AGE AND SEX

Previous Censuses (both in Australia and overseas) have observed that young adults aged 20-29 years are more likely to be missed, while older adults are more likely to be counted or overcounted in a Census and young children are often mistakenly omitted from Census forms. Males are also traditionally more likely to be missed than females.

These trends hold true for the 2016 Census.

The highest net undercount rate in 2016 was for the 0-4 year olds (5.1%), followed by the 20-24 and 25-29 year age groups (5.0% and 4.9%, respectively). This is a change from 2011, where the young adults had the highest net undercount rates. In contrast, the older age groups all showed low net undercount rates, with the lowest rate for people aged 70-74 years (-3.9%; i.e. a net overcount). Males were also more likely to be missed in the Census compared with females, with net undercount rates of 1.5% and 0.4%, respectively.



For males, those aged 25-29 years had the highest net undercount rate of all age groups (6.6%) followed by 20-24 year olds (6.1%). These age groups also had the highest net undercount rates in 2011.

For females, the 0-4 year olds had the highest net undercount rate (4.8%), followed by the 20-24 year olds (3.8%) and the 25-29 year olds (3.3%). The net undercount rate for females aged 60-64 years was the lowest of all age groups (-4.4%), although a net overcount can be seen for all female age groups 40 years and over.

Children showed a higher net undercount rate in 2016. Compared with 2011, the net undercount rates were higher for 0-4 year olds (from 1.2% to 5.1%), for 5-9 year olds (from 1.5% to 3.4%), and for 10-14 year olds (from 0.4% to 1.9%).

Care should be taken when comparing the 2016 and 2011 net undercount estimates for age, due to an inconsistency in the calculation of age for estimation purposes. The 2016 PES collected the respondent’s age as at the time of the PES interview and this was (correctly) backdated to 9 August 2016 (i.e. their age on Census night) using date of birth, where provided. In 2011, however, age was not backdated to Census night for estimation purposes. This would have produced a small downward bias to the 2011 PES population estimate (and hence to the net undercount rate) for 0-4 year olds and a small upward bias for the 85 and over age group.