3.5 Aboriginal and Torres Strait Islander peoples

Report on the quality of 2021 Census data: Statistical Independent Assurance Panel to the Australian Statistician

An independent view of the quality of statistical outputs from the 2021 Census of Population and Housing

Released
28/06/2022

3.5.1 Population counts

Decorative banner

3.5.1.1 Australia and states and territories

In the 2021 Census, there were 812,728 people who reported Aboriginal and Torres Strait Islander origins, an increase of 25.2% from 2016, compared to an increase of 18.4% from 2011 to 2016 (see Table 3.5.1).

  • The number of people reporting as being of Aboriginal origin increased from 590,056 in 2016 to 742,882 in 2021, representing an increase of 25.9% from 2016 to 2021, and up from 19% between 2011 and 2016.
  • The count of people of Torres Strait Islander origin was 33,765 in 2021 up from 32,345 in 2016, an increase of 4.4%.
  • The number of people reporting that they were of both Aboriginal and Torres Strait Islander origin grew by 34.8% between 2016 and 2021, from 26,767 to 36,083.

Similarly, large increases in the counts of Aboriginal and Torres Strait Islander peoples have been observed in previous censuses. These increases have been explained by factors of demographic change as well as other factors including changes in census coverage and census response rates and a changing propensity to identify as being of Aboriginal and Torres Islander origins. The question on Indigenous status in the Census is reliant on self-identification and people may change their identification between censuses.

In the 2021 Census, there were 172,387 people who responded to the Census but did not provide a response to the Indigenous status question resulting in a non-response rate of 0.7% for this data item. This is lower than the rates for 2016 and 2011 which were 1.0% and 1.4%, respectively.

In every census, there are people who do not complete a census form but are in a dwelling that was considered occupied on Census night. In such cases, ABS imputes people into these dwellings. Imputation is the process that attempts to fill in basic information for the people who didn’t respond to the Census. Currently, this information is limited to a person’s age, sex, usual residence and marital status, but Indigenous status is not imputed. In 2021, there were 1,061,722 people who did not respond to the Census but whose basic demographic information was imputed. For these people their Indigenous status is unknown.

In 2021, as in previous censuses, New South Wales and Queensland had the largest counts of usual residents reporting they were of Aboriginal and Torres Strait Islander origins (with counts of 278,043 persons and 237,303 persons, respectively) with increases of 28.6% and 27.3%, respectively between 2016 and 2021. The largest proportional increases were in the Australian Capital Territory (37.5% increase or an increase of 2,441 persons) and Victoria (37.4% or an increase of 17,858 persons) while the smallest was in the Northern Territory (4.9% or an increase of 2,867 persons). Every five years following the Census, the ABS undertakes analysis of the changes in the counts of Aboriginal and Torres Strait Islander peoples. The ABS’s analysis of the 2021 counts will be released in 2023.

Table 3.5.1 Census counts of Aboriginal and Torres Strait Islander identification by state/territory of usual residence

2021

Australia (a)

NSW

Vic.

Qld

SA

WA

Tas.

NT

ACT

Aboriginal and Torres Strait Islander peoples subtotal

812,728

278,043

65,646

237,303

42,562

88,693

30,186

61,115

8,949

Aboriginal

742,882

267,067

61,865

193,405

40,592

85,004

27,738

58,566

8,425

Torres Strait Islander

33,765

5,127

2,083

21,772

994

1,625

1,225

692

230

Both Aboriginal and Torres Strait Islander

36,083

5,844

1,692

22,122

967

2,068

1,229

1,852

300

Australians of other descent

23,375,949

7,404,499

6,148,188

4,635,042

1,669,314

2,431,204

501,521

152,705

429,520

Not stated (responding person) (b)

172,387

55,632

42,081

37,050

12,395

16,072

5,197

2,064

1,850

Not Stated (imputed person) (c)

1,061,722

333,981

247,582

246,742

57,251

124,052

20,659

16,714

14,182

Total (d)

25,422,788

8,072,163

6,503,491

5,156,138

1,781,516

2,660,026

557,571

232,605

454,499

 2016

Australia (a)

NSW

Vic.

