TECHNICAL NOTE 1 IMPROVEMENT IN COLLECTION OF INDIGENOUS STATUS
OVERVIEW
1 The ABS has undertaken an additional and extensive quality assurance process to understand the components of Indigenous net undercount in 2006 and 2011. This process identified that the main explanatory factor behind the change was the improved PES methodology and procedures, which led to better identification of Aboriginal and Torres Strait Islander persons in the 2011 PES and a noticeably different net category change outcome. Net category change is explored in detail in Components of net undercount and in the Differences in classification between the PES and Census Technical Note.
2 As a consequence of this important improvement in methodology and procedures, the 2006 and 2011 PES estimates are not directly comparable, with the 2011 methodology providing an improved estimate of net undercount for Aboriginal and Torres Strait Islander people.
CHANGE TO THE QUESTIONNAIRE
3 In order to ensure that Indigenous status was effectively collected in the 2011 PES and that all Aboriginal and Torres Strait Islander persons were identified, the ABS removed a household 'screening' question that had been previously used in the mainstream PES questionnaire. This meant that the Indigenous status question was collected for every person in the dwelling. This aimed to address a suspected response bias in the 2006 questionnaire.
4 This represented a departure from the format of previous PES questionnaires, but was considered by the ABS to be a necessary departure based on trials of the change in the 2010 PES Dress Rehearsal. It was also well supported by a robust Interviewer training program and support materials, which reinforced the importance of collecting and recording Indigenous status correctly for all respondents.
IMPACT
5 The impact of the improved collection of Indigenous status in the PES is evident through analysis of the components of the 2006 and 2011 net undercount for Aboriginal and Torres Strait Islander people. As the following table shows, the net category change in 2006 effectively reduced the net undercount rate in the contact sector by around 6% (or 30,797 persons), in contrast to 2011, where it resulted in a slight increase of 0.8% (or 5,128 persons). The remainder of the net undercount rate was therefore similar between 2006 (17.5%, or 11.5% plus 6%) and 2011 (16.5%, or 17.2% minus 0.8%, in rounded terms), a non-statistically significant difference.
COMPONENTS OF NET UNDERCOUNT(a), Indigenous status - 2006 and 2011 |
|
 | | | | | |
| | PES population estimate | Persons missed in the Census | Net difference in classification | Persons with Census category not-stated | Total undercount in
the contact sector (b) | Undercount in non-contact sector | Net undercount |
|
| 2006 | | | | | | | |
| Indigenous (no.) | 513 977 | 54 056 | -30 797 | 7 989 | 31 248 | 27 930 | 59 178 |
| Proportion of PES population estimate (%) | 100.0 | 10.5 | -6.0 | 1.6 | 6.1 | 5.4 | 11.5 |
| 2011 | | | | | | | |
| Indigenous (no.) | 662 335 | 56 650 | 5 128 | 7 009 | 68 787 | 45 402 | 114 188 |
| Proportion of PES population estimate (%) | 100.0 | 8.6 | 0.8 | 1.1 | 10.4 | 6.9 | 17.2 |
|
(a) 2006 and 2011 estimates are included in this table to support decompositional analysis only, and are not strictly comparable.
(b) Total undercount in the contact sector is the sum of Persons missed in the Census, Net difference in classification and Persons with Census category not-stated. |
6 The increase in the number of respondents in the 2011 PES who were identified as Aboriginal and Torres Strait Islander also resulted in a higher incidence of differential identification of Indigenous status between the PES and Census, upon which the category change is based. As the following table shows, in 2006 there were 116 PES respondents who were Aboriginal and Torres Strait Islander in the PES and non-Indigenous in the Census, compared with 283 respondents with the reverse combination. In contrast, in 2011 there were 268 and 316 respondents. While the category change is based on weighted estimates, rather than unweighted sample counts, the change in the relationship between these differential classification categories is clear.
7 In 2006, the identification of Indigenous status in the PES suggested that Census counts of Aboriginal and Torres Strait Islander persons were greater than they should have been and a net category change of -30,797 was applied to the estimate of net undercount in the contact sector. In 2011, the relationship between the two categories of differential classification was close to balanced, with a net category change of just 5,128 applied to the estimate of net undercount in the contact sector.
DIFFERENCE IN CLASSIFICATION, Indigenous status, Responding PES persons - 1991-2011 |
|
| PES Indigenous, Census non-Indigenous | PES non-Indigenous, Census Indigenous | Ratio of two differentially identified groups | Net difference in
classification |
| no. | no. | % | no. |
|
2011(a) | 268 | 316 | 0.85 | 5 128 |
2006(b) | 116 | 283 | 0.41 | -30 797 |
2001 | 133 | 246 | 0.54 | na |
1996 | 86 | 227 | 0.38 | na |
1991 | 174 | 192 | 0.91 | na |
|
| (a) Care should be taken when comparing 2011 estimates with previous years due to changes made to PES linking and matching methodology. For more information see Linking and Matching. |
| (b) Care should be taken when comparing estimates from 2006 onwards with previous years due to a new estimator, Prediction Regression (PREG), being implemented in 2006, as well as the inclusion of remote areas and discrete Indigenous communities in the PES sample from 2006. For further information see the Explanatory Notes of Census of Population and Housing - Undercount, 2006 (cat. no. 2940.0). |
8 The 2011 unweighted sample count ratio of the two differentially identified groups (0.85) was closest to the ratio seen in the 1991 PES (0.91) and was noticeably different to the ratio seen in the previous 3 surveys. This suggests that in 2011, prior to estimation, both types of differential identification occurred roughly a similar number of times, reflecting a different classification relationship between the two collections over time.
9 In weighted terms, the differential classification in 2011 contributed to the net category change of 5,128. That is, in 2011 PES identification of Indigenous status suggested the Census counts of Aboriginal and Torres Strait Islander persons were less than they should have been. This accounts for most of the difference between 2006 and 2011 estimates of net undercount.
Print Page
Print All