The Australian Bureau of Statistics (ABS) has defended its decision to continue using a biased question on religious affiliation in the Census, saying that providing religious organisations with comparable data was “a significant enough consideration” to keep the question.

In a letter to the Rationalist Society of Australia, the Australian Statistician, Dr David Gruen, said there was a “strong desire” from religious organisations for the ABS not to change the religious affiliation question.

However, Dr Gruen (pictured) did not answer the RSA’s question on why the ABS chose to ignore the overwhelming feedback of its formal public consultation process that called for removal of the bias that assumes all respondents have a religion.

In February, ABS announced that it would keep the religion question, ‘What is the person’s religion?’, for the 2026 Census.

The decision followed a public campaign by Catholic church bishops and Liberal Party figures, including former prime minister John Howard, that urged the Albanese government to block any changes to the question.

Following the announcement, RSA Executive Director Si Gladman sought answers from Dr Gruen on why the ABS decided to continue with a fundamentally flawed question despite the ABS having proposed changing the question to “support more accurate data collection” in response to concerns raised in the public consultation.

Mr Gladman told Dr Gruen that the decision undermined public trust and did “immense reputational damage” to the ABS.

In response, Dr Gruen also did not answer the RSA’s question of whether it was the ABS’ job to deliver accurate data. Instead, he said “several groups” placed a “high value” on data being comparable between censuses.

“That is, they put a high value on being able to compare proportions of different religious affiliations in successive Censuses without these proportions being affected by changes in the design of the religion question,” said Dr Gruen.

“Religious organisations with a strong desire for comparable data requested the ABS provide additional support if the question changed, to help them understand the impact on response patterns.

“The ABS decided that adequately supporting data users with a strong desire to compare religious affiliation data from successive Censuses was a significant enough consideration to retain the question in its current form.”

Following the public consultation, the ABS proposed, during 2023, a reformulated question – ‘Do you have a religion?’

Catholic bishops and Liberal Party figures spoke out against the proposed change and called for the Albanese government to block it. A decision by the government in August to scrap questions about gender and sexuality derailed the ABS’ testing night in September.

In his letter, Dr Gruen said the RSA that the ABS required more extensive testing of the changes that it had proposed for the religion question.

“Doing justice to an assessment of the impact on response patterns of a changed question design required a more extensive testing program than was possible in the leadup to the 2026 Census,” he said.

“As I stressed in public commentary to which you have linked on your website, it is not possible to design a Census question or questions on religious affiliation that will satisfy all groups with an interest in this topic.

“I accept this will be disappointing for those members of the community who requested a change to the question.”

Speaking today, Mr Gladman said the ABS had chosen to ignore the many users of religious affiliation data who require accurate data instead of flawed data that can be compared with the flawed data of the past.

“The ABS has a history of changing questions when there is a need to. Yet, on the religion question, it appears that it has wanted to keep religious organisations happy instead of listening to data users in the wider public,” he said.

“We know, for example, that many elected representatives in parliaments and local governments across Australia need accurate data on religious affiliation so that they can make policies that meet the needs of the community and address human rights concerns. Yet, they cannot access any accurate data on religious affiliation from the ABS.”

If you want to support our work, please make a donation or become a member.

Si Gladman is Executive Director of the Rationalist Society of Australia. He also hosts ‘The Secular Agenda’ podcast.

Image: Australian Bureau of Statistics

Letter from David Gruen, 25 March 2025

Dear Mr Gladman

Thank you for your correspondence of 19 February 2025 regarding the design of the religious affiliation question in the 2026 Census.

As you are aware, the ABS recently published outcomes from the 2026 Census topic review, which included the decision to retain the 2021 Census religious affiliation question design for the 2026 Census. As part of the review, the ABS assessed all questions against a set of pre-determined criteria, including whether there was a need for continuing data, and the strength of the desire among groups in the community for data that was strictly comparable between Censuses. This assessment was informed by extensive consultation, engagement and testing.

Several groups provided feedback that they placed a high value on data on the religion question that was strictly comparable between Censuses. That is, they put a high value on being able to compare proportions of different religious affiliations in successive Censuses without these proportions being affected by changes in the design of the religion question.

We also received feedback making the case for a change to the religious affiliation question for the 2026 Census. The Rationalist Society of Australia was one of the groups providing this feedback to us.

Religious organisations with a strong desire for comparable data requested the ABS provide additional support if the question changed, to help them understand the impact on response patterns. Doing justice to an assessment of the impact on response patterns of a changed question design required a more extensive testing program than was possible in the leadup to the 2026 Census.

The ABS decided that adequately supporting data users with a strong desire to compare religious affiliation data from successive Censuses was a significant enough consideration to retain the question in its current form.

As I stressed in public commentary to which you have linked on your website, it is not possible to design a Census question or questions on religious affiliation that will satisfy all groups with an interest in this topic. I accept this will be disappointing for those members of the community who requested a change to the question.

Information on the topic review process, including the outcomes from different stages in the process, is available at https://www.abs.gov.au/census/2026-census-topic-review. Thank you for interest in this important issue.

Yours sincerely

Dr David Gruen AO

Australian Statistician