4705.0 - Population Distribution, Aboriginal and Torres Strait Islander Australians, 2006  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 15/08/2007  Final
   Page tools: Print Print Page Print all pages in this productPrint All

INTERPRETING THE DATA


INTRODUCTION

In the 2006 Census, as in previous Censuses, there was an increased count of Aboriginal and Torres Strait Islander Australians. In addition, the Indigenous status of a significant number of Census records was unknown, and some Indigenous people were missed. When producing population estimates by Indigenous status, adjustments are made to take account of these factors. Some care should be taken when analysing Census counts both because Indigenous status is unknown for some records, and because Indigenous people are more likely than other people to not be counted on Census Night.


This chapter describes key features of the 2006 Indigenous Enumeration Strategy (IES) designed to improve both the count of Indigenous people and their identification as being of Indigenous origin; and then describes three different measures of the population: place of enumeration Census counts; usual residence Census counts; and estimated resident population. It also discusses the size and distribution of Census records with unknown Indigenous status to assist readers to interpret and apply the information contained in this publication. Net undercount and the Post Enumeration Survey (PES) are also briefly discussed.


Additional information on data quality is available in Appendix 3: Quality of Indigenous Status Data in the 2006 Census. Further information will be included in forthcoming ABS publications from the 2006 Census, including Population Characteristics, Aboriginal and Torres Strait Islander Australians, 2006 (cat. no. 4713.0) and Experimental Estimates and Projections, Aboriginal and Torres Strait Islander Australians, 1996 to 2016 (cat. no. 3238.0).



COUNTING THE ABORIGINAL AND TORRES STRAIT ISLANDER POPULATION

Indigenous Enumeration Strategy

The ABS invests considerable resources to ensure a high quality count of all people in Australia during the Census enumeration period. Counting the Indigenous population presents a number of challenges. To maximise the quality of the count, the IES has been designed as a supplementary strategy to regular Census collection procedures.


Since the 1976 Census, the ABS has used an evolving set of procedures tailored to the enumeration of Indigenous people. The 2006 IES built on this experience and had two main components:

  • Alternative collection procedures to overcome potential barriers (e.g. language) to the effective counting of Aboriginal and Torres Strait Islander peoples.
  • Census awareness activities aimed at encouraging Aboriginal and Torres Strait Islander peoples to both participate in the Census and be identified as Indigenous.

New approaches introduced in the 2006 IES to improve the Indigenous Census count include:
  • Using Indigenous Engagement Managers, who have an ongoing role in the ABS's Indigenous Community Engagement Strategy, as Census State Indigenous Managers or assistants to State Indigenous Managers. The 2006 Census was the first Census conducted since the implementation of the ABS's Indigenous Community Engagement Strategy.
  • Developing more detailed strategies for the enumeration of Indigenous people in urban, regional and remote areas.
  • Easing the administrative burden for interviewers in discrete communities by using a streamlined Interviewer Household Form in place of the separate Special Indigenous Household Form and Personal Form used in 2001.
  • Improving procedures for validating Interviewer Household Forms in the field.
  • Developing specific strategies at the Census Data Processing Centre to improve the processing of Census forms completed by, or on behalf of, Indigenous people.

Within the IES framework, different approaches were used to enumerate discrete Indigenous communities and Indigenous households within the wider community. Indigenous people were generally enumerated using standard procedures and forms in urban and regional areas, although specially trained Census collectors were available to assist if required.


In the majority of discrete Indigenous communities in remote areas, an interview form (rather than a self-completion form) was used and local Indigenous people were employed to interview households. Discrete communities were enumerated over a period of weeks around the Census date. This is known as a 'rolling enumeration' and as a result some communities were enumerated at different times.


In 2006, 75,351 or 17% of all Indigenous people both counted in the Census and identified as Indigenous were enumerated using the Interviewer Household Form. The proportion varied across jurisdictions: Northern Territory (66%), Western Australia (20%), Queensland (17%), South Australia (12%) and New South Wales (3%).



CENSUS COUNTS

There are two types of Census counts:

  • Census counts based on where people are when counted (place of enumeration).
  • Census counts based on where people usually live (place of usual residence).

Both types of counts can be used for analysis. Usual residence counts are generally preferred for geographic analysis since, for any given location, they include people who were away from home at the time of the Census and exclude temporary visitors. While place of enumeration and usual residence counts are shown in tables 2, 5 and 44, usual residence counts are generally shown elsewhere in this publication.


Nationally, 30,853 Indigenous people (7%) were identified and counted at a location away from their usual home at the time of the 2006 Census. While, at the state/territory level, between 4% and 8% of Indigenous people were counted away from home at the time of the Census, there was little difference between the total place of enumeration and usual residence Census counts for the states and territories because those away from home were usually counted and identified elsewhere within the same state or territory.



