The way Australia asks about religion shapes the answer it receives.
Every five years, Australia takes stock of itself through the census, and how it asks about religion quietly shapes what it finds. A new survey suggests that the current question format — a list of faiths with 'no religion' as one option among many — may be systematically undercounting the secular population by as many as two million adults. The Australian Bureau of Statistics has declined to change the format for 2026, caught between the demands of historical continuity and the possibility that decades of data rest on a subtly tilted frame. In this, Australia confronts a quiet truth about self-knowledge: the questions a society asks about itself are never entirely neutral.
- A new poll reveals that how Australia asks about religion may matter as much as what Australians actually believe — a yes-or-no question yields 54% non-religious, while the current list-based format produces only 43%.
- That eleven-point gap represents roughly two million adults, and census data directly drives funding allocations, policy decisions, and the nation's official understanding of its own identity.
- The Australian Bureau of Statistics considered reforming the question after 2021 but backed away, citing fears of breaking decades of comparable data — and, critics allege, pressure from religious organisations with a stake in the status quo.
- For 2026, only cosmetic changes have been made; the fundamental question design remains intact, meaning the potential undercount will persist into another cycle of national self-portraiture.
- Campaign groups are already mobilising for 2031, armed with survey evidence and a growing generational tide — among Australians aged 18 to 34, only 34% reported having a religion under the cleaner format.
In August, Australia will conduct its five-yearly census, and a new survey raises an unsettling possibility: the nation may be considerably more secular than its official statistics suggest. The issue is not what Australians believe, but how they are asked.
An Essential Media poll tested two question formats on the same respondents. Under the current census design — a pre-printed list of religions alongside 'no religion' and 'other' — 43% chose no religion. When asked instead a simple yes-or-no question first, followed by a text box for those who answered yes, 54% said they had no religion. The difference amounts to roughly two million Australian adults.
The stakes are real. Census figures shape funding, policy, and public self-understanding. At the 2021 census, 38.9% selected 'no religion' under the existing format, while Christianity held at 43.9% — declining but still dominant. Campaign spokesperson Michael Dove argues the census is the 'gold standard' institutions depend on, and that a biased question design corrupts the entire foundation.
The Australian Bureau of Statistics examined a format change after 2021 and declined. Officials pointed to data continuity — altering the question would sever the thread of historical comparison — and acknowledged feedback from religious organisations opposed to any redesign. Dove believes those organisations carried undue influence. 'I think they have been compromised,' he said. For 2026, the ABS has made minor adjustments to ordering and instructions, but the core format is unchanged.
The religion question has always attracted friction. In 2001, over 70,000 Australians declared themselves Jedi Knights; Pastafarians have used the census to register quiet protest. These moments of levity point to a genuine tension: the question may not capture what people actually believe or have ceased to believe.
Age patterns sharpen the concern. Among 18-to-34-year-olds, only 34% reported a religion under the yes-no format. Secularisation appears generational, suggesting the gap between official figures and lived reality will only grow. The campaign has turned its attention to 2031, hoping to press for genuine reform before another decade passes. For now, Australia will count itself the way it always has — and the frame, as ever, will shape the picture.
In August, Australia will conduct its five-yearly census, and a new survey suggests the nation may be significantly more secular than official statistics have claimed. The question is not about what Australians actually believe—it's about how they're asked.
The Essential Media poll, commissioned by the Census – Not Religious? Mark No Religion campaign, tested two different ways of asking about religious affiliation. Under the current census format, where respondents choose from a pre-printed list of religions plus "no religion" and "other," 43% selected "no religion." But when the same people were asked a simpler yes-or-no question first—"Do you have a religion?"—followed by a text box for those who answered yes, the result shifted dramatically. Fifty-four percent said no. That difference represents roughly 2 million Australian adults.
