Variations in the way polls account for undecided voters leave some hope for the ‘Yes’ campaign as it battles sliding support a month out from the Voice referendum. By Mike Seccombe.

What the ‘Yes’ polls really mean

Prime Minister Anthony Albanese has ochre smeared on his forehead by a man in a blue blazer at Parliament House.
Prime Minister Anthony Albanese has ochre smeared on his forehead at Parliament House on Thursday at the completion of Indigenous leader Michael Long’s 20-day walk for the “Yes” campaign.
Credit: AAP Image / Mick Tsikas

The presentation by Labor national secretary Paul Erickson to Tuesday’s caucus meeting had something of the air of a sports coach geeing up his team at half-time.

Sure, they were behind: the public polls showed a big lead for those advocating a “No” vote in next month’s referendum on the proposal for a constitutionally enshrined Indigenous Voice to Parliament.

Yet the deficit was not insurmountable, Erickson said, citing extensive research by the Yes23 campaign, as well as the party’s own polling, which showed 30 per cent of people, some five million voters, had not yet made up their minds. About 20 per cent were “soft” voters for one side or the other, and 10 per cent had no idea which way they would vote.

So, said Erickson, “Yes” could still win. Rather than being distracted by speculative or predictive commentary about what the result might be, Labor should resolve, once parliament rose on Thursday, to get out on the field and get the job done.

It’s very hard to ignore the negative commentary, however, particularly when it comes from former team members. Kos Samaras, director of strategy and analytics with the research and advocacy firm RedBridge Group, and a former deputy campaign director for Labor in Victoria, agrees there are millions of soft and undecided votes. He just fears that, given the momentum of the contest and the way his team is playing, the “Yes” campaign will not be able to win them.

RedBridge’s latest poll, released last Saturday, was the worst yet in a succession of deteriorating results, showing the “No” vote ahead by 22 points, 61 to 39.

Still, it is mathematically possible for the Voice to pass – if the data Erickson presented to caucus is correct. In fact, Samaras suggests the number of voters who have yet to irrevocably decide is even larger.

In its poll, RedBridge found 41 per cent of respondents were “very certain” they would vote “No”, compared with 22 per cent in the corresponding category for “Yes”, which leaves 37 per cent not locked in to either side – of these, 14 per cent said they were “somewhat certain” to oppose, and 15 per cent somewhat certain to support, leaving only 8 per cent not at all sure, one way or the other.

“So the rusted-on ‘No’ vote is almost double the rusted-on ‘Yes’,” says Samaras.

For the supporters to win, they need to convince more than three-quarters of undecided voters, and many of those – the “almost certains” – are going to be very hard to persuade. As for the genuinely undecided, who tend to be low-information voters, they are likely to oppose what they don’t understand.

“The problem is that as they tune in, they jump on the ‘No’ pile,” Samaras says.

Not all the polls are as bad for the “Yes” side as the RedBridge Group one, although they are still pretty bad. The latest Resolve political monitor survey, published in the Nine media this week, had “No” ahead 57-43. Last week’s Newspoll in the Murdoch media scored it 53-38. The Essential poll in Guardian Australia has it 53-47.

These are big variations, particularly if you compare them with the results we see in an election. At this stage of the process – about a month from polling day – the major polls tend to show very similar numbers, within a couple of points.

There is a reason for this, Professor Simon Jackman of the University of Sydney says: a referendum is not like a general election, where more people have firm views.

“They know how they’ll vote 10 years out, in some cases,” he says. “This, though, for a lot of people, is something that they’re being asked to weigh in on that they probably haven’t given a lot of thought to. The interesting thing from a political campaigning perspective is the number of soft views. People say yes, but they’re still persuadable, or no, and are still persuadable.”

The big differences between the various polls, says Jackman, come down in large part to how their differing methodologies account for voter uncertainty. Newspoll, for example, presses those who identify as uncommitted with a follow-up question: “Voting at this referendum will be compulsory. While you may change your mind, if you had to choose now, do you approve of this proposed alteration of the Australian Constitution to recognise the First Peoples of Australia by establishing an Aboriginal and Torres Strait Islander Voice?”

Others, such as RedBridge Group, record gradations of voter opinion. Others still accept soft yes and no answers and don’t knows at face value. The result is wildly differing estimates of the number of people who are uncommitted. Newspoll says 9 per cent; Erickson says 30.

Jackman has built a model that attempts to average out the poll variations, weighting them according to their margins of error and other factors.