ESTIMATED RESIDENT POPULATION

When producing estimates of the resident Indigenous population (ERP) the ABS adjusts the Census count to take account of unknown Indigenous status and undercount (together with other factors such as the difference in reference periods). Overall, the preliminary Indigenous estimated resident population at 30 June 2006 was 517,200, 14% greater than the 2006 Census count of people identified as being of Indigenous origin. While Indigenous population estimates for each state and territory were greater than the Census counts, the largest adjustments for ERP purposes were made for Western Australia (33%) and the Northern Territory (24%).

Comparison of Usual Residence Census Count and ERP, Indigenous Population - 2006

Usual residence
Census count,
8 August 2006
Indigenous ERP,
30 June 2006(a)
Difference
between ERP
and Census
count
Percentage
increase
from Census
count
State/Territory
no.
no.
no.
%

New South Wales
138 507
148 178
9 670
7.0
Victoria
30 143
30 839
696
2.3
Queensland
127 580
146 429
18 849
14.8
South Australia
25 556
26 044
488
1.9
Western Australia
58 710
77 928
19 218
32.7
Tasmania
16 768
16 900
132
0.8
Northern Territory
53 661
66 582
12 921
24.1
Australian Capital Territory
3 875
4 043
169
4.4
Australia(b)
455 028
517 174
62 146
13.7

(a) Preliminary.
(b) Includes Other Territories. See Glossary.
Tables 1 and 2.


For further information see Appendix 2: Indigenous Estimated Resident Population - Method of Calculation.



UNKNOWN INDIGENOUS STATUS

Indigenous status is not recorded for all people counted in the Census. While some of the people with unknown Indigenous status will be of Indigenous origin and some will be non-Indigenous, the exact proportions cannot be determined from the Census data.


There are two situations which result in Indigenous status being unknown:

  • Incomplete Census forms, with no answer to the Indigenous status question, are returned to the ABS (item non-response).
  • The ABS cannot obtain Census forms from people or dwellings and as a consequence imputes basic demographic information (but not Indigenous status) for these records.

While the non-response rate to the Indigenous status question has remained relatively constant at around 1.7% in the 1996 and 2006 Censuses and 2.0% in the 2001 Census, the proportion of imputed records has increased from 1.3% of the total Census count in 1996 to 4.1% in 2006.



Overall distribution

The number of records with unknown Indigenous status in 2006 was 1,133,446 (5.7% of the total count) compared with 767,757 (4.1%) in 2001. The difference of 365,689 records represented a 48% increase over the 2001 level.

Indigenous status unknown(a), Usual residence Census counts - 1996-2006

1996
2001
2006
Intercensal change
1996-2001
2001-2006
State/Territory
no.
%
no.
%
no.
%
%
%

New South Wales
178 074
3.0
290 192
4.6
391 274
6.0
63.0
34.8
Victoria
131 841
3.0
191 884
4.1
266 028
5.4
45.5
38.6
Queensland
101 353
3.1
131 425
3.7
224 909
5.8
29.7
71.1
South Australia
36 878
2.6
45 031
3.1
69 318
4.6
22.1
53.9
Western Australia
46 902
2.7
70 638
3.9
127 328
6.5
50.6
80.3
Tasmania
15 241
3.3
16 390
3.6
22 903
4.8
7.5
39.7
Northern Territory
8 655
4.9
11 544
6.1
16 505
8.6
33.4
43.0
Australian Capital Territory
6 400
2.1
10 538
3.4
15 025
4.6
64.7
42.6
Australia(b)
525 403
3.0
767 757
4.1
1 133 446
5.7
46.1
47.6

(a) Percentages based on total Census count.
(b) Includes Other Territories. See Glossary.
1996, 2001 and 2006 Census data.



Areas most affected by unknown Indigenous status

In 2006, the Northern Territory had the highest rate of unknown Indigenous status (9%) followed by Western Australia (7%) and New South Wales (6%).


Among Indigenous Regions, South Hedland (16%) and Broome (14%) in Western Australia had the highest incidence of unknown Indigenous status (table 4).


Some Indigenous Locations in mining and agricultural areas, which also had a large percentage of non-private dwellings, recorded a high incidence of unknown Indigenous status. In non-private dwellings such as staff quarters, a nominated person in the institution completed a summary form that listed all residents and personal forms were distributed to individual residents. It is possible that in some cases personal forms were not returned or not completed properly for all staff listed on the summary form. Examples of Indigenous Locations affected include: Laverton (S) - Rem, Wyndham-East Kimberley (S) - Rem, East Pilbara - Rem and Meekatharra (S) - Rem in Western Australia. In these Indigenous Locations, people in non-private dwellings accounted for 50% to 100% of the records with unknown Indigenous status.


