PRELIMINARY ERP BY STATE AND TERRITORY – OTHER QUALITY CONSIDERATIONS
RESIDENTS TEMPORARILY OVERSEAS
Estimated Resident Population or ERP is based on the usual residence (UR) population concept. Therefore, Census counts for persons who were not 'usual residents' need to be adjusted so they align with the usual resident concept when deriving Australia's official population estimates. Australian residents who are temporarily overseas on Census night are one group that need to be accounted for in the ERP.
Preliminary ERP uses a preliminary RTO method because it takes an extended period of time to determine actual RTO based on traveller behaviour. If a resident leaves Australia on Census night, based on the '12/16 month rule' that is used for determining overseas migration, it can take up to 16 months to determine whether it was a temporary trip or a non-temporary relocation. [The '12/16 month rule' for determining overseas migration specifies that a resident has permanently moved out of Australia if they are overseas for a total 12 months or more during a 16 month period.]
Quality assurance testing for the RTO adjustment shows that close to 97–98% of RTOs return to Australia within the first six months of the Census night and have their RTO status resolved in time for the preliminary ERP publication. This pattern is the same across all states and territories. Therefore, the ABS needs only to model that proportion of the remaining 2–3% of residents who have not returned within six months who will eventually be classed as temporary after the 16 month period. This model has proven to be very robust and can be tested on any 16 month time period for stability and validity.
Additional information can be found in Information Paper: Improving Net Overseas Migration Estimation, Mar 2010 (cat. no. 3412.0.55.001), Overseas Arrivals and Departures, Australia (cat. no. 3401.0) and Migration, Australia (cat. no. 3412.0).
Census quality assurance processes are broad and detailed, and more extensive than those which are required for ERP purposes. This is because Census contains over 50 variables, only four of which are used for the production of ERP. These variables are sex, age, Indigenous status and place of usual residence.
The extensive quality assurance processes undertaken for the 2011 Census are well documented in How Australia Takes a Census, 2011 (cat. no. 2903.0)
POST ENUMERATION SURVEY
The Post Enumeration Survey (PES) delivers the Census net undercount estimates that are used in rebasing Australia's official population estimates. Undercount estimates are also used to assess the completeness of Census counts, and evaluate the effectiveness of Census collection procedures so that improvements may be made for future Censuses.
The introduction of Automated Data Linking (ADL) in 2011 was a significant improvement compared with previous years. ADL allows a more accurate match between people included in the Census and people surveyed in PES in order to estimate undercount, overcount, and therefore net undercount.
Like the Census, the quality assurance process for the PES is extensive and well documented. More information on 2011 quality assurance processes can be found in Census of Population and Housing – Details of Undercount (cat. no. 2940.0) and Information Paper: Measuring Net Undercount in the 2011 Population Census (cat. no. 2940.0.55.001).