You are here

Should We Have a Census? Issues of Design and Use

As far as I can tell only two news services, the Australia Financial Review and the Chinese newspaper XINHUANET, carried reports on comments from David Gruen, the head of the Australian Bureau of Statistics (ABS), that the ABS might do away with the Census and replace it with other data collection processes.

The mandatory Census of Population and Housing has been defined as the most significant statistical event in Australia and we are considered fortunate to have one every five years while some other countries, such as the UK and America, only have one every ten years. But it is not sacrosanct. What is essential is good data.

Are there alternatives to the Census? 

The census is expensive (one estimate is $0.5 billion) so in a world awash with data, is there an alternative? This can only be answered if we have a thorough and professional understanding of what data we would like, what data we need, what data we can handle and how representative is the information.

We do not need a decision based mainly on cost, or on getting figures that are so general, they fail to give clarity and can easily become narrow political weapons. Nor do we do not need the debate to turn into a media circus of short sound grabs where one number is said to tell the whole story.

Personal views on a social issue are irrelevant.

Furthermore, we do not need to judge the whole census from a personal view. More media coverage has been given to the Australian Bureau of Statistics release of new guidelines on gender and sexual orientation, aimed at improving data collection on diverse gender identities. Any of our opinions on gender identity is irrelevant. The only issue should be whether this information is sufficiently helpful to warrant the cost of inclusion.

The quest to determine the need for a census or to find a better way, is made all the more difficult as we live in a time where a professional’s opinion is often demeaned and is overshadowed in the public debate. 

At the time of writing, America has struggled to gain broad acceptance of professional input into the pandemic debate and is also going through an angry dispute about whether their census has been politicised: the director of the U.S. Census Bureau has just resigned. I will write a separate article on some thoughts on overseas experiences at a later date.

We need to approach this with the full input of data scientists and practical users, not just emotional comments or an accounting bottom line.

We need the will and the resources to make use of the data

The process of conducting a census is so large that it often drains the energy from reviewing the on-going trends and consequences. One of the great elder statesmen of the planning profession, Bob Myer Fellow at Cox Architecture in Sydney has said that, in his opinion based on reviewing all the major transport studies in Sydney from 1909 to the present, government departments should be made to review the conclusion made on census data every year (he even said, tongue-in-cheek, that not updating to the latest trends should be a criminal offence!). The 1948 County of Cumberland plan underestimated the population influx of immigration while the 1968 Sydney Region Outline Plan underestimated the increase in housing stock particularly in the west of Sydney. This is not challenging the Census data but the effort to review the situation was under-resourced.

Whichever process we evaluate it must be based on thoughtful consideration of what we really need and how we plan to achieve good planning even after we get good data.

Data Science – Good for business; good for government 

There is a huge growth in data science in businesses, as they strive to refine their processes. At the same time, a considerable amount of government and media energy is being given to oversimplifying issues into slogans and shouted opinions.

Using data just because it is easily available does not guarantee success.

I believe we should take a reverse engineering approach to our review of the census process. Rather than theoretically consider what would be nice, we should look at how the information is currently being used and what additional information may be helpful. Then work out how to collect it. 

And of course, we need to be collecting data not just for the things that are happening now but as a basis on which future changes can be measured. The ABS is considering adding a question about any significant health issues with people in the household. This is more than just a question on COVID and should continue even if and when we get on top of the current pandemic.

I would think a list to help guide us in any review should consider the following:

  1. It is the professional users of the information and data scientists who should assess the value of data obtained through the current census and alternative processes.
  2. Our assessment of what we need must consider the value of recorded historical trends and whether continuing such data series would be helpful.
  3. The value of general census type data can be greatly enhanced if we can consistently link it to other, more detailed, information.
  4. We must provide the resources to analyse what we collect. A large amount of data is often requested but then “sits in a draw” in the hope that we get around to analysing it some time. 
  5. If we propose new approaches, we should trial them in parallel to the current census to determine the differences (which could include identifying weaknesses in the current system).
  6. Technology for technologies sake is a catastrophic approach. Just because we can get large numbers from modern technology does not mean it is accurate or representative. We need to define the requirements THEN look at how we achieve them.
  7. Technology is not necessarily appropriate or accurate simply because the system can present colourful graphics. Presentation should not overshadow the need to base our understanding on good data principles.

References

  1. SBS News: “Australian census to include a gender non-binary option for the first time” BY Maani Truu 15/01/2021 https://www.sbs.com.au/news/australian-census-to-include-a-gender-non-binary-option-for-the-first-time
  2. Australian Financial Review: “Next census could be last five-yearly population snapshot” By Tom Mcllroy  - 12/8/2020
  3. APnews: “Time, transparency needed as Biden inherits frazzled census”: By MIKE SCHNEIDER January 22, 2021 https://apnews.com/article/joe-biden-census-2020-9fcf5cac65553490f8544f74026189e7
  4. The American Association for the Advancement of Science: “Statisticians hail early departure of U.S. Census director” By Jeffrey Mervis Jan. 19, 2021 https://www.sciencemag.org/news/2021/01/statisticians-hail-early-departure-us-census-director
  5. Bloomberg City Lab: “The 2020 Census Is Being ‘Sabotaged,’ Says Leading U.S. Statistician” By Kriston Capps October 15, 2020 https://www.bloomberg.com/news/articles/2020-10-15/top-statistician-u-s-census-is-being-sabotaged
Author

John Reid

Managing Director, Austraffic

From the beginning of his career in local government and then when he established Austraffic in 1983, John realised that data collection is not just about numbers but about understanding people and the activities that serve the community's needs.  Poor or even bad data is counter-productive.  Even if results fit our preconceived ideas that doesn’t mean it is accurate. John has seen how good data expands our perceptions and thinking and can be surprising in its results. Connect with John on LinkedIn.

John Reid