Another cataclysmic failure of opinion polls

“I have always believed in miracles,” said Australian Prime Minister Scott Morrision late Saturday evening after his government survived and won a clutch of seats from its opponents.

It was supposed to be the unlosable election for the Australian Labor Party (ALP). For three years the conservative LNP government, a coalition of the Liberal Party and the National Party (largely rural constituencies) appeared to be on the nose with voters. In 56 successive polls, not a single one in the entire term of government showed the LNP leading.  

It had been in office since 2013, when it was elected with Tony Abbott as leader. But Abbott was a “climate change denier” and opposed same-sex marriage. His nemesis, another Liberal, Malcolm Turnbull, pointed out that the LNP government had lost 30 opinion polls in a row. So in 2015, the party room dumped Abbott and elected the more progressive Turnbull as Prime Minister.

In the 2016 election, Turnbull scraped through with a narrow victory but soon began treading in Abbott’s poor poll footsteps. In April last year came the day when he, too, had lost 30 opinion polls in a row. Now desperate, Turnbull’s colleagues turned on him and elected Scott Morrison as leader. The polls continued to show that the ALP would win the upcoming 2019 election.

With the polls devastating and defeat looming, senior politicians began jumping ship, announcing that they would not stand at the next election. Turnbull resigned from Parliament, triggered a by-election and began undermining the new leadership. The leader of the National Party resigned after a sordid scandal. The Coalition was imploding. "To say the government is a shambles is insulting to shambles," was one of the better headlines. 

The polls continued to show that voters were going to ditch the LNP on May 18. Two days before election day, one of Australia’s leading bookmakers, Sportsbet, paid out A$1.3 million to punters who had backed a Labor victory. On the morning of election day, another agency was offering odds of $1.08 on Labor and $11.50 on the Coalition. On the evening of election day, TV pundits were still predicting a landslide to the ALP – 80 or 81 seats in the 151-seat Parliament.

The result? Votes are still being counted in a handful of close seats, but the Coalition appears to have won 78 seats to the ALP’s 67. (Six independents were also elected.) The polls – all of them -- could not have been more wrong. It was Trump and Brexit all over again, an apollcalypse. 

"I think this is really a cataclysmic era of polling in this country," said political scientist Andy Marks ruefully.

"It's like one poll can be three per cent out and that's what you would sort of expect now and then by random chance,” says electoral analyst Kevin Bonham. “But all the polls being out by that amount in the same direction and getting all the same results is something that absolutely cannot happen by random chance.".

This is more than a technical question for politics addicts. It affects all of us. The public is constantly being told to support issues because polls show that this is the majority view – whether the issue is euthanasia, climate change, abortion, or transgender rights. More than a political tool, polls have become a source of moral authority.

The number-crunchers themselves are deeply perplexed. Despite scorn on Twitter from American guru Nate Silver, Australian pollsters are quite sophisticated. The scandal is that the five major polling organisations, one and all, had epic fails.

What went wrong?

The professionals have a few answers.

Pollsters focus on educated professionals. “Australian pollsters ask for age and gender of those they survey, but not for education levels,” says Adrian Beaumont, a statistician at the University of Melbourne. “People with higher levels of education are probably more likely to respond to surveys than those with lower levels.” This means that the inner-city latte sippers and upper-income progressives will be over-represented.

Herding. Polling is a highly competitive industry. If pollsters have a result contradicting the consensus, they might adjust the figures for fear of being wrong. “Nobody wants to be the lone pollster with the completely wrong result while all the others are right,” observes Bonham.

Bad sampling. The industry recognises that this is a serious challenge. In the past, they rang random land-line numbers in the White Pages. Everyone uses mobile phones nowadays but polling methodology has still not caught up with changing demographics.

Language problems. One in six Australians speaks a language other than English at home. This may be leading to underrepresentation of foreign-born voters.

But the most thoughtful answer comes from British sociologist Frank Furedi in Spiked. He argues that in an era of political correctness, the pollsters were blinded by voter self-censorship.

In a world where language is systematically policed, where people are continually warned ‘You can’t say that!’, it is inevitable that many people choose to keep their opinions to themselves...

The reluctance to express one’s true beliefs is particularly widespread among people who feel that their views are stigmatised and scorned in wider society. People of faith are sometimes reluctant to express their moral concerns in public. In an age when conservatives are automatically branded as ‘right wing’, and sometimes casually dismissed as ‘far right’, many voters feel unsure about expressing conservative views to pollsters.

One of the explanations put forward by the professionals is the “shy Tory” – voters who are ashamed to voice their views. “But in truth,” Furedi contends, “the fault doesn’t lie with people who are reluctant to state their true opinions – it lies with the dominant culture of intolerance towards dissident opinion ... Challenging the culture of ‘You can’t say that!’ is the precondition for diminishing the pressure to conform and self-censor.”

For those sceptical that a “culture of intolerance” exists, the latest sighting came last week in The Guardian newspaper, which is published in the UK but has a widely-read online presence in Australia.

The Guardian has just modified its style guide – the way that language should be used in its publications.

So, instead of “climate change” Guardian journalists will write “climate emergency, crisis or breakdown” and “global heating” rather than “global warming”.

Those opposed to the scientific “consensus” are no longer to be called “climate sceptics” but “climate science deniers” or “climate deniers”. Why? Because a sceptic is “a seeker of the truth”. Climate deniers deny the truth “in the face of overwhelming scientific evidence”.

The word “denier”, of course, immediately evokes “Holocaust denier”. So The Guardian is not-so-subtly associating climate change sceptics with Nazis and the loopiest of loopy extremists. This comes from a newspaper which prides itself on its use of language.

So what would you tell a pollster from The Guardian who asks, “do you consider yourself a climate science denier”? You would probably tell a fib. Perhaps something similar is happening with polling, in Australia and elsewhere.

Michael Cook is editor of MercatorNet.


Join Mercator today for free and get our latest news and analysis

Buck internet censorship and get the news you may not get anywhere else, delivered right to your inbox. It's free and your info is safe with us, we will never share or sell your personal data.

Be the first to comment

Please check your e-mail for a link to activate your account.