From 2021 there should be a full online census every 10 years and an increase in the use of government and survey data, according to national statistician Jil Matheson.
Her recommendations, backed by the Office for National Statistics, are supported by the results of a public consultation on the future of the census in England and Wales. The results and Matheson’s statement were both published on 27 March. National Records of Scotland has also announced similar plans to carry out a primarily online census in 2021 and make more use of emerging technologies.
The census has been used to survey all households in England and Wales every 10 years for more than two centuries. The most recent was carried out in 2011. However, in 2010, the government commissioned a review of the system to take into account population changes and the availability of better technologies for gathering and analysing data.
The review, which included a public consultation, statistical research and international comparisons, considered two broad approaches. One was to carry out the 10-year census online. The other was to use administrative data—information gathered routinely by the government, such as health and tax records—and annual surveys of 4 per cent of households to generate annual population statistics.
Matheson recommended a combination of the two. “This would make the best use of all available data to provide the population statistics required and offer a springboard to the greater use of administrative data and annual surveys in the future,” she said.
“I am very pleased with the recommendations,” says Dave Martin, professor of geography at the University of Southampton. “The ONS has clearly heard and acted on the concerns of users expressed through the consultation. The stark dichotomy originally presented—of only administrative data linkage or only an online census—posed major risks to future data.”
The results of last year’s public consultation, which received more than 700 responses, showed that policymakers, researchers and citizens were concerned that the annual surveys would not provide sufficient information on small areas or populations.
Further to this, the ONS said that respondents felt existing methods for using administrative data showed “considerable potential” but were “not yet mature enough” to offer a sufficient replacement for the 10-year census.
More than a third of respondents wanted to see a hybrid approach. The ONS’s report acknowledged that this would increase short-term costs but said it would provide more time to develop the use of administrative data and surveys for the future.
Martin says this is important, because comparisons with other nations show that a move away from a conventional census “takes at least two decades of development and cannot be done in a single leap”. He adds that the extra time protects against “the huge risk of switching off the census model before we can be confident that the alternatives will deliver what is required”.
Cary Cooper, chairman of the Academy of Social Sciences, also welcomes the recommendations. Carrying out an online census will require more work to ensure that people with limited internet access or experience of using technology can respond, he says but this is achievable before 2021.
However, Les Mayhew, a professor of statistics at City University London, says his big concern is that the cost and time involved in moving to an online census will divert resources from improving administrative data methods and surveys. "In a way, I’m not surprised, as the ONS is fairly wedded to the census paradigm," he says. "It’s a step in the right direction, but I think the ONS could be going faster towards administrative data than it has been."
Matheson also said that the increased use of administrative data would depend on public consent, and that legislation on data sharing would be “required to maximise the benefits of this approach”.
John Pullinger, president of the Royal Statistical Society, welcomed the recommendations. He said in a statement that there would be a public meeting at the RSS to discuss the implications and future work required.
The government has the final say on whether the recommendations are to be implemented.