Polls predicted Labor landslide but Scott Morrison performed a ‘miracle’

An election night panel has slammed the polls that predicted — for two entire years — a Labor victory. How did they get it so wrong?

For two whole years, Labor outperformed the Coalition in the only real tangible measurement the public had to rely on.

The polls were emphatically skewed towards a future Bill Shorten government.

They were the catalyst for the overthrow of former Prime Minister Malcolm Turnbull.

Forty-three consecutive Newspoll results put Labor ahead of the Coalition on a two-party preferred basis.

But none of it mattered yesterday when the nation returned Scott Morrison to the top job.

ABC political editor Andrew Probyn called the discrepancy between polls and actual votes “a shambles” on the national broadcaster’s election night coverage.

“When it comes to the opinion polling, something’s obviously gone really crook with the sampling both internally and externally,” he said.

Labor Senator Penny Wong said simply: “I think we all agree the polls are not reflective in what we’ve seen.”

What a remarkable outcome if this Libs pull this off. They have not won a single Newspoll since the 2016 election #ausvotes pic.twitter.com/gqXZ06Ql4h

— Luke Henriques-Gomes (@lukehgomes) May 18, 2019

Columnist Nikki Savva from The Australian — publisher of Newspoll — tackled the issue this morning on Insiders. She said the polling externally was not in line with polls conducted by Labor of the Coalition.

“The published polls were certainly very different from the internal polling the Liberal Party were doing … it was picking up big swings in Queensland,” she said.

ABC journalist Patricia Karvelas agreed. “They could see what was happening. It just wasn’t being reflected in the public polls,” she told the program.

“The Labor Party polling showed them behind in every key seat in Queensland. Strategists there were very worried about their polling. In terms of gaining seats, they weren’t confident.”

The same message was shared by The New Daily journalist Samantha Maiden, who wrote that Labor refused to buy into external poll results, but still never saw the bloodshed coming.

“Labor’s own polling was always far worse than Newspoll and never got above the 76 seats they needed to secure a majority,” she said.

“Mr Shorten knew all along. But even Labor was shocked by the result on the night.”

The result left the public frustrated. Many wondered why they should trust polls going forward when they were proven irrelevant.

How can every published poll, from every polling company, for months, including exit polls taken today, not just be wrong but be so spectacularly wrong? It makes no sense. #ausvotes

— Paul Kidd (@paulkidd) May 18, 2019

Why does the actual election produce such a different outcome than polling indicates? Is there some flaw in the polls, or is something happening on Election Day? Which one is the ‘wrong’ answer?

— Simon (@simonblight) May 18, 2019

Oh wow what's this? I'm sitting at home watching ANOTHER election where all the polling was wrong.

Sick of it pic.twitter.com/7dueeVhfBj

— Thomas Kelsall (@Thomas_Kelsall) May 18, 2019

US election, polls suggested big Hillary win, Donald won. Australian election (just now) polls suggested large Shorten (Labour) win, Scott has just won. What has gone wrong with the polling businesses?

— Eric Douglas Gribble (@Eric3Gribble) May 18, 2019

Is polling broken in the 21st century? Few predicted trump winning us election. Less thought brexit would happen. Now this. Seems strange it keeps getting it so wrong. Is the methodology no longer relevant?

— Ed Jackson (@edjacko) May 18, 2019

How were the polls so wrong? Who were they polling? #auspol #ausvotes

— 🐲 Robyn (@rzester) May 18, 2019

Newspoll, which is administered by Galaxy Research, is one of a number of polling companies relied on in Australia. Others include Roy Morgan, Nielsen (based in the US), Ipsos and Essential Poll.

Polling across the board predicted a primary vote of 39 per cent for the Coalition, but as of Saturday night the Coalition was at 41 per cent.

Peter Lewis from Essential Research said part of the problem was with the way polls were carried out — with robo-calls to fixed phone lines.

He told The Guardian it might be time to move away from our reliance on external polling.

“We treat polls with more import than they may deserve given the margin of error,” he said. “Is it that the polls were wrong or is it the way we read polls that is flawed? We obviously need to go back to model and see how it held up.”

It’s not the first time polls have been so far off the mark. An analysis of polling ahead of the US election that saw Donald Trump voted into the White House showed it was not even close.

Sam Wang, a Princeton University data scientist, wrote on the night Mr Trump secured victory that the “entire industry” needed a close inspection.

“The entire polling industry — public, campaign-associated, aggregators — ended up with data that missed tonight’s results by a large margin,” he wrote.

“There is now the question of understanding how a mature industry could have gone so wrong.”

Continue the conversation on Twitter: @ro_smith | [e-mail]