Home > Life in General > Census crowdsource (repost)

Census crowdsource (repost)

July 16th, 2017

In 2012, I crowdsourced an analysis of the census results, looking at the extent to which the increase in religion was driven by changes in stated affiliation from religious to non-religious, as opposed to the demographic replacement of older more religious cohorts by younger, less religous ones. A couple of wrinkles on this

* I didn’t mention immigration last time. It appears (unsurprisingly) that those born overseas are more likely to be religious, but less likely to be Christian, than the Australian born.

* As the ABS notes

The religious pattern of those under 18 is most similar to the 35-49 year olds, suggesting the form may be completed with their parents’ beliefs.

It seems likely that when they report for themselves, these young people will be more like the 19-34 age group. It’s hard to say whether we should call this an affiliation change or a cohort effect.

I’d like to ask again for a crowdsourced analysis. It may be useful to read the comments thread to my previous request.

Repost from 2012

I’ve seen a bunch of reports from the census saying that the proportion of Australians reporting “no religion” has increased substantially, to around 22 per cent. I’d be interested to know if this is mainly a cohort effect (non-believing younger generations entering the population) or the result of people who previously reported a religious affiliation switching to reporting none. I’d be surprised if much of it was the result of people abandoning previous religious beliefs, as opposed to nominal affiliations, but I don’t think the data allows a test of this.

I just had a brilliant idea for how to motivate this effort. The first person to give a good answer gets to nominate the next topic for crowdsourcing. As a hint, the ideal way to answer the question would be to compare responses from a given age group in 2006 with the same group, now 5 years older, in 2011, adjusting, if possible for migration effects.

Update: The evidence, collected in the comments threads, suggests that cohort and conversion effects each account for about half of the shift.

The prize goes to David Barry, with honorable mentions to Aldonius and Luke Elford. I’ll give Dave first shot at proposing a new topic (in comments), but also invite suggestions from Luke and Aldonius. Meanwhile, I’m going to suggest something a bit more challenging for crowd-sourcing. If anyone would like to use the data to develop a simple model to project likely changes in stated Census affilations over the next two decades, with a specific focus on the question “When will (Census reported) Christian affilation become a minority response in Australia”, I’ll add a write up and send it as a joint post to The Conversation, the new(ish) academic-focused website.

Categories: Life in General Tags:
  1. David Barry
    July 16th, 2017 at 19:36 | #1

    You did post a follow-up crowd-sourced effort with my question! http://johnquiggin.com/2012/06/24/crowdsourcing-contest-global-research-on-disease/

    It was another David who gave the best answer (comment #12).

  2. John Quiggin
    July 16th, 2017 at 20:26 | #2

    Thanks, I should have looked more carefully. I’ll delete that part of the post.

  3. Luke Elford
    July 16th, 2017 at 23:50 | #3

    To estimate the relative significance of changing affiliation, consider those Australian-born and aged between 25 and 99 in 2016, broken down into five-year groups. These were adults in 2011 (aged between 20 and 94) and presumably able to fill out the census in accordance with their actual beliefs in both years. The percentage choosing the no religion/no religious affiliation option for the 25-99 age group (NOT cohort) increased from 21.7% in 2011 to 31.1% in 2016, an increase of 9.3% (using unrounded data). The corresponding figures for all Australian-born respondents are 24.5%, 33.8% and 9.3%.

    For the five-year cohorts, we can ascribe the 2011 responses in place of the 2016 responses for the 2016 groups—e.g. replace the response for those aged 25-29 in 2016 with the response given by those aged 20-24 in 2011. Doing so implies that, if affiliation were unchanged since 2011, 23.4% of those aged between 25 and 99 would have chosen the no religion option at the 2016 census. Of the 9.3 percentage point change that occurred, only 1.6 percentage points can be explained by the entry and exit of cohorts into and out of the 25-99 age group, leaving a 7.7 percentage point change that results from changing affiliation.

    Of course, some of the change ascribed to affiliation above could reflect entry into the cohorts (e.g. returning from overseas) and exit from the cohorts (e.g. death).