In other remote mining and agricultural areas a high incidence of unknown Indigenous status may reflect difficulties in contacting people due to distance and irregular working patterns. Areas affected include Indigenous Locations in:

  • Mining areas such as the Pilbara region and the central west of Western Australia.
  • Centres such as Wyndham in the Kimberleys and Tennant Creek in the Northern Territory.
  • Very large remote areas, such as those near Petermann on the Northern Territory/South Australia border, the Gulf country in northwest Queensland and the far west of Bourke in New South Wales.

In these areas, collectors may have experienced difficulties such as determining whether or not a dwelling was occupied, delivering and collecting the Census form, and ensuring all questions were completed.


Discrete Indigenous communities, where the Interviewer Household Form was generally used, did not tend to contribute to unknown Indigenous status. The Indigenous Location of Horn Island in the Torres Strait, in which 11% of records had unknown Indigenous status, was an exception.


Indigenous Locations in major population centres with high rates of unknown Indigenous status include: Sydney (C) - Rem (24%) and Redfern (19%) in New South Wales; Melbourne (C) (14%) and Port Phillip (C) (12%) in Victoria; Cairns (C) - City (18%) and Cairns (C) - Central Suburbs (15%) in Queensland; and Perth (C)/Vincent (T)(13%) in Western Australia; and City Inner (21%) and Gray (18%) in the Northern Territory.


Records from non-private dwellings, including tourist accommodation, contributed significantly to the level of unknown Indigenous status in inner areas of some major population centres. Examples include: City/Inner (19%) and Larrakeyah/The Gardens (16%) in Darwin; Cairns (C) - City (15%); and Melbourne (C) (10%).



Imputed records

Imputation is a statistical process used to predict a value where no response has been provided. While Census records are imputed when the ABS does not obtain a Census form for residents of either private or non-private dwellings, most imputation is for non-response in private dwellings.


Imputation is undertaken as a result of the following situations:

  • A collector believes a dwelling to be occupied but is unable to contact the householders.
  • A collector makes arrangements for the householder to either mail the Census form or complete the eCensus, but the form does not arrive at the ABS.
  • The householder refuses to participate in the Census.
  • The collector fails to collect the form.
  • People in non-private dwellings do not receive the assistance they need (e.g. due to age or disability) to complete the Census form.
  • People in non-private dwellings are listed on the dwelling summary form as being present on Census Night but personal Census forms are not completed for all of them.

The rate of imputation was 1.3% (or 234,552 records) of the total Census count in 1996, 2.1% (or 402,189 records) in 2001 and 4.1% (or 807,062 records) in 2006. Increases were observed in all states and territories. Furthermore, the proportion of Census records with unknown Indigenous status as a result of imputation has increased from 45% in 1996 to 71% in 2006.


At the state/territory level, the highest increases in the number of imputed records over the period 1996-2001 were in New South Wales and Victoria, while over the period 2001-2006 the highest increases were in Queensland, Western Australia, South Australia and Tasmania.


The reasons for the increase in imputed records are currently being investigated. They include changes in field procedures and difficulties in contacting householders.

Indigenous status unknown, Imputed records(a)(b) - 1996-2006

1996
2001
2006
Intercensal change
1996-2001
2001-2006
State/Territory
no.
%
no.
%
no.
%
%
%

New South Wales
86 226
1.4
170 283
2.7
280 805
4.3
97.5
64.9
Victoria
54 766
1.2
98 618
2.1
179 743
3.6
80.1
82.3
Queensland
39 826
1.2
59 255
1.7
165 161
4.2
48.8
178.7
South Australia
14 633
1.0
18 135
1.2
44 538
2.9
23.9
145.6
Western Australia
23 906
1.4
36 115
2.0
97 737
5.0
51.1
170.6
Tasmania
5 050
1.1
6 022
1.3
13 523
2.8
19.2
124.6
Northern Territory
6 151
3.5
7 109
3.8
14 059
7.3
15.6
97.8
Australian Capital Territory
3 949
1.3
6 637
2.1
11 399
3.5
68.1
71.7
Australia(c)
234 552
1.3
402 189
2.1
807 062
4.1
71.5
100.7

(a) Usual residence Census counts.
(b) Percentages based on total Census count.
(c) Includes Other Territories. See Glossary.
1996, 2001 and 2006 Census data.