Qld

SA

WA

Tas.

NT

ACT

Aboriginal and Torres Strait Islander peoples subtotal

649,171

216,176

47,788

186,482

34,184

75,978

23,572

58,248

6,508

Aboriginal

590,056

207,256

44,592

148,943

32,616

72,924

21,570

55,805

6,140

Torres Strait Islander

32,345

4,839

2,024

21,053

938

1,434

1,119

744

183

Both Aboriginal and Torres Strait Islander

26,767

4,080

1,171

16,493

629

1,628

889

1,699

183

Australians of other descent

21,341,231

6,826,286

5,532,275

4,211,020

1,557,001

2,237,541

455,137

147,327

370,748

Not stated (responding person) (b)

227,589

71,080

61,114

45,337

17,916

20,228

6,845

2,500

2,492

Not Stated (imputed person) (c)

1,183,899

366,686

285,449

260,345

67,554

140,663

24,411

20,760

17,656

Total (d)

23,401,892

7,480,228

5,926,624

4,703,193

1,676,653

2,474,410

509,965

228,833

397,397

 2011

Australia (a)

NSW

Vic.

Qld

SA

WA

Tas.

NT

ACT

Aboriginal and Torres Strait Islander peoples subtotal

548,368

172,625

37,992

155,826

30,432

69,664

19,625

56,779

5,184

Aboriginal

495,754

164,611

34,947

122,897

28,831

67,063

17,742

54,571

4,859

Torres Strait Islander

31,407

4,770

2,158

20,095

1,039

1,309

1,171

675

190

Both Aboriginal and Torres Strait Islander

21,205

3,244

881

12,831

559

1,296

713

1,532

138

Australians of other descent

19,900,765

6,402,111

5,069,155

3,952,704

1,503,204

2,038,783

456,343

137,773

338,029

Not stated (responding person) (b)

290,022

94,836

77,508

55,467

21,408

27,485

7,446

2,696

3,120

Not Stated (imputed person) (c)

768,561

248,084

169,388

168,737

41,524

103,236

11,932

14,696

10,882

Total (d)

21,507,719

6,917,656

5,354,039

4,332,737

1,596,569

2,239,171

495,351

211,943

357,218

(a) Includes Other Territories.

(b) Person was included on a Census form but did not answer the Indigenous status question.

(c) Person was not included on a Census form.

(d) Excludes overseas visitors.

3.5.1.2 Age distribution

Figure 3.5.1 shows the age distribution of people who reported they were of Aboriginal and Torres Strait Islander origins in the 2021 Census and in the 2016 Census. It shows that much of the increase in the counts is in the childhood years (0-14 years). In 2016, there were 73,625 children aged 0-4 years of Aboriginal and Torres Strait Islander origins while in 2021 the number of 0-4 year olds had increased to 85,941. Similarly, the count of 5-9 year olds increased from 75,755 to 89,137 while the largest increase was observed for 10-14 year olds where the count increased from 71,378 in 2016 to 90,473 in 2021.

Note: Includes Other Territories.

Figure 3.5.2 shows the same information presented in a different way. It shows people in both the 2016 and 2021 Censuses at the age they were in 2016. In 2016, there were 73,265 Aboriginal and Torres Strait children aged 0-4 years. Five years later in the 2021 Census, they would be between 5-9 years of age but this group had now grown to 89,137. The difference in the two lines reflects changes in coverage and response in the Census, and changes in identification. There are increases in the counts across almost all age groups.

Notes: Includes Other Territories.

2021 age group has been adjusted (i.e. five years subtracted) to indicate age group at 2016.

3.5.1.3 Counts for capital city and rest of state

Below state level, the results have been analysed for greater capital cities and rest of state/territory. In 2021, the largest counts of Aboriginal and Torres Strait Islander peoples were in the rest of New South Wales (a count of 185,873 persons, increasing by 28% since 2016) and the rest of Queensland (159,179 persons increasing by 21% since 2016) (see Figures 3.5.3 and 3.5.4). These areas are outside the greater capital city boundaries in each of their respective states.