The implications are substantial. Census data shapes policy decisions, funding allocations, and public understanding of who Australians are. If the current question format is systematically undercounting the non-religious population, then the nation's actual religious composition has been misrepresented for decades. At the 2021 census, 38.9% of respondents selected "no religion" using the existing format. Christianity remained the largest category at 43.9%, though it has declined steadily over the past half-century. The trend toward secularization is clear, but the magnitude may be obscured by the way the question is posed.
Michael Dove, a spokesperson for the campaign and self-described "demography nerd," frames this as a matter of data integrity. The census, he argues, is the "gold standard" that governments and institutions rely on to make informed decisions. When the question design itself may be biasing the results, the entire foundation becomes questionable. The campaign has compiled evidence from multiple surveys suggesting that the non-religious population is consistently higher than census figures indicate.
The Australian Bureau of Statistics considered changing the question after 2021 but ultimately rejected the proposal. Officials cited two concerns: any change would break the chain of historical comparison, making it impossible to track trends reliably across decades, and they had received feedback from religious organizations opposed to reformatting. The ABS consulted with both religious and secular groups and concluded it could not design a question that satisfied everyone's needs while maintaining data continuity. For the 2026 census, the bureau has made smaller adjustments—reordering categories, providing clearer instructions, and changing how data is processed—but the fundamental question format remains unchanged.
Dove believes the ABS has been swayed by lobbying from religious groups with the strongest vested interests in maintaining the current system. "I think they have been compromised," he said. "It needs to be fixed." But he acknowledges the moment has passed for 2026. The campaign is already looking ahead to 2031, when they hope to push for a genuine redesign.
The history of the census religion question reveals how even technical decisions can become cultural battlegrounds. In 2001, over 70,000 Australians declared themselves Jedi Knights, inspired by Star Wars. Pastafarians—members of the Church of the Flying Spaghetti Monster—have deliberately selected "no religion" to avoid being counted as Christians. These moments of humor mask a deeper frustration: the question itself may not be capturing what people actually believe or don't believe.
The poll also found that religious affiliation varies by age and education. Among those aged 18 to 34, only 34% reported having a religion when asked the yes-no format, compared to 50% among those with university education. These patterns suggest that secularization is generational, and the gap between the current census figure and the true proportion of non-religious Australians may only widen as younger cohorts age into the population.
For now, the 2026 census will proceed with the existing question. But the survey has documented, with precision, that the way Australia asks about religion shapes the answer it receives. Whether that gap between 43% and 54% represents a flaw in measurement or a genuine ambiguity in how people understand their own religiousness remains contested. What is certain is that the nation's official portrait of itself—religious or secular—depends partly on the frame.
Notable Quotes
The census is the gold standard of data needed to inform debate, policy and funding decisions. We trust the ABS to deliver high-quality data we can rely on.— Michael Dove, Census – Not Religious? Mark No Religion campaign
I think they have been compromised by lobbying from religious groups with the strongest vested interests. It needs to be fixed.— Michael Dove
The Hearth Conversation Another angle on the story
Why does the format of a question matter so much? Isn't the answer the same regardless of how you ask it?
Not quite. When you see a list with "Christianity," "Islam," "Buddhism," and then "no religion" at the bottom, your eye lands on the religions first. The question assumes religion is the default. A yes-or-no gate changes the cognitive path entirely.
So the 11-point gap—from 43% to 54%—is just psychology?
It's not "just" anything. It's real. If 2 million more Australians are non-religious than the census shows, then policy makers are making decisions based on a distorted picture of the country.
Why did the ABS reject the change?
Two reasons, officially. One is genuine: changing the question breaks historical comparison. You can't track trends if the measuring stick changes. The other is political. Religious organizations lobbied hard to keep the current format.
Could both be true?
Yes. The technical concern is legitimate. But it's also convenient for those who benefit from the current numbers.
What happens now?
The 2026 census uses the old question. The campaign is already preparing for 2031. They've proven their case with data. The question is whether the ABS will listen next time.