Item non-response

Item non-response occurs when Census forms are returned to the ABS but not all questions are answered. Item non-response accounted for 29% of total unknown Indigenous status in 2006.


The question on the standard household form asking about a person's Indigenous origin was moved forward to Question 7 in 2006, from Question 17 in 2001. This was reflected in the reduced item non-response rate for the question (based on all Census forms returned to ABS, i.e. excluding imputed records), from 2.0% ( 365,568 people) in 2001 to 1.7% ( 326,384 people) in 2006, as early questions are more likely to be completed than those appearing later in the form. Between the 2001 and 2006 Censuses , there were decreases in the item non-response rate in all states and territories with the largest decrease observed in the Northern Territory (from 2.5% to 1.4%).

Indigenous status unknown, Item non-response(a)(b) - 1996-2006

1996
2001
2006
Intercensal change
1996-2001
2001-2006
State/Territory
no.
%
no.
%
no.
%
%
%

New South Wales
91 848
1.6
119 909
1.9
110 469
1.8
30.6
-7.9
Victoria
77 075
1.8
93 266
2.0
86 285
1.8
21.0
-7.5
Queensland
61 527
1.9
72 170
2.1
59 748
1.6
17.3
-17.2
South Australia
22 245
1.6
26 896
1.9
24 780
1.7
20.9
-7.9
Western Australia
22 996
1.4
34 523
1.9
29 591
1.6
50.1
-14.3
Tasmania
10 191
2.2
10 368
2.3
9 380
2.0
1.7
-9.5
Northern Territory
2 504
1.5
4 435
2.5
2 446
1.4
77.1
-44.8
Australian Capital Territory
2 451
0.8
3 901
1.3
3 626
1.2
59.2
-7.1
Australia(c)
290 851
1.7
365 568
2.0
326 384
1.7
25.7
-10.7

(a) Usual residence Census counts.
(b) Percentages based on all Census forms returned to ABS, i.e. excluding imputed records.
(c) Includes Other Territories. See Glossary.
1996, 2001 and 2006 Census data.



Characteristics of non-respondents to the Indigenous status question

The item non-response rate in 2006 was higher among people aged 75 years and over (4.7%) and lower (1.4%) for people under 65 years of age. People aged 65 years and over accounted for 29% of total item non-response to the Indigenous status question, although they represented only 14% of the people counted in the Census when a form was returned.


Among people for whom a response to the Indigenous status question was not recorded, the item non-response rate for questions on each of birthplace, ancestry, language spoken at home or religious affiliation was greater than 15%. Some 24,545 records did not have a response to any of these questions.


Responses to other Census questions may indicate that a record with Indigenous status item non-response is actually for an Indigenous person. There were 679 records that did not indicate Indigenous status, but did report Australian Indigenous ancestry (in some cases together with an Indigenous language and/or religious affiliation), while an additional 201 people reported speaking an Indigenous language and/or having an Indigenous religion but did not report Indigenous ancestry. The ABS does not edit these records during processing and they are therefore left with Indigenous status unknown.


The Indigenous status question was not pre-coded on the Interviewer Household Form because it must cater for different Indigenous responses (i.e. 'Aboriginal', 'Torres Strait Islander', or 'Both'), and for non-Indigenous people living in discrete Indigenous communities. Therefore, while the item non-response rate for Indigenous status is lower for Interviewer Household Forms (0.7%) than the overall rate of 1.7%, interviewing does not eliminate item non-response.



UNDERCOUNT

Each Census, some people are missed and others are counted more than once. In Australia, a greater number of people are missed than are counted more than once and the overall effect is called net undercount. The Post Enumeration Survey (PES), conducted about one month after Census Night, is used to measure net undercount. This is an important indicator of the quality of the Census count because it is an independent measure of Census coverage.


The ABS expanded the scope of the PES in 2006 to include both remote areas of Australia and discrete Indigenous communities. Care should be taken when comparing the 2006 Census net undercount with that of previous Censuses due to changes in the scope of the survey and changes in how the undercount was estimated from the survey data.


The PES estimate of the size of the Indigenous population (excluding Other Territories) at the time of the 2006 Census was 513,977 people compared with the Census count for Australia (excluding Other Territories) of 454,799. The Census count of Indigenous people excludes people whose Indigenous status was unknown in the Census. The difference of 59,178 represents 11.5% of the PES estimate. See also Appendix 2: Indigenous Estimated Resident Population - Method of Calculation.


For information about PES methodology and undercount see:

  • Information Paper: Measuring Net Undercount in the 2006 Population Census (cat. no. 2940.0.55.001)
  • Census of Population and Housing - Undercount, Australia (cat. no. 2940.0)