The counts have increased across all capital cities and rest of state areas with the exception of the rest of the Northern Territory where the counts have remained flat over the last three Censuses (with a small decrease in 2021).

Notes: Capital city counts for the Australian Capital Territory are for the whole of the Australian Capital Territory.

Excludes Other Territories and Migratory, Offshore and Shipping categories.

Notes: Capital city counts for the Australian Capital Territory are for the whole of the Australian Capital Territory.

Excludes Other Territories and Migratory, Offshore and Shipping categories.

3.5.2 Mode of response

Decorative banner

Aboriginal and Torres Strait Islander peoples have increasingly taken up the online option when completing their Census with almost two thirds of responses in 2021 being online. There has been a corresponding decline in the proportion of paper responses. A large number of responses from Aboriginal and Torres Strait Islander peoples are also received via Interviewer household forms which are used in particular discrete communities (communities comprising mainly Aboriginal and Torres Strait Islander peoples and which are managed on a community basis) and where the use of the standard self-enumeration form may be impractical. Almost 9% of Aboriginal and Torres Strait Islander peoples were enumerated on this type of form in 2021, down from 14.2% in 2011 (see Figure 3.5.5).

As has been noted elsewhere, the online Census form delivers data of a high quality due to its ease of use and electronic assistance within the form, so an increase in its use can be expected to deliver associated increases in data quality.

Note: Includes Other Territories.

3.5.3 Ancestry

Decorative banner

Ancestry is part of the cultural diversity suite within the Census and provides information about countries or cultures to which people feel they have an affiliation. It is subjective by its very nature and a person’s cultural identity may not be well known amongst unrelated household members. The question allows provision of up to two ancestries, which may limit people with more complex ethnic backgrounds.

The number of Aboriginal and Torres Strait Islander peoples reporting that they had Australian Aboriginal ancestry or Torres Strait Islander ancestry in their first Ancestry response increased dramatically in the 2021 Census (see Figure 3.5.6). The number reporting Aboriginal ancestry increased from 90,673 in 2011 and 89,984 in 2016 to 694,644 in 2021. Similarly, there was a large increase in the number of people reporting they were of Torres Strait Islander ancestry, up from 6,709 in 2011 and 7,149 in 2016 to more than 28,000 in 2021. There have been corresponding decreases in the number of people reporting Other Australian ancestry (down from 298,283 in 2016 to 27,497 in 2021), and in reporting Other ancestries (down from 220,982 in 2016 to 34,630 in 2021). The non-response rates for Ancestry as reported by Aboriginal and Torres Strait Islander peoples have declined in recent censuses from 7.5% in 2011, 5.1% in 2016 and 3.4% in 2021.

The very large increases observed in 2021 reflect a change to the layout of the Ancestry question in the 2021 Census in both the paper and online form. This change was made following feedback from the Round Table on Aboriginal and Torres Strait Islander Statistics that revising response categories for topics such as Ancestry would improve engagement and participation without adding questions to the Census.[1] Firstly, 'Aboriginal' and 'Torres Strait Islander' were added as specific response categories. In addition, on the online form, these categories appeared at the top of the pick list if a respondent answered the Indigenous status question with a response of Aboriginal, Torres Strait islander or both. These changes would have the effect of acting as an overt prompt for respondents.

While these increases were observed on the online and paper form responses in 2021, this was not evident in responses via the Interviewer household form. This is because those more overt prompts have been in use in this type of form in past censuses.

 

[1] Australian Bureau of Statistics (2018). Census of Population and Housing: Topic Directions, 2021. Retrieved from https://www.abs.gov.au/ausstats/abs@.nsf/Lookup/2007.0.55.001main+features12021.

(a) Other Australian includes Australian, Australian South Sea Islander, Norfolk Islander and Australian Peoples, not further defined.

Note: Includes Other Territories.

