A5.1. Estimates of the Indigenous population are experimental in that the standard approach to population estimation is not possible because satisfactory data on births, deaths and internal migration are not generally available. Furthermore, there is significant volatility in census counts of the Indigenous population, thus adding to the estimation problems. This volatility can in part be attributed to changes in the propensity of persons to identify as being of Indigenous origin. As a result, a method based on the use of life tables is used to produce time series data.
A5.2. Indigenous estimates based on the 1996 Census of Population and Housing are significantly higher than those which would have been expected if the Indigenous population in 1991 had been subject to expected levels of natural increase and migration over the 1991 to 1996 intercensal period.
A5.3. Estimates for Jervis Bay Territory are included in the total population estimates. Counts from the 1996 Census of Population and Housing include 24 persons usually resident in the Cocos (Keeling) Islands and Christmas Island who had Indigenous origin reported. These persons have been excluded from these estimates because of the small numbers involved.
1986 Census based estimates
A5.4. Indigenous estimates as at 30 June 1986, based on the 1986 Census of Population and Housing, have been provided in the latest ABS publication Experimental Estimates of the Aboriginal and Torres Strait Islander Population (3230.0). These data were previously published in Experimental Estimates of the Aboriginal and Torres Strait Islander Population, 1986 to 1991 (3230.0).
1991 Census based estimates
A5.5. Indigenous population data as at 30 June 1986, 30 June 1991 and 30 June 1996 have been reproduced in the latest ABS publication of Experimental Estimates of the Aboriginal and Torres Strait Islander Population (3230.0), based on the 1991 Census of Population and Housing. The 1991 data have previously been published in Experimental Estimates of the Aboriginal and Torres Strait Islander Population, 1986 to 1991 (3230.0) and subsequently revised in Experimental Projections of the Aboriginal and Torres Strait Islander Population, June 1991 to June 2001 (3231.0).
1996 Census based estimates
A5.6. Indigenous estimates based on the 1996 Census have been provided for 30 June 1991 and 30 June 1996 in Experimental Estimates of the Aboriginal and Torres Strait Islander Population (3230.0). This publication also provides Indigenous estimates at the ATSIC Region level for 30 June 1996. The method described in the remainder of this appendix specifically refers to the 1996 Census-based Indigenous estimates.
A5.7. Indigenous population estimates have been produced for each year from 30 June 1996 back to 30 June 1991. The source of these statistics are Indigenous person counts from the 1996 Census of Population and Housing based on place of usual residence. In the census, place of usual residence refers to the address at which a person has lived or intends to live for a total of 6 months or more in the census year.
A5.8. The method of estimation was a two stage process. Firstly, various adjustments were applied to census counts to produce the estimated Indigenous population as at 6 August 1996. Secondly, these 6 August Indigenous estimates were initially 'survived' back to 30 June 1996 and then back to 30 June 1991 using life tables. This method is known as the reverse survival method. Zero net internal and external migration were assumed.
A5.9. Before census counts as at 6 August 1996 were survived back to 30 June 1996, various adjustments were made. The adjustment process involved four factors: persons recorded as Indigenous whose parents were born overseas; non-response to the Aboriginal and Torres Strait Islander origin question in the census; net census undercount; and adjustments based on registered Indigenous births.
A5.10. The first adjustment relates to Indigenous persons where both parents were overseas born. The census counts in 1996 included Indigenous persons where both parents were born overseas. An adjustment was applied to the census counts of Indigenous persons to exclude those Indigenous persons whose parents were born overseas. A person was deemed to have been born overseas if their birthplace was not recorded as either 'Australia' or 'Not Stated'.
A5.11. The second adjustment dealt with those persons for whom basic details were available but there was no response to the Aboriginal and Torres Strait Islander origin question on the census form. This non-response may be due to two factors. One is that a form was completed but there was no response to the Indigenous origin question. The other is where the census collector considered that a person did reside at a particular residence on census night, but was not able to obtain a completed census form for that person. In the latter case, a 'dummy' census form was created where age, sex and state of usual residence on census night are imputed, but all other characteristics (including Aboriginal and Torres Strait Islander origin) remain 'Not Stated'.
