The latest Newspoll, published last night, showed on federal voting intentions, Labor and One Nation lost 1 percentage point to both sit on 29%, while the Coalition also dropped 1 percentage point to 18%. All of these polls are dutifully reported not just by the media outlet that pays for them, but by other media too.
Author
- Daniel Casey
Lecturer, Politics and International Relations, Australian Catholic University
Yet, we are just over one year out from the most recent federal election, and two years away from the next. Given this, the way the recent rise in One Nation’s polling in particular has shaped political debate, do we have too many polls?
Voice of the people
“Vox populi, vox Dei” is the 1709 idea that the voice of the people is the voice of God. Surely, then, opinion polls – which accurately, scientifically and systematically measure the voice of the people – are one of the most important tools in democracy? Or are voting intention polls a shallow measure of public opinion – noisy, but telling us little about why opinion is shifting, or what politicians should be doing? Is the obsession with the latest polling distracting us from the actual issues and the long-term problems facing Australia?
According to respected poll tracker William Bowe (aka The Poll Bludger ), around 12 polls were released in July 2026 – roughly three every week. We have now had around 140 polls from ten different pollsters since the 2025 election – roughly two every week.
If it seems it wasn’t like this before, well, you’re half right. While it’s difficult to track precisely the number of published opinion polls, until the early 1970s Australia had only one major polling organisation, Roy Morgan Research, which produced polling for The Bulletin.
According to the Parliamentary Library in the 1990s, there were roughly 38 polls per year – fewer than one a week. That doubled in the 2000s, before jumping to around 130 a year between 2010 and 2016.
This coincided with the Australia’s place as ” coup capital of the democratic world ” – several leadership coups in both major parties. While there was then a dip for a couple of election cycles, by the 2022-25 election cycle we were back up to the current average of about two polls a week.
This growth is mirrored elsewhere. In the United Kingdom between 1992 and 2005 there were around 1.5 opinion polls per week. Since then, it has more than tripled to around five per week since the 2024 general election.
How does it affect political debate?
This isn’t simply a story about volume, though – it’s what happens to all that polling once it reaches the public that matters. So, what does the research say about the effects of all this horse-race-style coverage?
With a limited amount of space for political news, an increase in such journalism almost certainly comes at the expense of substantive issues coverage.
US research indicated only 10% of news coverage in the 2016 presidential election focused on issues. The rest was horse-race. Explaining the complexities of taxing trusts and capital gains tax – rather than reporting on the impact of these announcements on public opinion – is difficult, time-consuming, resource-intensive, and unlikely to grab eyeballs and clicks on social media.
How does horse-race journalism impact the public? Does it turn them off, making them even more cynical about politicians who seem completely poll-driven? The assumption is this increases political alienation, negatively impacting democracy. Others have suggested this sort of coverage is accessible, and encourages civic engagement.
Unfortunately there is limited Australian research on the topic. However, research across the United States and half a dozen European countries all seems to point in the same direction. A meta-analysis of 32 studies by Alon Zoizner found that strategic, horse-race-style coverage raises cynicism towards both politicians and the media covering them, and lowers audiences’ evaluation of the credibility of the news itself.
There are similar findings in Sweden , while a Danish study found journalists routinely report small, statistically meaningless shifts in the polls as if they were real, substantive changes. This manufactures a sense of movement and news value out of pure statistical noise.
Taken together, an increase in horse-race-style reporting leads to increased distrust in politicians and media, and an uninformed electorate – one being fed movement in the polls that, statistically, often isn’t there at all.
Journalists are not entirely to blame. There has been a significant hollowing out of newsrooms, with thousands of journalist positions disappearing, leading to a decline in the depth of subject-matter knowledge, and fewer reporters with the grounding to interrogate a set of numbers rather than simply relay them.
Beyond newsroom cuts, there has been a range of structural shifts in the media industry that rewards immediacy and clickbait headlines. In these circumstances, it’s easier to reach for the latest pre-packaged opinion poll or press release than to commission original reporting or spend a week working through the fine print of a tax policy.
However, there are some concrete steps that could be taken. Journalists and editors could choose to downplay opinion polls and improve how they are covered. While the Australian Press Council has guidelines on reporting on opinion polls, the media’s coverage often falls short of these standards and can be significantly misleading. There is also evidence that providing journalists with training on improving community engagement can reduce horse-race journalism.
Politicians reject or embrace polling depending on what suits them – former prime minister Malcolm Turnbull relied on bad polling to justify challenging his predecessor Tony Abbott, then came to regret citing it once his own numbers turned.
My analysis of MPs’ e-newsletters shows the same instinct more broadly: polls are routinely wielded as proof of momentum, or as a scare tactic to mobilise supporters. “Recent polling suggest[s] if an election were called today Peter Dutton would win,” Katy Gallagher warned supporters in February 2025.
If you think it’s bad now…
So, are there too many polls? Probably – but volume alone isn’t really the villain. The problem is volume combined with how it is interpreted, and how readily politicians weaponise it, whether by brandishing the numbers or theatrically dismissing them. Neither move is really about the polling itself; both are performances for an audience that the sheer glut of polls has made possible in the first place.
And the glut is likely to get worse before it gets better. Researchers have already built AI “respondents” good enough to pass as human in a survey – synthetic voters that cost cents rather than dollars to poll, and can be tuned to return whatever result their operator wants.
If today’s polling problem is expensive, noisy and occasionally sloppy, tomorrow’s could be cheap, invisible and engineered. The current barrage of too many polls may end up looking like the good old days.