3.5.4 Australian Defence Force service

Decorative banner

This is one of two new topics introduced in the 2021 Census. The counts of people of Aboriginal and Torres Strait Islander origins who reported that they were currently serving in the Australian Defence Force in either a regular or reserve capacity varies between the Census and the Department of Defence (see Figure 3.5.7). Differences between the two data sources may be due to different levels of Aboriginal and Torres Strait Islander self-identification reported in the Census and in the Department of Defence data. The Panel has limited capacity to assess the quality of Census data on previous Australian Defence Force service, as there is neither a definitive independent source nor historical Census data. The Census counted 11,610 Aboriginal and Torres Strait Islander peoples who had previously served in the Regular or Reserves Service.

Note: Census data includes Other Territories.

3.5.5 Long-term health conditions

Decorative banner

Long-term health conditions is the second new topic included in the 2021 Census. The question asked people whether they have ever been told by a doctor or nurse that they have any selected long-term health conditions. With a non-response rate of 8.1%, a relatively high number of Aboriginal and Torres Strait Islander peoples did not answer this question, most likely due to sensitivity issues and a desire not to disclose a condition to others in the household or a possible lack of knowledge of others’ health conditions in the household.

As observed in Figure 3.5.8, there are differences in the levels of long-term health conditions reported in the Census and in other surveys and this is due to differences in scope and methodology. Such differences are apparent when comparing the Census and 2018-19 National Aboriginal and Torres Strait Islander Health Survey. While the levels of Aboriginal and Torres Strait Islander peoples reporting long-term health conditions are different between the two data sources, the patterns of reporting are similar, with the conditions with the highest levels reported being mental health conditions and asthma in both surveys. The proportion reporting mental health conditions is much lower in the Census than in the National Aboriginal and Torres Strait Islander Health Survey, and this pattern was observed in the total population (see Section 3.7).

It is likely that more detailed and targeted surveys such as the National Aboriginal and Torres Strait Islander Health Survey will provide better estimates of the level of reporting, noting that this survey also includes self-diagnosed conditions. However, the utility of the Census is in the ability to provide cross classifications by a range of characteristics for small geographic areas. Further investigation will be conducted by the ABS into the benefits and limitations of long-term health conditions information collected in the Census.

(a) All National Aboriginal and Torres Strait Islander Health Survey (NATSIHS) data used in this figure includes self-diagnosed conditions.

Note: Includes Other Territories.

Figure 3.5.9 indicates a high level of non-response to this question for people under 15 years on the Interviewer household form. The placement of the question on this form may have influenced non-response.

Note: Includes Other Territories.

3.5.6 Net undercount

Decorative banner

The Post Enumeration Survey estimated that 983,257 Aboriginal and Torres Strait Islander peoples should have been counted in the Census, compared to the 812,505 who were actually counted. This is equivalent to a net undercount of 170,752 people, or a rate of 17.4% of the Post Enumeration Survey population estimate. This is similar to the level in the previous two Censuses (see Table 3.5.2).

Table 3.5.2 Net undercount, Post Enumeration Survey population estimates and Census counts, by Indigenous status

 

PES population estimate

Census count (a)(b)

Net undercount (c)

Net undercount rate

 

no.

Standard error

no.

no.

Standard error

%

Standard error

2021

Aboriginal and Torres Strait Islander peoples

983,257

21,590

812,505

170,752

21,590

17.4

1.8

Australians of other descent

24,624,765

41,448

23,371,978

1,252,787

41,448

5.1

            0.2

Not stated

 -

 -

1,233,495

 -

 -

 -

 -

2016

Aboriginal and Torres Strait Islander peoples

786,689

19,776

648,939

137,750

19,776

17.5

2.1

Australians of other descent

22,837,014

46,483

21,337,326

1,499,688

46,483

6.6

0.2

Not stated

 -

 -

1,411,031

 -

 -

 -

 -

(a) Includes imputed persons in non-responding dwellings. These were all given an Indigenous status of 'not stated'.

(b) Refers to Census counts which correspond to the scope of the Post Enumeration Survey (PES) and may differ slightly from aggregate counts in other Census products.

(c) Net undercount is based on Census counts for a category. In the Census, Indigenous status is 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.