“The basic idea is that more information is better than less,” he says.

He has been working with the ABC to provide a running trend in the polls. The results are not good for the “Yes” campaign. Last October, two-thirds of people planned to vote in support of the Voice. The average of polls now is 56.3 opposed and 43.7 in favour. Like Erickson, however, Jackman still thinks “around 20 to 30 per cent” are uncommitted.

“Ascertaining who is genuinely undecided is hard, even for us, and we’re doing qualitative research,” says Rebecca Huntley, director of research at 89 Degrees East.

Her experience across many 90-minute focus group sessions suggests there is a difference between a soft yes and a soft no.

“Soft yes is just that: soft. Soft no is more likely a firm no,” she says.

Her worry – and she has been working for the “Yes” campaign – is that those who claim to be “soft no” voters often hold harder views than they admit.

“People are terrified of being described as racist,” she says, adding that “this is distorting what happens when you actually ask people where they sit”.

The other thing that stands out from her focus group work is that most people accept there is a real problem of Indigenous disadvantage that should be addressed. When the facts of the Voice proposal are explained, she says, those who are genuinely uncommitted are apt to gravitate to the “Yes” side.

“I’ve seen people in groups go from being pretty resistant to the idea to being supportive…” she says.

The problem, Huntley notes, is that many people are not across the facts. She acknowledges the “Yes” campaign faces a tough uphill battle but will not say it has lost.

In truth, while the published polls look grim for “Yes”, many imponderables remain when it comes to accurately measuring public opinion, relating to the fact the referendum must record a double majority to succeed: a majority of votes nationally, and in at least four of the six states.

Getting a fix on the national vote is easier than working out which way the various states will fall, because it is hard to get a representative sample in smaller states. The consensus among most pollsters and pundits is that Queensland and Western Australia are lost causes for “Yes”. New South Wales and Victoria could still be in play, but are looking less likely over time. The polling samples from the two smallest states, South Australia and Tasmania, are unreliably small.

“My political scientist self,” says Jackman, “tells me it’s probably not going [to] matter what happens in Tassie and South Australia, given where the national polls are, but if this got back to being closer – and we’ve still got about a month to go – it would be very, very helpful if our friends commissioning media polls put more effort there.”

Proponents of the Voice argue the slide in the polls is due to disinformation spread by the other side. Some also privately acknowledge their side has not run a good campaign. There have been arguments over the sharing of research data, making it harder to identify the best lines of argument to use. The “No” advocates have been sharper in their messaging and better targeted.

Andrea Carson, a professor of political communication at La Trobe University, notes the scattergun messaging of the “Yes” camp. Her analysis found 33 different “Yes” messages running in paid media.

The “No” campaign had about six different messages – but really only two, because the others weren’t in high use or rotation. They mainly used the lines that the Voice was divisive and that not all Indigenous people want it.

“The other thing is, it’s a negative message. And we know, from years of political science literature, that negativity works,” she says. “As human beings, we respond to messages of fear.”

Her monitoring of social media showed the “No” side, aware that younger people were more likely to vote “Yes”, had been far more effective in its use of sites young people employ, TikTok and Instagram. “They have recognised that if you want things to go viral, you don’t just cut and paste news clips, or give sermons that young people can’t relate to. They use humour really well.”

With a month to the official polling day, and only a couple of weeks before early voting begins, Carson says, the prospects of turning support around seem remote.

And should the referendum fail, says Huntley, no one will really win. People, she thinks, will be angry with both sides for their having subjected us all to an ugly campaign that achieved nothing.

But worse than the anger, there will be “a kind of sadness and a sense of, oh, we actually can’t do hard things”. 

This article was first published in the print edition of The Saturday Paper on September 16, 2023 as "What the ‘Yes’ polls really mean".

For almost a decade, The Saturday Paper has published Australia’s leading writers and thinkers. We have pursued stories that are ignored elsewhere, covering them with sensitivity and depth. We have done this on refugee policy, on government integrity, on robo-debt, on aged care, on climate change, on the pandemic.

All our journalism is fiercely independent. It relies on the support of readers. By subscribing to The Saturday Paper, you are ensuring that we can continue to produce essential, issue-defining coverage, to dig out stories that take time, to doggedly hold to account politicians and the political class.

There are very few titles that have the freedom and the space to produce journalism like this. In a country with a concentration of media ownership unlike anything else in the world, it is vitally important. Your subscription helps make it possible.

Select your digital subscription

Month selector

Use your Google account to create your subscription