TREATMENT OF NON-RESPONSE
Some care should be taken when analysing Census counts, as Indigenous status and other characteristics are unknown for some records. Data quality statements that discuss non-response and editing procedures are available for all Census variables on the ABS website.
The level of non-response should be taken into account when interpreting Census results. Where the level of non-response is low, the distribution of stated responses may reasonably be assumed to be representative of the distribution which would have resulted from the whole population, or sub-population. As non-response rates increase, this assumption becomes less valid.
Non-response to Indigenous status
There were 1,133,466 records with unknown Indigenous status in the 2006 Census, or 5.7% of the total count. These records are generally omitted from the analysis throughout this report, as comparisons are made between known Indigenous and non-Indigenous populations. Where possible, records for which Indigenous status is unknown are included in the total population count, and it is possible to derive the number of these records by subtracting the Indigenous and non-Indigenous populations from the total population.
Non-response to other characteristics
In this report, analysis of population characteristics is limited to known responses only. Where possible, the numbers of unknown responses are noted in the tables, allowing further analysis if required.
The previous edition of this report, produced following the 2001 Census, reported non-response to particular questions in the 'Not stated' or 'Unknown' category, and information on characteristics was presented as a proportion of the total including non-response. As noted above, this convention was not followed in this report. Accordingly, readers should exercise caution when making comparisons with results from the 2001 Census.
For more information on non-response rates see Appendix 1: 2006 Non-response rates.