 – Nil or rounded to zero.

Note: Excludes Other Territories and overseas visitors.

In collaboration with the ABS, the Panel has reviewed the methodology used to obtain the estimates of the undercount. The estimates are sensitive to:

  • accurate matching of persons between the Post Enumeration Survey and the Census; and
  • correlated response bias (see Appendix C).

With respect to accurate matching, the ABS puts considerable effort into matching Aboriginal and Torres Strait Islander peoples given it is more difficult because of less consistency in the use of names, for example. Regardless of the efforts made, it is difficult to determine whether there is a match or non-match between the Post Enumeration Survey and the Census. To the extent that there are remaining linking errors, there will be an upward bias in the estimate of the undercount.

With respect to correlated response bias the adjustment does not remove all the biases and there are reasons why this may be relatively higher for Aboriginal and Torres Strait Islander peoples. This will lead to a downward bias in the estimate of the undercount.

Sensitivity analysis undertaken by the ABS suggests that the impact of inaccurate matching is larger than correlated response bias so there may be a net upward bias in the estimate of the undercount. However, this would also apply to the previous two Censuses. This is confirmed by an independent estimate of the undercount using the Multi-Agency Data Integration Project (MADIP) data which suggests it might be about 15%. Even if you could make an adjustment for that bias, the undercount for Aboriginal and Torres Strait Islander peoples is clearly higher than for Australians of other descent.

As noted in earlier sections, the 2021 Census had a higher response rate than the 2016 Census in the capital cities and major urban areas. However, the opposite occurred in the more remote areas particularly in Western Australia where there were recruitment and access difficulties. The proportion of Aboriginal and Torres Strait Islander peoples is relatively high in these areas so they may not have benefitted to the same extent from the general increase in response rates.

Looking at the reasons for the undercount of Aboriginal and Torres Strait Islander peoples showed some differences between 2016 and 2021. Unpublished data indicates that there was an increase in the number of persons who, in the Post Enumeration Survey, said they were elsewhere on Census night but could not be found. The increased lag between the timing of the Census and the Post Enumeration Survey may be a partial explanation. It also indicates an increase in the number of persons who were in dwellings missed by the Census.

Another change has been the relative importance of the contact and non-contact sectors between the 2011 and the 2016 and 2021 Censuses. In 2011, two-thirds of the undercount was due to the contact sector. In 2016, the non-contact sector dominated (55%) possibly because of issues with occupancy determination and in 2021 it was about 50%. If data from MADIP could be used as an alternative source for determining Aboriginal and Torres Strait Islander origins, the undercount might be reduced considerably and warrants further investigation for future censuses.

Table 3.5.3 Net undercount rate in the contact and non-contact sector for Aboriginal and Torres Strait Islander peoples

 

Contact sector (CS)

Non-contact sector (NCS)

Total

 

Gross undercount

Gross overcount

Net difference in classification

Census 'not stated'

CS - Net undercount

NCS -Net undercount

Total - Net undercount

 

%

%

%

%

%

%

%

2011

11.0

1.4

0.8

1.1

11.5

5.7

17.2

2016

11.6

2.1

-2.5

0.8

7.8

9.7

17.5

2021

11.0

1.5

-1.3

0.7

9.0

8.4

17.4

 

3.5.7 Implications for data quality

Decorative banner

The growth in the counts of Aboriginal and Torres Islander peoples seen in previous censuses has continued. Changes in the counts have an impact on other statistics, such as population measures and some performance indicators used for government reporting. As after previous censuses, ABS will further analyse the increase in the counts to understand the changes observed in the 2021 Census. The Post Enumeration Survey estimates that net undercount for Aboriginal and Torres Strait Islander peoples remains high (17.4%) despite considerable investment in and efforts to improve enumeration.

A large increase in online form response by Aboriginal and Torres Strait Islander peoples is expected to have associated increases in data quality due to the reasons mentioned elsewhere (see Section 3.3.4). A change to the layout of the Ancestry question for 2021 has resulted in increased numbers of people reporting Aboriginal and Torres Strait Islander ancestry.