Voting in Australia’s postal plebiscite on same-sex marriage closes on November 7; the results will be released on November 15.

The consensus in the media and amongst politicians is that a majority will vote Yes. According to a Newspoll report published on November 1, among those who have already posted their ballots, support has risen from 59 to 62 percent.

According to The Australian, “While the No campaign made a strong start in the early weeks of the survey, it has slipped to 35 per cent support among those who have already voted, down from 38 per cent two weeks ago.”

But there are many odd features of the plebiscite which could derail the great expectations of same-sex marriage campaigners.

It’s not compulsory, unlike Australian elections, so opinion polls may not translate into votes. People vote by posting an envelope instead of slipping a paper into a ballot box as they normally do. Older people, who generally support the No vote, are used to posting letters; younger voters, who generally support the Yes vote, may never have posted a letter in their lives.

The pressure to vote Yes has been so intense in the media and workplaces that some voters may lodge a quiet protest by backing the No side.

Still, the consensus has been that Yes will win comfortably.

However, a study published yesterday in The Conversation upsets the apple cart. It predicts that supporters need to prepare themselves for a narrow loss.

Using data analytics, David Tuffley and Bela Stantic, academics at Griffith University, in Queensland, studied social media posts about the same-sex marriage campaign. Their conclusion? “Our detailed study of the opinions expressed on Twitter shows the result could be a narrow defeat of the Yes campaign, with 49.17% support.”

Since their method also predicted Trump’s upset victory in last November’s presidential election, it’s worthwhile heeding. Here's how they crunched the numbers.

First they examined 458,565 Australian tweets which mentioned same-sex marriage in October, using software which allowed them to identify the person’s opinion and the strength of that opinion, as well as their gender, age and education.

A initial analysis of the tweets showed overwhelming support for the Yes vote, with 72 percent in favour. However, a closer look showed that some passionate Yes-men had sent as many as a thousand tweets. In real life, though, what matters is not the volume of tweets but how unique users plan to vote.

Although there were 458,565 Tweets, only half of these represented unique users — 207,287. Amongst these, support for Yes ran at 57 percent.

And drilling down further, only 15 percent of the tweets came from over-55s, even though they represent 36 percent of the population. And over-55s overwhelmingly support the No case. Taking this into account, Tuffley and Stantic estimate that the Yes vote will dip as low as 49 percent.

Why the big difference from the opinion polls? Tuffley and Stantic believe that people are afraid to say what they really think, even to pollsters.

One of the problems with predicting poll outcomes is that people are often reluctant to say out loud what they really think about issues. What people say online can often be more accurate than what they say to each other in this age of political correctness.

In the lead-up to the recent US presidential election, the polls pointed to a Hillary Clinton win because many people were publicly saying “No” to Trump when asked by pollsters. But in the privacy of the booth, people quietly voted according to what they actually thought.

BettingSite.com.au  is offering odds of $1.60 for a Yes vote and $2.20 for a No vote. William Hill is offering $1.12 for a Yes vote and $5.50 for a No vote. The bookies often know best and they are still predicting victory for the Yes vote. But an upset win by the No vote is still credible.

We’ll just have to wait until November 15.

Michael Cook is editor of MercatorNet.  

Michael Cook

Michael Cook is the editor of MercatorNet.