piece by Time Magazine’s Zeke Miller gives an insight into the phenomenon, and shows why turnout from previous non-voters proved decisive:

Chris Reilly, a commissioner in York County, Pennsylvania, has lived in the heavily Republican area north of Baltimore for 28 years. On the day in September after Mike Pence spoke to some 800 folks in downtown York, Reilly scanned a panoramic picture of the crowd in the local paper and had a shock. “I recognized one face,” he said. That’s when the party stalwart knew something was going on.
Then, on a recent Friday, Reilly got word that the county had received 9,000 absentee-ballot applications in a single day. It had to mail them out by Monday but had no money for extra help. So Reilly turned up at the election office on Saturday to stuff the applications into envelopes himself. As he did, he noticed something surprising. The applications were running 10 to 1 male. And when he peeked at the employment lines, he saw a pattern. “Dockworker. Forklift operator. Roofer,” Reilly recalled. “Grouter. Warehouse stocker. These people had probably never voted before. They were coming out of nowhere.”

‘Not an insane error’

Meanwhule, the folks at FiveThirtyEight.com have been speaking to pollsters from a range of companies to figure out what happened.

“We may be looking at a 4-point or so national miss – which [...] is not an insane level of error, but it is real error and the public’s right to question polls is justified,” Nick Gourevitch of Global Strategy Group said.

A number of pollsters rejected the suggestion that Trump supporters were too shy to tell interviewers who they were backing – although one said the GOP candidate did better when a recorded voice was asking the questions, rather than an in-person caller.

Women who supported the businessman may have been reluctant to tell pollsters, another expert said – while it’s likely some backers of Libertarian Gary Johnson, the third party candidate, switched their allegiance to the Republican at the last minute.

Johnson’s decline, in the days leading up to polling day, coincided with an improvement for Trump.

Source: Fox Business/YouTube

What happens now?

Pollsters, you may recall, also took some heat back in June, when Britain voted to leave the EU – although findings had been wavering back and forth in the days before the Brexit vote, so the upset wasn’t quite on the same scale.

But this week’s US election is just the latest in a string of missteps by established polling companies: pollsters were also showered with criticism in the wake of the Conservatives’ win in last year’s UK election, remember?

As to what happens next, experts reckon it will take months to properly analyse the data and figure out how to fix the problem.

“There’ll be one or more blue-ribbon panels to come up with something that works,” Sabato said.

But the University of Virginia political professor refused to throw in the polling towel.

“Analysis by anecdote is not academic,” he stressed.

You don’t want to just rely on instincts, you have to rely on data.

The professor pointed to a steady decline in willingness among the public to involve themselves in telephone polls. In the future, “most polling will be done online,” he added, dismissing concerns that internet surveys can be easily skewed.

They’re not unreliable – not if you do them well.

“We’ve reacted well to failure before,” Jon Cohen, the vice president of survey research at SurveyMonkey told Politico.

Polling is too important to go away. The way that we are going to understand what happened in the election, and the contours of where we sit as a country, is through polling.
- with reporting by AFP

Read: How the US president-to-be picked up a Clare golf resort for a song >

Read: Is it time the US Electoral College voting system was done away with? >

About the author:

Read next:

COMMENTS (75)