Published on September 18th, 2012 | by Emma0
What do statistics mean to the national census?
9 August 2011 was Census night in Australia. Conducted every 5 years, the Census of Population and Housing takes a snapshot of our population. It helps us find out where Australians live and what they do, so we can answer such questions as ‘How fast is our population growing?’ ‘What languages do we speak?’ and ‘Where do we need to build schools and improve essential services?’
But surveying 22 million people is not easy. Firstly, it’s expensive, but more importantly, it’s hard to make sure everyone actually participates. One key concern in any national census is that some of the most vulnerable parts of the population – the homeless and recent migrants, for example – may not be able to complete the survey.
So how can we find out about these groups? A few weeks ago I spoke to Professor Christian Robert, a statistician based at the Université Paris-Dauphine, France, about his work in improving the accuracy of census data overseas.
“When we do the Census, we miss the same group of people every time,” he said. “It’s not random at all; it’s very biased.”
Professor Robert says that it’s often more effective to survey a sample of people who are representative of the total population. In France, for example, the national census now surveys only 20 percent of the population.
“With proper statistical techniques, you can interview fewer people and be more precise,” he said, “and the savings are enormous.”
So how do you find out about the people who don’t respond to the census? One way is to use the capture-recapture method, a technique used in ecology where animals are tagged and released back into the wild. This allows researchers to recognise animals that they have seen before.
“The repetition tells you how the population is behaving. If you interview a few people over a few nights, you can get a good idea of the behavior of the whole population very quickly.”
Of course they don’t literally put tags on people – I needed to clarify this one for myself – but this method does raise some interesting questions. In particular, how do you estimate the size of the group you’re working with?
“That’s one of the big challenges in this sort of work,” Professor Robert said. “In fact, the number of things we don’t know is one of the things we don’t know!”
Find out more about the Census of Population and Housing here.[subscribe2]