2077.0 - Census of Population and Housing: Understanding the Increase in Aboriginal and Torres Strait Islander Counts, 2016  
Latest ISSUE Released at 11:30 AM (CANBERRA TIME) 17/10/2018   
   Page tools: Print Print Page Print all pages in this productPrint All RSS Feed RSS Bookmark and Share Search this Product

TECHNICAL NOTE 2 – THE UNDERCOUNT IN THE CENSUS AND THE PES


The Census Post Enumeration Survey (PES) is run shortly after each Census as a way to independently measure how well the Census has counted the Australian population. The PES estimates how many people should have been counted in the Census, how many were missed (undercount) and how many were counted more than once or in error (overcount).

The Census count includes:

    • All responding persons who were counted on at least one Census form, with multiple counts for persons who were included on more than one form
    • Persons who have been imputed into non-responding Census private dwellings during data processing
    • Imputed records for persons who spent Census night at a non-private dwelling (e.g. a hotel or caravan park) but did not complete a Census form there.

All people present in Australia on Census night, with the exception of foreign diplomats and their families, should have been included on a Census form at the place where they stayed. However, due to the size and complexity of the Census, it is inevitable that some people will be missed and some people will be counted more.

Some of the reasons people are missed in the Census include:

  • They were travelling and difficult to contact.
  • They mistakenly thought they had been counted elsewhere.
  • The person completing the form mistakenly thought that certain people (e.g. young babies, the elderly or visitors) should not be included.
  • A Census form or online Census code was not delivered to the address or the dwelling was mistakenly recorded as being empty (unoccupied) on Census night.

Some of the reasons why people may have been counted more than once include:
    • They were included on the Census form at the dwelling where they usually live, even though they stayed and completed a form elsewhere on Census night.
    • They moved during the Census period and completed forms at both their new and old address.
    • They intentionally or unintentionally completed more than one form, e.g. an electronic form and a paper form.

The PES collects information on the characteristics of respondents including Aboriginal and Torres Strait Islander status, which allows the ABS to understand more about people who did not respond to the Census and those who were overcounted.

Using the information provided by the PES, the ABS produces the net undercount for each Census. Net undercount for any category of person is the difference between the actual Census count and the PES population estimate (how many people should have been counted in the Census).

Net undercount is typically presented as a rate. The rate is the net undercount as a percentage of the PES estimate for a given population (i.e. as a percentage of the number of people who should have been counted in the Census). The Census is typically associated with a net undercount. Rates of net undercount or overcount can vary significantly for different population groups depending on factors such as sex, age, Indigenous status, and geographic location, and on whether these characteristics have changed or been misclassified between the Census and PES.

The total net undercount rate for Australia in 2016 was 1.0%, which was lower than in 2011 (1.7%).

Aboriginal and Torres Strait Islander Persons Net Undercount

The 2016 PES estimated that the net undercount rate for Aboriginal and Torres Strait Islander persons was 17.5%, equivalent to 137,750 persons. This is only slightly higher than the 2011 rate (17.2% or 114,188 persons).

Western Australia had the highest net undercount rate in 2016 (23.0%) whilst in New South Wales, the net undercount rate was similar to the national net undercount rate (17.3% compared to 17.5%). In the Australian Capital Territory (ACT), the net undercount rate was the smallest – 12.1% however this is offset by a comparatively large standard error (7.3%).


Aboriginal and Torres Strait Islander Net Undercount Rate and 95 per cent confidence intervals by State/Territory, Empirical Bayes Method(a), 30 June 2016(b)

Graph: Aboriginal and Torres Strait Islander net undercount rate with upper and lower 95 per cent confience intervals by State and Territory, Empirical Bayes method, 30 June 2016

(a) The Empirical Bayes method takes the original PES undercount estimate (the ratio of the PES population estimate to the Census count) for each region and smooths this towards a prediction based on the Census characteristics of the region (specifically the proportion of Aboriginal and Torres Strait Islander persons and the level of Census non-response in the region). Further information on this methodology can be found in Estimates of Aboriginal and Torres Strait Islander Australians, June 2016 (ABS cat. no. 3238.0.55.001).
(b) Net undercount is based on Census counts for a category. In the Census, Indigenous status was set to not stated where the response was blank or where imputed person records were created for non-responding dwellings. Hence components of undercount for Indigenous status do not sum to the Australia total.
Source: Estimates of Aboriginal and Torres Strait Islander Australians, June 2016 (ABS cat. no. 3238.0.55.001).