A5.12. It was assumed that some of these non-respondents were actually of Indigenous origin. The imputation procedure allocated non-respondents on each type of form to being of Indigenous origin according to the proportion of Indigenous in the total stated response on that form type (ie. household, person, Special Indigenous Form). Imputation was applied individually for males and females in each Statistical Local Area.
A5.13. The third adjustment involved correcting for net undercount of the Indigenous population. Results from the 1996 Post Enumeration Survey (PES) showed that there was a significant difference between the net undercount rates for Indigenous and non-Indigenous persons. As a result, it was decided to incorporate the Indigenous net undercount into Indigenous estimates. These estimates are the first to use a specific Indigenous undercount rate. Estimates based on data from the 1991 Census of Population and Housing used the undercount rate corresponding to the total population. The net Indigenous and non-Indigenous undercount rates from the 1996 PES have been reproduced previously in Table 4.5.
A5.14. Total undercount rates vary considerably by age, sex and State. For example, undercount is consistently highest for males aged 20 to 29 years. Thus it was considered that a more accurate estimate of the Indigenous population would be achieved by applying an age, sex and State specific undercount distribution instead of a uniform 7.1% undercount rate. However, high sampling errors precluded the use of PES results of indigenous undercount by age, sex and State. Instead, the age, sex and State undercount distribution for the total Australian population was applied to the total Indigenous undercount rate of 7.1%. Although the PES was not conducted in remote areas or Indigenous communities where special procedures were used in the Census, the undercount was applied to all geographic areas in the absence of other information.
A5.15. Lastly, adjustments were made based on registered Indigenous births. Following the first three adjustments, the numbers at age zero were compared with registered births. The births registered in South Australia, Western Australia, Tasmania, the Northern Territory and the Australian Capital Territory were higher than the numbers at age zero. In these States and Territories, the numbers at age zero were raised on the basis of the number of registered births. The following table shows these registrations during the financial year 1995–96 and subsequent adjustments.
A5.16. Ages 0, 1, 2, 3 and 4 were then smoothed for all States and Territories to maintain relativities between ages. The new total for the age group 0 to 4 remained unchanged.
A5.17. After adjustments had been made to produce Indigenous population estimates as at 6 August 1996 (the Census date), the estimates were backdated to 30 June 1996 (30 June is the standard reference date for population estimates). In normal circumstances, the cohort component method would have been employed. This method updates a base estimated population (stock figure) by the use of data on demographic flows (flow data). These demographic events are births, deaths and migration. However, satisfactory flow data for the Indigenous population were not generally available. As a result, the estimates of the Indigenous population as at 30 June 1996 were obtained by surviving the adjusted census counts back using life tables. This method is known as the reverse survival method. No attempt has been made to adjust for any change in propensity to identify as Indigenous. Instead, independent series have been produced using both the 1991 and 1996 Censuses.
A5.18. While net internal migration has increased in the 1991–1996 intercensal period compared with the 1986–1991 intercensal period, the absolute size of the migration still does not warrant a specific interstate migration assumption. As such, zero internal migration has been assumed. Census data also indicates that the level of external migration of Indigenous persons is negligible. Hence zero external migration has also been assumed.
A5.19. The life table used to backdate the population is listed in ABS publication Experimental Estimates of the Aboriginal and Torres Strait Islander Population (3230.0). This life table was produced using the Preston-Hill method by comparing the change in counts of the Indigenous population between the two censuses with the number of Indigenous deaths registered for each five-year cohort. This enables the rate of increase in identification and the level of under-enumeration of deaths to be calculated. With this data, it is possible to estimate the number of deaths and the population on a consistent basis and to produce a life table for the period 1991–1996.
