This document was added or updated on 08/05/2020.
Modelled estimates for small areas
1 Data cubes 2015 SDAC SA2 modelled estimates and 2015 SDAC LGA modelled estimates contain modelled estimates of disability and carers for small areas based on data from the 2015 Survey of Disability, Ageing and Carers (SDAC), 2016 Australian Census of Population and Housing, the Estimated Resident Population (ERP) as at 30 June 2015, and aggregated administrative data sourced from the Public Health Information Development Unit (PHIDU), 2018 release.
2 These modelled estimates of characteristics associated with disability and carers at a small area level for the Australian population were produced by the ABS. This core set of tables is similar to consultancies that the ABS undertook in 2015 (based on 2012 SDAC data) and 2010 (based on 2009 SDAC data), which produced a wider set of modelled small area estimates.
3 The modelled estimates for small areas can be interpreted as the expected value for a typical area in Australia with the same characteristics. There will be differences between the disability or carer characteristic prediction and the actual number of people with that characteristic not accounted for in the measure of accuracy. One explanation for this is that significant local information about particular small areas exists, but has not been included in the model as it is not available to the ABS. It is important to consider local area knowledge, such as information on disability or carer related facilities and businesses in the area, when interpreting the modelled estimates for that region.
4 Used in conjunction with an understanding of local area characteristics and their reliability limitations, modelled estimates for small areas can assist in making decisions on issues, such as the requirement for services, relevant to disability and carer populations at the small area level. Care needs to be taken to ensure decisions are not based on inaccurate estimates. It is recommended that the provided modelled estimates for small areas are aggregated to larger regions (such as regional planning regions) as this will improve the accuracy of the estimates upon which decisions may be based.
5 The modelled estimates for small areas are applicable to private dwellings in scope of the SDAC 2015 private dwellings collection. Data for special dwellings (approximately 15.6% of the total 2015 SDAC sample) was excluded. Please refer to the Notes tab within each spreadsheet for the population group each table of data relates to.
6 Modelled estimates for small areas have been produced at the Statistical Area Level 2 (SA2) and Local Government Area (LGA) level for all jurisdictions.
7 SA2s are defined within the main structure of the Australian Statistical Geography Standard (ASGS). They are medium-sized general purpose areas built up from whole Statistical Areas Level 1 (SA1s). Their purpose is to represent a community that interacts together socially and economically.
8 LGAs are not defined or maintained within the main structure of the ASGS. They are an ABS approximation of gazetted local government boundaries as defined by each State and Territory Local Government Department. For more detailed information see the 2015 SDAC Small Area Estimates Explanatory Notes PDF on the Downloads tab.
9 To produce accurate and detailed estimates of disability and carer characteristics at the small area level, models are created using detailed SDAC data, in conjunction with Census data, and ERP data to produce modelled estimates for small areas. The modelling method assumes that the relationships observed at the higher geographic level (as available in SDAC) between the characteristics of interest and known characteristics also hold at the small area level. Section 3 of the 2015 SDAC Small Area Estimates Explanatory Notes PDF on the Downloads tab provides full details of the process used to produce modelled estimates for small areas.
Reliability of estimates
10 The errors associated with the modelled estimates for small areas fall into four categories. Sampling error, non-sampling error, modelling error, and prediction error. The relative root mean squared error (RRMSE) provides an indication of the deviation of the modelled estimate from the true value. The RRMSE is primarily a measure of prediction error, but in its calculation it also inherits some aspects of modelling and sampling error. Section 4 of the 2015 SDAC Small Area Estimates Explanatory Notes PDF on the Downloads tab provides details on the accuracy of results. Instructions for aggregating modelled estimates (with worked examples) are provided in the relevant example tabs of each spreadsheet.
11 Estimates have been confidentialised to ensure they meet ABS requirements for confidentiality.
12 Because SDAC population benchmarks have been used in the modelling process, the modelled estimates provided here can also be considered perturbed. Users should note that due to perturbation, the summing, or aggregation, of the modelled estimates to derive a total (e.g. at state level) will not necessarily give the same result as the published total. In these cases, the difference between the sum of modelled estimates for small areas and the published total will be small and will not impact the overall information value of the aggregate total or any individual component.
13 Aggregation of the modelled estimates of small areas to capital city or state/territory level is not recommended. If you require capital city or state/territory level data for the characteristics of disability and carers provided here at small area level, the appropriate source is published survey data (and/or use of the TableBuilder product).
14 The tables of modelled estimates include a 'population' count created solely for analysis of the small area data; these are not official ABS population statistics.
15 The 2015 SDAC Modelled Estimates for Small Areas Explanatory Notes PDF accompanies the modelled estimates for small areas and can be found on the Downloads tab. We recommend reading the full content of these explanatory notes to ensure the best and most appropriate usage of the data.