Submission: Phase 2 of the ABS 2026 Census topic consultation

28 August 2023

This is the Census21 – Not Religious? campaign’s submission to phase 2 of the Australian Bureau of Statistics’ 2026 Census topic consultation.

Public submissions can be made by 8 September 2023 here on the Australian Bureau of Statistics website.

Census21 Group – Recommendations

Who we are

We are a group of secular and other organisations comprising Atheist Foundation of Australia, Sydney Atheists, Rationalist Society of Australia, Humanists Australia, National Secular Lobby and Humanists Victoria. We represent many thousands of Australians and have a common interest in accurate and meaningful government statistics to ensure fair distribution of scarce public resources, for the public good of religious and non-religious Australians.

Overall Context – Question on Religion

Our collective interest/focus for this submission relates to the question on Religion.

We also appreciate that a related ABS project is likely to result in a revised Australian Standard Classification of Religious Groups (ASCRG) and the Religious Affiliation Standard (RAS) which will be used as the coding and reporting frame for all future ABS surveys where religion is the subject of research, including the 2026 census. We refer to our previous submission in November 2022 and look forward to participating in the next phase of consultation on this major review when the proposed changes are released (now expected on 18th September).

We are grateful for this opportunity to provide feedback on the current wording of the question.

Feedback on Proposed Change

We are pleased and relieved that the ABS is considering a re-wording of the flawed question on religion.

We strongly support the move to a two-part question with the first part of the question filtering out respondents without a religion.

Currently, the data does not fully meet our needs, or the needs of all organisations and researchers who expect Australia’s largest survey to deliver accurate and meaningful data on religion. The proposed change will go some way to address this.

Our motive is to have access to the most accurate and meaningful data possible. The current question “What is the person’s religion?” is an example of a question loaded with ‘acquiescence bias’.

Acquiescence bias means the question presumes the person has a religion and the respondent is more likely to respond with a religion, rather than ‘No Religion’, even where there is a tick box with that label.

Advantages of adding a filter question

  • Adopting this change will deliver an improved data outcome for data on religion. This is because the proposed change will remove some inherent bias in the question.
  • Improved data will provide the Australian community with a more acceptable evidence-base to support policy and funding decisions related to religion.
  • It will help reduce discrepancies with evidence gained from other robust and reliable sources of data on religion, such as the Australian Electoral Study and the Australian Values Survey.
  • It is not appropriate for the largest, most comprehensive survey in Australia to perpetuate an elementary error in questionnaire design. It is even less acceptable for our country’s leading statistical agency to be the author of such a fundamental flaw.

Disadvantages of adding a filter question

  • The benefit of direct census-to-census comparison will be somewhat compromised. It is difficult to quantify the extent to which this will occur, but it is not the first time that the question presentation and the coding of responses have been changed with a similar unquantified impact. However, the loss of directly comparable time-series statistics is a small price to pay for the benefit of superior data quality.
  • A marginal increase in space required for the question.

One small step

We are also of the view that this small fix to the question is the minimum required to improve the outcome. It will not fully address the problem of misrepresenting the role that religion plays in Australian society.


In the guidance given to census respondents, ABS background information muddies the water by relying on the imprecise concept of religious ‘affiliation’. This is far weaker than ‘belonging to’ and requires clarification. The ‘operational definition’ of religion used by the ABS, which introduces the word ‘Affiliation’, should be removed. The ‘nominal definition’ of religion indicated in the background information is unambiguous and sufficient.

In our view, ‘affiliation’ is akin to asking someone to declare the football team they ‘support’. Family, neighbourhood, or other peer-group influences will encourage people to declare an allegiance with a club irrespective of whether they have ever been to a match, or if they have actively engaged in practical support of that club. There would be an outcry if a club received funding, solely based on the results of a survey about which club they ‘liked’, rather than committed memberships or attendance at matches.

Given that census data is used as the basis for decisions on funding and other government policy, the question should not be open to a variable interpretation of ‘affiliation’. The question should promote a consistency of understanding among respondents.

Sequencing of questions

Further, the placement of the question in the census introduces an additional problem. Currently it flows on from a series of questions asking about a person’s background – for example, where they were born, where their parents were born, and the ancestries they identify with. So the respondent is conditioned into a mindset about their background. A person is therefore much more likely to choose the religion they were born with, or the one in which they were brought up.

This problem could be addressed in two ways.

  1. The question could be asked at the beginning of the sequence of questions relating to cultural diversity, and/or;
  2. Introducing an additional question that differentiates between the current and former religion. This occurred in the 2021 census in Northern Ireland. where a two-part question was:
    1. What religion, religious denomination or body do you belong to?
    2. What religion, religious denomination or body were you brought up in?

As with political ‘affiliation’, beliefs are dynamic and, especially in a multicultural pluralist society, change over time. As the aim of a five-yearly census is primarily to capture a snapshot of current Australian society, the question should encourage the respondent to record their current relationship with religion.

Summary of additional steps required for better quality data

Conceptual confusion due to continued use of the insubstantial concept of ‘affiliation’ together with the inappropriate placement of the question, both encourage respondents to record their past belief, rather than the one that is relevant today.

This means that, even with the proposed change to the question, the result will still overstate the role that religion plays in Australian society.

While we support the proposed change, we believe the ABS should consider the broader scope of change recommended above to enable users of the data to be given a true understanding of the role that religion plays in Australian society. This will better inform decision-makers in their choices when allocating scarce resources.