Persons with Census Category Not Stated

As discussed in Technical Note 1, respondents can choose not to answer the Indigenous Status question in the Census. Where no answer is provided, the Census does not impute for the missing response (which is also the case for imputed persons). While the person (real or imputed) will continue to be counted in broad-level Census counts, they will not be included in the Census counts for Indigenous status.

When calculating net undercount for Aboriginal and Torres Strait Islander persons, these ‘not stated’ responses are treated as people missed from the Census for this population and therefore form part of the undercount estimate.

Net Difference in Classification

Occasionally, the answers obtained for a person in the PES are not consistent with how they completed their Census form. This is expressed as the net difference in classification. There are a number of reasons why responses can differ:

  • A person may have difficulty answering a question for themselves or another household member, either in the Census or in the PES.
  • A person may interpret the question differently in the Census, where the majority of forms are self-enumerated (completed by the respondent unassisted), compared with the PES, which is administered by a trained interviewer.
  • Different people in the household may provide the Census and PES responses.
  • The Census may contain a not stated or imputed response while PES has a valid response.

The following analysis is based on unweighted PES counts. For further information, see Technical Note 1: Differences in Classification Between the PES and Census (ABS cat. no. 2940.0).

When the PES sample was matched to the 2016 Census, 94.6% of records that identified as Aboriginal and/or Torres Strait Islander in the 2016 Census also did so in the PES. A small number of PES records (282 or 4.6% of the PES sample) identified as non-Indigenous in the 2016 Census and as Aboriginal and/or Torres Strait Islander in the PES. There were a small number (381) of records that had identified as Aboriginal and/or Torres Strait Islander in the 2016 Census but non-Indigenous in the PES.

There were 848 PES records that had an unknown Indigenous status in the 2016 Census. In the PES, the majority of these records (803 records) identified as non-Indigenous – not unexpected given that non-Indigenous persons made up the majority of the Australian population in 2016. This suggests that Aboriginal and Torres Strait Islander people are not disproportionately represented in records with an unknown Indigenous status.


Difference in Classification(a) by Indigenous status, 2016

PES response(b)
Aboriginal and/or Torres Strait Islander
Non-Indigenous
Total

no.
%
no.
%
no.
%

Aboriginal and/or Torres Strait Islander
5 741
94.6
381
0.4
6 122
6.0
Non-Indigenous
282
4.6
93 972
98.8
94 254
93.1
Not stated
45
0.7
803
0.8
848
0.8
Matched Persons
6 068
100.0
95 156
100.0
101 224
100.0

(a) Excludes 1,554 Indigenous persons and 8,942 non-Indigenous persons in the PES who were not matched to Census.
(b) The data presented in this table is unweighted PES counts of all responding people in the PES who have been matched to their Census record.
Source: Australian Census of Population and Housing: Details of Overcount and Undercount, Australia, 2016 (ABS cat. no. 2940.0)

Why is the Aboriginal and Torres Strait Islander net undercount important?


The Census and the net undercount (together with other administrative data) are used by the ABS in the compilation of Australia’s official population estimates known as the Estimated Resident Population (ERP) (cat. no. 3238.0.55.001).The ERP is also used as a benchmark for ABS population surveys that take place between Censuses such as the National Aboriginal and Torres Strait Islander Health Survey. The net undercount is therefore an essential component of all statistics collected by the ABS for the Aboriginal and Torres Strait Islander population.

Further information

Further detail on the 2016 undercount and overcount is available in Census of Population and Housing: Details of Overcount and Undercount, Australia, 2016 (ABS cat. no. 2940.0).