Editorial: NZ's first online census missed too many

Last year's census missed too many people. Photo / Ron Burgin

Editorial

COMMENT: A year ago this week New Zealand held its latest five-yearly census. A year on, we are still waiting for the results. It was the first census to be held largely online and about 10 per cent of the population were missed. Social researchers are calling this a "disaster" and it is.

Census figures are the basis of our knowledge of our society. Every allocation of public money for health, education, social assistance and services of all sorts is based in some way on the data compiled from the information people provide in every locality in the country. The census is by far the main source of research for scholars of social problems and public policy.

Those missing from the returns were probably elderly, poor, transient or ill, some of the people most in need of the programmes the census informs.

Statistics NZ says it will "provide an update in April about the release of the 2018 Census data". Its website admits, "It always takes time for us to analyse and produce the dataset but this time we are taking longer than usual because the overall individual response was lower than we had aimed for."

It spent much of the past year desperately trying to backfill the gaps in its census night data "using other government data about real people, such as birth, tax, health and education records".

To do this, it explains, it is using methods based on work it has been doing for a "census transformation programme" which is investigating the feasibility of a future census based on administrative data.

Meanwhile, the Government is preparing a Budget based on many more measures of well-being than the GDP of economic activity. Among those will be the reduction of child poverty, mainly measured as a proportion of the mean household income. The previous Government was staking just as much policy on sound social data. Bill English, as finance minister, built his "social investment" approach on a belief in the precision of data available to governments these days.

The conduct of last year's census is being independently reviewed but Statistics NZ has already suggested a few reasons for the poor response. It says many households that did not want to do the forms online received them by post after census day and probably felt less inclined to respond.

It also admits that though it made close to a million follow-up visits, they did not make direct contact with enough households as close to census day as possible.

And in those households that did choose to do the census online, not all returned a form for each individual living in the house.

It is hard to avoid the conclusion there is simply no substitute for a person coming to the door of every home, explaining what needs to be done on the night of the count and returning in person. Not only does it help the confused, it puts more pressure on the careless. It is easy to forget a message on a computer, harder to ignore a bundle of papers in the house that a person is coming back to collect.

Sometimes we trust technology too much.