MAJOR CHANGES TO THE 2016 PES
1 There were a number of changes implemented for the 2016 PES, some of which resulted from the changes made to the 2016 Census.
2 Key changes made to the PES field collection phase were:
- A sample size increase of 20%. This was done to ensure the high level of accuracy for the PES estimates (as measured by the sampling error) was maintained, as well as to manage the risk of a number of operational unknowns when the PES sample was being designed (e.g. potential overlap in Census and PES enumeration) that could have significantly impacted the PES response rates and sample size
- The focused sampling strategy to improve estimates for the Aboriginal and Torres Strait Islander population was removed, as a review in 2015 indicated that most of the improvement in the estimates for this population was achieved through other means
- Follow-up phases for the 2016 Census started and finished several weeks later than the 2011 Census. Accordingly, the 2016 PES enumeration started later than in 2011 and commenced immediately following conclusion of the Census follow-up phase
- The PES enumeration period was extended to cater for the larger sample and the increasing difficulty in making contact with householders
- A telephone interview option was introduced to provide more convenience and choice for PES respondents. Respondents were able to register for a telephone interview via a number of methods (online or via telephone) by following instructions provided in the Primary Approach Letter. The introduction of telephone interviewing improved the effectiveness of enumerating the large 2016 PES sample in a short timeframe.
Key changes made in the PES processing and estimation phase include:
- Address text matching was introduced to provide an opportunity to link PES dwellings to a specific entry on the ABS Address Register
- For the purposes of PES estimation, the respondent’s age (which is collected as at the time of the PES interview) was backdated to 9 August 2016 (i.e. their age on Census night) using date of birth, where provided
- Census records that had insufficient personal identifier information, thus reducing the likelihood of linking to PES, were identified and moved to the Census non-contact sector for the purposes of PES estimation.