












































AUSTRALIAN POPULATION STUDIES 2018 | Volume 2 | Issue 2 | pages 59-61 
 

© Reimondos, Gray and Evans 2018. Published under the Creative Commons Attribution-NonCommercial licence 3.0 
Australia (CC BY-NC 3.0 AU). Journal website: www.australianpopulationstudies.org 

How moving a box changed the 
religious landscape of Australia 

Anna Reimondos*  Australian National University 
Edith Gray  Australian National University 
Ann Evans  Australian National University 

* Corresponding author. Email: anna.reimondos@anu.edu.au. Address: School of 
Demography, ANU College of Arts and Social Sciences, The Australian National University,     
9 Fellows Road, Acton, ACT 2601, Australia 

Paper received 1 June 2018; accepted 20 July 2018; published 12 November 2018 

This data visualisation is the winner of the Australian Population Studies 2018 population data 
visualisation competition, kindly sponsored by Geografia Pty Ltd. 

 

In this data visualisation we use the 2011-2016 Australian Census Longitudinal Dataset (ACLD) to 
examine how answers to the question on religious affiliation changed between 2011 and 2016. A 
question on religious affiliation has been included in every census since the first national census of 
1911. It is the only question in the census which is optional, as outlined in the Census and Statistics 
Act 1905, and it is also the question that has arguably attracted the most controversy over the years. 

Up until 1991 the question on religion was in the form of an open text question. In 1991 the format 
changed to pre-defined check boxes listing several main religions including Catholic, Anglican (Church 
of England) and Uniting Church. The check boxes were followed by a space for an open-text answer 
where people could write any ‘Other’ religion that was not included in the pre-defined list. Finally at 
the bottom was a check box for ‘No religion’. Although the content and order of the pre-defined list 
of religions changed between 1991 and 2011, in all years up to and including 2011 ‘No Religion’ was 
always on the bottom after ‘Other’.  

In a public consultation on the content of the 2016 Census, nearly half of the 915 submissions related 
to the question on religious affiliation (ABS 2013). Following a review, the format of the question 
changed and for the first time in 2016 ‘No Religion’ was placed at the very top of the list of pre-
defined religions. 

Although the percentage of Australians identifying as not having a religion has been increasingly 
steadily since the first census in 1911, between 2011 and 2016 the percentage stating ‘No Religion’ 
increased dramatically from 22.3% to 30.1%. This 35% increase is significantly higher than the 20% 
increase between 2006 and 2011 and the 19% increase between 2001 and 2006. 

To observe the change in behaviour following the re-formatted question in the 2016 Census we used 
the ACLD, which is a representative 5% sample (1.2 million records) from the 2011 Census linked to 
the corresponding records from the 2016 Census. We used the ABS TableBuilder to cross-tabulate 
people’s responses to the religion question from 2011 and 2016. This table was re-formatted in Stata 

D
em

oG
ra

ph
ic

 

 

http://www.australianpopulationstudies.org/
mailto:anna.reimondos@anu.edu.au


60 Reimondos, Gray & Evans Australian Population Studies 2 (2) 2018   

 
to create an individual record of each combination of possible religions in 2011 and 2016. The data 
was then inputted to and open source data visualisation framework at RAWGraphs1. The resulting 
SVG code was then further edited to create the desired image. 

The visualisation is shown in Figure 1. On the left of the diagram is the distribution of religious 
affiliation in the 2011 Census, in order of magnitude, and on the right are the answers given in 2016. 
In 2011, Catholic was the dominant religious affiliation answered by 26% of the ACLD sample. In 2016 
‘No Religion’ had become the prominent religious belief with Catholics in second place.  

Census respondents who answered ‘No Religion’ in 2016 had previously identified with a wide variety 
of other religions. Not surprisingly, the majority of those stating ‘No Religion’ in 2016 had also stated 
‘No Religion’ in 2011. However among those who did state a religion in 2011, the percentage who 
switched to ‘No Religion’ varied considerably. For example only 11% of those who previously stated 
they were Catholic chose ‘No Religion’ in 2016. In contrast a fairly large proportion (21%) of Anglicans 
changed to ‘No Religion’. Of those who did not state a religion in 2011, just under half (46%) chose 
‘No Religion’ in 2016. For some religions such as Eastern Orthodox, Islam and Hinduism only a small 
proportion switched to ‘No Religion’.  

For questions involving unordered categorical responses, such as the religion question in the census, 
the order in which the options are presented has been shown in previous research to be important. 
The primacy effect suggests that people may choose the first acceptable answer presented to them, 
particularly if the list of response categories is long (Tourangeau et al. 2000). Evidently the change in 
question format, and moving the box titled ‘No Religion’ to the top of the list of pre-defined options 
had a significant impact on the way in which the 2016 question on religious affiliation was answered. 

 

References 

ABS (2013) Census of Population and Housing: Consultation on Content and Procedures, 2016. Catalogue 
No. 2007.0. Canberra: ABS. 

Tourangeau R, Rips L J, and Rasinski K (2000) The Psychology of Survey Response. Cambridge: Cambridge 
University Press. 

 

                                                           
1 http://app.rawgraphs.io 



Australian Population Studies 2 (2) 2018 Reimondos, Gray & Evans 61 

 

 

Figure 1: Answers to the question on religion in the 2011 and 2016 censuses 
Source: Australian Bureau of Statistics, 2011-2016 Australian Census Longitudinal Dataset 

 


