In a world awash with data, is the census still relevant?

Liz Allen, Australian National University

How we track our economy influences everything from government spending and taxes to home lending and business investment. In our series The Way We Measure, we’re taking a close look at economic indicators to better understand what’s going on.


The Australian Census came under intense scrutiny in the wake of #censusfail. Parliament conducted a review, the Senate an inquiry, and some in the media questioned the entire point.

But cost and privacy concerns aside, population is one of the three pillars of the economy.

Understanding population characteristics is vital to inform us of challenges and opportunities, and is a necessary input in other economic indicators. The quality and timely population data found in the census is not gathered through any other means. If changes need to be made, it’s in the discussion around the census.

So we know who is where

The census is unique in that it is a total survey of the population, covering a range of social and economic variables. At present, it is the only way such data is obtained in Australia.

Without the census, we wouldn’t know how many we are, who we are and where we live. This means important planning and policy issues couldn’t be addressed. The location of schools and hospitals, provision of medical facilities, funding for major infrastructure would all be done without an accurate idea of who is where.

In fact, local, state and federal governments rely heavily on data only available in the census. The number of children, working age population, travel to work information, occupations, housing suitability and vulnerable populations is all data only found in the census.

The census also allows for sub-national analyses to be performed, particularly legislated population estimates and projections. These estimates form the basis of economic indicators such as labour force statistics and gross domestic product per capita.

The estimates and projections also highlight inequalities within society, and provide opportunities for policy responses and development at a regional level.

But the purpose of taking a census goes beyond informing resource allocation, taxation and electoral representation.

The statistical benchmarks used in surveys and studies, research and analysis, and, most importantly, lower level aggregates and groups of interest can only be informed by census data. Low level aggregates allow identification of need. Identification of areas with high proportions of young people who cannot access employment or education can provide much insight into barriers to economic participation.

Quality information about homelessness, minorities, and Indigenous populations is only truly obtained via a census.

The data we already collect won’t do

One of the arguments against the census is that we can get the same data elsewhere from the multitude of service providers that already come in contact with the public.

The problem is that these data collections are administrative. They’re collected for a reason and with limited scope.

Centrelink data is collected to provide a service. Information we provide to the tax office ensures tax compliance. Medicare doesn’t keep information about overseas nationals and people who have never had their birth registered, which is an issue in remote and Indigenous communities.

Australia’s large immigrant population would become a blind spot if we were to rely on the data the government already collects, as many aren’t eligible for certain government services. The data collected by Centrelink, the tax office and Medicare don’t provide sufficient scope. So far the census is the only data source that fits the bill.

Some alternatives

Population registers are a viable alternative to our five-yearly censuses. Finland uses a computerised system to record population data including births, deaths, marriages, migration and so on. The Netherlands, on the other hand, conducts a virtual census by pulling together digital data from a number of different sources.

These registers offer real-time data, but they require ongoing maintenance and verification and often exceed the cost of our census. Ironically, they also need to be checked against a census. And Germany’s experience shows population registers are not always accurate.

Further, major legislation changes would have to go through for Australia to be able to pool data like this. The establishment of a national population register would be costly and demand interdepartmental government coordination.

We could also look to the United States’ method of conducting surveys in between a 10-yearly census. This mixed methodology was suggested by the ABS in 2015 to cut costs.

However, limited financial upside, together with lower quality data, makes it a risky alternative for Australia. Plus we shouldn’t think of the census as an unrecoverable cost. The Office of National Statistics in the United Kingdom estimated the costs of their 2011 census were recovered in just over a year.

The future is data

So how can we improve our census?

Online census completion will save money, improve data quality and reduce data processing time. However, online collection must be balanced to ensure disadvantaged populations aren’t excluded. The end of the census collector hasn’t arrived just yet.

More importantly, we must define contemporary data needs moving into the future. An informed public conversation about migration, employment, families and our changing population is much needed to gain social licence to collect and use relevant data.

Whether the methodology of census continues as is or we introduce an alternative method of data collection, the key going forward is the question of legitimacy. Steps must be taken to justify the need to take a census, and to assuage privacy and security concerns. Without social license we’ll see the failings of the 2016 census play out over and over again.

Australia’s future relies on strong evidence we can agree on. This isn’t solely the domain of researchers. We all have a stake.

