Many media outlets have noted the drop in religiosity indicated by the latest Australian Bureau of Statistics (ABS) census results, and I tend not to reiterate the same stuff the mainstream media have come up with unless there is something else that can be said. However, the disparity between this census and other recent surveys is something that has piqued my interest.
Over the last few years, I have written a couple of articles about the decline of religion in bits of the western world. The first was written in the middle of 2019 and its Australian data were based on the results from the 2016 census, where 30.1% of the populace reportedly had no religion1. The second article, written in late 2021, looked at a couple of more recent surveys. The first was the Household, Income and Labour Dynamics in Australia (HILDA) survey run by the University of Melbourne. The most recent data from that survey was from 2018, two years after the most recent census mentioned in the previous article. The HILDA survey indicated that about 40% of the sample had no religion. Also mentioned in that article was a Roy Morgan survey from 2020 which indicated that 45.5% of Australians had no religion2. The third article was mostly about a Pew Research Center survey into religiosity in the United States3.
As almost everyone knows, the results of the 2021 census were released today, and that included the proportion of those who voluntarily responded to the religion question. This census indicates that 38.9% of the Australian populace have no religion. In 2016, that proportion was 30.1% (see above), so that is an increase of 8.8% in 5 years. This 38.9% is not a lot less than the proportion identifying as Christian (43.9%). In the 2016 census, those identifying as Christian comprised 52.1% of the populace. That is a decrease of 8.2% in 5 years4. The religion question in the census is the only question that is voluntary, and despite it being so, there was an increase in the proportion of people answering the question. In 2016, it was about 91%, whereas last year it was 93%4.
Why there is such a disparity between the HILDA and Roy Morgan surveys and the most recent census is something that struck me as odd. However, there has been much written about the accuracy of the census and what it actually measures. Researchers know some respondents will answer based on their religious beliefs (or lack thereof), while others will answer based on cultural identity, or their parents’ religious belief. Liz Allen, a demographer at the Australian National University said: “The current census question doesn’t necessarily reflect religiosity – the level of importance that we invest in religion, and whether or not we practise our religion, and even belief in the faith that corresponds to that religion.” The ABS said the question is about a person’s “affiliation to a religion” and is part of a section about cultural identity. Groups such as the Rationalist Society of Australia have suggested this means the official figures for religious affiliation are inflated5.
In 2021, the National Secular Lobby (NSL) released the results of two polls run during July of that year which attempted to analyse the census’s measurement of religiosity in Australia. The organisation stated that the approach taken by the ABS of asking for a person’s religion without first asking people whether they had one, will suggest to some that they are expected to respond with a religion, even if it’s one they were only exposed to in childhood and with which they do not currently identify6.
The poll, conducted by Essential Research in two consecutive periods throughout July, compared the responses from two different questions:
- The question on religion as asked by the ABS in the national census; and
- An alternative two-part question designed to be non-leading by first asking people whether they currently had a religion, before (in the case of a positive response) asking which one.
The NSL’s reproduction of the ABS’s question, designed to model the increase in the ‘No Religion’ demographic in the five years since the last census, returned a ‘No religion’ figure of 41% and a religious figure of 56%.
Their non-leading alternative question resulted in a ‘No religion’ figure of 52% and a religious figure of 44%6.
The result from the first ‘leading’ question (41%) was roughly in line with the result from the 2021 census itself (38.9%), whereas the result from the non-leading question was very different, and seems more or less on trend with the HILDA and Roy Morgan surveys, given that they were taken in 2018 (40%) and 2020 (45.5%) respectively. So, the 52% may more accurately represent reality than does the 38.9% of the census.