    For further detail, here are the increases, for each cohort, in the percentage of respondents choosing the no religion option between 2011 and 2016. The listed age ranges are for 2016.
    25-29 14.3%,
    30-34 9.9%,
    35-39 8.0%,
    40-44 7.5%,
    45-49 7.6%,
    50-54 7.3%,
    55-59 6.9%,
    60-64 6.2%,
    65-69 5.4%,
    70-74 4.6%,
    75-59 3.9%,
    80-84 3.9%,
    85-89 4.0%,
    90-94 4.0%,
    95-99 5.3%.

  4. Melinda Barnard
    July 17th, 2017 at 04:32 | #4

    Hello John and others who will read this. I firstly apologize if this is the incorrect space to communicate with you John and if so I hope my following reasons will explain.
    I refer to a comment that I have just read which you posted on this forum in 2006 when the shareholders of Earth Sanctuaries were advised that the company was being de-listed. Like you, I was also a shareholder. Your comments interested me, as I believe that this was actually before its time and I am considering a similar concept again now. John’s visions, similar to my own, was to give back 1% of Australia to the native wildlife through feral fencing and providing sanctuaries. Many people invested, or ‘made a donation’ as you say with what I would hope to be a similar belief in the pure vision of John Wamsley. I am in a position where I have been fortunate enough to continued John’s dream as an owner of an Earth Sanctuary. I am exhausting all avenues to protect land that was part of John’s dream. When visions die and no body is left fighting for them, the things that should be at the forefront of everyone’s minds, like protecting our environment are ignored and over time completely overlooked. I am taking a stand and wondering whether you would consider becoming an investor once again? Do you know of any forums that I may be able to promote and further explain my cause? Any comments, advise, direction, or alternatives from all would be extremely appreciated.
    Thanks for taking the time to read.
    Melinda

  5. Svante
    July 17th, 2017 at 20:33 | #5

    @Melinda Barnard

    Would government come to the party? If at all suitable, what would it take to bring government in as a partner? Change.org petitions? Queensland had spent what seemed sufficient on the Currawinya National Park bilby fence, but after the decimation of bilby numbers by feral cats following flood damage to the fence they appear now to be literally doubling down on it.

    http://www.abc.net.au/news/2016-03-22/bilby-repopulation-predator-proof-fence-repair/7266616

    http://www.abc.net.au/news/2015-06-15/bilbies-currawinya-national-park-size-doubled/6547308

  6. Greg McKenzie
    July 18th, 2017 at 09:00 | #6

    Be careful with trend analysis of Census data. As the Henderson Commision of Inquiry into Poverty and Wealth in Australia showed all too well, assuming that a current trend will continue leads to erroneous outcomes. accurate Statistical analysis requires no inherent biases and prejudices to cloud the vision of raw data. Census data, in particular, may not follow current trends for countries with large immigrant intakes. To presume that one trend will continue over many census periods, is to establish a blind spot for the statistical analysis of census data. Economists, in particular, should have become aware, after 2008, that assumptions are dangerous to the accuracy of any predictions.

  7. pablo
    July 18th, 2017 at 10:22 | #7

    One trend that was maintained in the latest census is the numbers identifying as indigenous, up 17.4 percent on the 2011 census. While a substantial increase was evident in NSW and Queensland seaboard trend figures, the area around Alice Springs/southern NT sowed a decrease.
    I am curious as to how valid this is and what possible reasons could explain it…low turnout on the night, actual population decline, or could fewer people previously identifying opt to be different in 2016. I take Greg McKenzie’s view that “assumptions are dangerous” but would appreciate a demographers view

  8. may
    July 18th, 2017 at 15:22 | #8

    pablo :
    One trend that was maintained in the latest census is the numbers identifying as indigenous, up 17.4 percent on the 2011 census. While a substantial increase was evident in NSW and Queensland seaboard trend figures, the area around Alice Springs/southern NT sowed a decrease.
    I am curious as to how valid this is and what possible reasons could explain it…low turnout on the night, actual population decline, or could fewer people previously identifying opt to be different in 2016. I take Greg McKenzie’s view that “assumptions are dangerous” but would appreciate a demographers view

    sowed or showed?

    while not a demographer,
    maybe the Alice and south NT people moved to NSW and QLD?

Comments are closed.