The Conversation

Liz Allen, Postdoctoral Fellow, Centre for Aboriginal Economic Policy Research, Australian National University

This article was originally published on The Conversation. Read the original article.

Peter, Paul, Kylie … David! Why we forget family members’ names

Fiona Kumfor, Neuroscience Research Australia and Stephanie Wong, Neuroscience Research Australia

Have you ever been called somebody else’s name? How often has your own mother forgotten your name? Does she ever cycle through the names of each of your siblings, and perhaps even the family pet, before getting to yours?

Don’t worry, it’s probably not because she loves them more than you.

According to researchers at Duke University, misnaming is a common cognitive slip-up. In fact, it seems to occur most frequently between family members and close friends.

The researchers examined survey data from more than 1,700 participants, who were either undergraduate students or older individuals from the community. Regardless of age, those who had been misnamed reported being misnamed by someone they knew well. Likewise, those who had misnamed someone reported doing so to a familiar person.

Misnaming usually occurs within the same semantic category. So, family members are misnamed with another family member’s name and friends are misnamed with another friend’s name.

Names are also more likely to be confused when they share phonetic similarities. For example, misnaming will potentially occur more often if you have children named Dan and Stan.

Notably, the study found misnaming has little to do with physical similarity – which is certainly reassuring if you have ever been called by the dog’s name.

A method to the madness

The finding that we often mix up names that are semantically and phonetically related, rather than at random, gives insights into the way our memories for names are organised in the brain.

The brain tends to group names that are related. They can be related because they belong to a similar semantic category (e.g. family members, school friends, work friends). Or, they can be related because they sound alike (e.g., names starting with “S”).

According to network theories of language, each individual word is linked to other words that share similar conceptual properties. For example, your brother’s name, Paul, might be linked to your other siblings’ names (Kylie, David), as well as other names that start with “P” (Peter).

Because these names share links and are stored in close proximity in the brain, saying one name may also bring to mind other semantically or phonetically similar names. This maximises efficiency, as the brain is able to retrieve closely related information faster.

What happens when these processes break down?

Like knowledge about other types of cognitive abilities, much has been gained from studying people where the ability has been disrupted due to damage in the brain.

Complex pathways in the brain are working to retrieve stored information on faces and names.
from http://www.shutterstock.com

Our research group has a special interest in studying individuals with semantic dementia. Like other types of dementia, semantic dementia is caused by the abnormal accumulation of proteins in the brain. This ultimately leads to cell death and shrinkage of different brain regions.

In semantic dementia, the anterior temporal lobe, a part of the brain situated behind the ear, is most affected.

Individuals with semantic dementia, as the name suggests, show a progressive loss of semantic knowledge (our knowledge about facts, places, things and names). One of the earliest symptoms in semantic dementia is difficulty in naming things.

Naming difficulties in semantic dementia

People with semantic dementia show very specific types of naming errors. For example, they may call a “zebra” a “horse”, or an “animal”. This suggests that as semantic knowledge is lost, our understanding of the world becomes less specific and more generalised.

People with semantic dementia also have difficulty in recognising and naming people. This depends on which side of the brain is more affected.

People with semantic dementia who have more atrophy/shrinkage in the left hemisphere of the brain commonly struggle to recognise names. In contrast, those with greater atrophy in the right hemisphere have more trouble recognising people’s faces.

This kind of research gives us important insights into how names are represented in the brain.

It’s harder than it looks

The ability to recognise someone and call them by their correct name is incredibly complex, even though it feels like second nature to us.

Calling a person by their correct name requires integration of information across both hemispheres of the brain to connect face and name knowledge in mere milliseconds.

Understanding how this process goes wrong – whether through misnaming someone, struggling to say that word on the tip of your tongue, or switching around two words or sounds – gives us important insights into how our brain stores and retrieves the seemingly endless number of names and faces we know.

What does this mean for those of us who frequently misname our family and friends? Based on what we know about how names are organised in the brain, this common cognitive slip-up could be the brain’s way of trying to make our life easier, rather than a sign of something sinister.

So next time you say the wrong name, spare a moment to consider how challenging this ability is and how much work our brains do in order for us to call someone by their name.

The Conversation

Fiona Kumfor, Postdoctoral fellow, Neuroscience Research Australia and Stephanie Wong, PhD Student, Neuroscience Research Australia

This article was originally published on The Conversation. Read the original article.