A5.20. Two modifications then took place. Firstly, female mortality rates were adjusted such that when the 1996 estimates were backdated to 1991, the sex ratios of the resultant population closely matched those recorded at the 1991 Census. Secondly, a revision was made at ages 0 and 1. The resulting table is a modification of that which appeared in Deaths, Australia, 1996 (3302.0).
Propensity to identify
A5.21. The difference between the 1991 and 1996 Census counts of Indigenous persons, as recorded on census forms, was significantly larger than expected despite similar collection procedures for the two censuses. This increase cannot be fully accounted for by natural increase (births minus deaths) and net migration over the intercensal period. While improvements in census procedures have taken place, it appears that there has been a large increase in the propensity to indicate Indigenous origin on census forms between the two censuses. This is consistent with the previous intercensal period, in which there was a significant difference between the 1986 and 1991 Censuses which could not be fully accounted for. Population estimates based on the 1986 and 1991 Censuses as well as the 1996 Census have been published to indicate the impact of this change in propensity to identify.
A5.22. The PES shows that the reported Indigenous origin of persons can change over a relatively short time. The PES is a survey conducted after the Census which asks similar demographic questions to the Census, including a question on Indigenous origin. While the PES is conducted only three weeks after the Census, some individuals still have different responses to the Indigenous question in both the Census and the PES. The following table shows the indicated Indigenous origin of people in the 1996 PES and their corresponding Census response. This table includes only those persons who stated in the PES that their birthplace was Australia. The PES did not collect data on 'Mixed Origin'.
A5.23. The number of people who were identified as being of Indigenous origin was 8.0% less in the 1996 PES compared to the Census. Results from 1991 and 1986 indicate that this relationship varies considerably. In 1991, there was a reduction in Indigenous persons of 1.7% in the PES compared to the Census. In 1986, the number identified as being of Indigenous origin was 10.6% higher in the PES than in the 1986 Census.
A5.24. Some differences can be expected because the Census and the PES use different collection procedures. For example, census forms are generally delivered to households and self-enumerated by those at the household while PES forms are administered by an interviewer who interviews any responsible adult in the household. As such, the difference in the Indigenous origin for people between the PES and Census can in part be attributed to the difference in methods.
A5.1: Indigenous Origin, 1996 Census and Post-enumeration Survey
|Torres Strait Islander|
(a) Persons of Mixed Origin are those indicating both Aboriginal and Torres Strait Islander origin.
Definitions of Aboriginal and Torres Strait Islander origin
A5.25. The 1996 Census form asked the following, as question 14, of each person:
The question appeared as question 13 in the 1991 Census form. It was identical except for the prompt for persons of mixed origin:
A5.26. The Census question is based on the first aspect of the definition adopted by the Commonwealth Government in 1978 of Aboriginal and Torres Strait Islander origin but implicitly measures the second aspect as well. The definition states that an Aboriginal or Torres Strait Islander is:
- a person of Aboriginal or Torres Strait Islander descent,
- who identifies as an Aboriginal or Torres Strait Islander, and
A5.27. For the 1996 Census, some assistance was available in a 'General Inquiry Guide' and through the Census Hotline which provided the following explanations:
- is accepted as such by the community in which he or she lives.
- For census purposes, 'Aboriginal' does not refer to everyone born in Australia, or to the aboriginal people of any other country.
- Mark 'Yes, Aboriginal' if you are of Australian Aboriginal descent and identify yourself as Aboriginal.
- 'Torres Strait Islander' means people who originally come from the Torres Strait Islands located between the Australian mainland and Papua New Guinea.
- Mark 'Yes, Torres Strait Islander' if you are of Torres Strait Islander descent and identify yourself as a Torres Strait Islander.
- Persons of both origins should mark both 'yes' boxes.
- If someone is from another group of islands, tell them to answer 'No', unless they have an ancestor who was Aboriginal or from the islands between Cape York and Papua New Guinea.
This page last updated 12 March 2007