Earlier this year, the New Zealand census was held. It is run every five years by Statistics New Zealand, a governmental department. The Census is important, the data collected in it feeds into all sorts of governmental decisions, including spending. For example, over NZD10 billion of funding for health each year is allocated to district health boards on the basis of census statistics. Electoral boundaries (due to be updated before the 2020 election) are redrawn on the basis of the census figures. With such important consequences at stake, it seems as if the NZD121 million price tag is a relatively small price to pay every five years.
However, it seems as if the latest census did not go according to plan. Indeed, the results of the census have been postponed by six months to March 2019 while statisticians go through other government departments (Internal Affairs, Business, Labour, Education, Inland Revenue etc) in a process known as “imputation” to ensure the results of the census are correct. The reason for this delay and extra work is that the number of people who did not respond to the census increased by 82 per cent from 2013. While nearly 95 per cent of the population responded to the last census, the number this time was only 90 per cent. Such figures mean that about 450,000 New Zealanders were not counted on census day. Now it is true that the last few census returns have seen a decline in the number of responses. The last twenty years have seen the response rate decline each census by about 0.9 per cent. But the response rate plummeted by five times that rate between 2013 and 2018.
The reason for this drop in responses is that 2018 was the first “online census”. Everyone was expected to answer the census online by default and could only answer on paper forms if they elected to do so. While I found the online process easy enough, there were a number of people who were less comfortable with the online format. Indeed, this blog post helpfully lists some of the major issues that the census had that may have contributed to the sharp decrease in responses.
Now, things are not irreversible. Canada apparently went from a disastrous census response rate of 68 per cent in 2011 to 98.5 per cent in 2016. So perhaps the next online census in 2023 will work better (and perhaps the overall population will be that much more tech-savvy). But in the meantime there are many concerns with the 2018 results. Ones that may take some time to be allayed.