|Page tools: Print Page Print All|
Usual residents are people who usually live in a particular dwelling and regard it as their own or main home. People usually resident in non-private dwellings, such as hotels, motels, hostels, hospitals, nursing homes, or short-stay caravan parks were not in scope.
Further scope exclusions for this survey were:
The 2014–15 NATSISS was conducted in remote and non-remote areas in all states and territories of Australia, including discrete Aboriginal and Torres Strait Islander communities. Coverage refers to the extent to which the defined scope is represented by the achieved sample. Whereas, undercoverage is the shortfall between the population represented by the achieved sample and the in-scope population. Undercoverage can be planned or unplanned. In this survey, coverage exclusions were explicitly applied to some people who were part of the in-scope population to manage enumeration costs. These people were not included in the sampling frame (based on where Aboriginal and Torres Strait Islander households were identified in the 2011 Census of Population and Housing). These exclusions were:
These coverage exclusions result in an estimated undercoverage of approximately 6% of in-scope Aboriginal and Torres Strait Islander persons in Australia. Although these areas were not enumerated, the final sample was weighted to population benchmarks to account for these exclusions. In addition to this, there were further unplanned exclusions which resulted in increased undercoverage. Further information on undercoverage is provided in the Non-sampling error chapter of this publication, and in the Explanatory notes of the summary publication.
For this survey, the population benchmarks were projections of the most recently released Estimated Resident Aboriginal and Torres Strait Islander Population (ERP) data, in this case, 30 June 2011. For information on the methodology used to produce the projected resident Aboriginal and Torres Strait Islander population see Estimates and Projections, Aboriginal and Torres Strait Islander Australians, 2001 to 2026 (cat. no. 3238.0). To create the population benchmarks for the 2014–15 NATSISS reference period, the Aboriginal and Torres Strait Islander ERP from 30 June 2011 was projected forward to 31 December 2014 using average annual growth rates observed between the 2006 and 2011 censuses.
The projected resident Aboriginal and Torres Strait Islander population at 31 December 2014, excluding those living in non-private dwellings, was 686,800.
The coverage of this survey included persons who self-identify (in a face-to-face question) as being of Aboriginal or Torres Strait Islander origin, whereas the benchmarks (scope) are based on persons who identified as being of Aboriginal or Torres Strait Islander origin in the Census and Post Enumeration Survey (PES) using different identification mechanisms. It is worth noting that whether a person identifies as being of Aboriginal or Torres Strait Islander origin or not can depend on the data collection methodology. For example, while the NATSISS was collected via face-to-face interviews, the 2011 Census was generally collected via a self-completion form.
The 2014–15 NATSISS was designed to produce reliable estimates at the national level and for each state and territory. For persons aged 15 years and over, the sample was allocated to produce estimates with a relative standard error (RSE) of no more than 25% for NSW, Vic, Qld, SA, WA , Tas, and NT, for characteristics that at least 3% of the population would possess. For the ACT and the Torres Strait Area, estimates for characteristics possessed by at least 6% of the the population were required to have an RSE no greater than 25%. For selected states and territories (NSW, Qld, WA and NT) the sample for children aged 0–14 years was allocated to produce estimates that have an RSE of no greater than 25% for characteristics that at least 5% of these populations would possess. The survey was also designed to provide reliable estimates at the national level for children aged 0–3 years and 0–14 years, with the same RSE requirements.
As with previous ABS Aboriginal and Torres Strait Islander surveys, additional sample was collected in the Torres Strait Area, to ensure data of sufficient quality would be available for the Torres Strait Area and the remainder of Queensland.
Funding was received from the Department of Health to enable the continued of collection of data on Aboriginal and Torres Strait Islander children aged 0–14 years.
Community and non-community samples
The sample design incorporated a random selection of:
The samples for community and non-community areas were designed separately, with each involving a multistage sampling process. These two designs were combined to ensure that all areas of the in-scope population had a chance of being selected in the survey. It was not possible for someone to be selected in both the community and non-community sample. The community and non-community samples included population from remote and non-remote areas of Australia.
The sample design differed by community and non-community areas in:
Corresponding with the method used in non-community areas of NSW, Vic, Qld, SA, WA and the NT, the sample design for the following states/territory was the same throughout:
The community sample was obtained from a random selection of discrete Aboriginal and Torres Strait Islander communities and out-stations using a specially developed Aboriginal and Torres Strait Islander Community Frame. This frame was constructed using counts from the 2011 Census of Population and Housing and information collected in the 2012–13 National Aboriginal and Torres Strait Islander Health Survey (NATSIHS). Communities on this frame were in both remote and non-remote areas of Australia. From this frame, 73 Community Sets (containing one main community and zero or more out-stations) were selected for enumeration. The number of communities selected for each applicable state/territory were:
A random selection of dwellings was made within the selected communities and out-stations, with different selection procedures applied to the main communities and out-stations. Depending on the size of the main community, up to 48 in-scope dwellings were selected for enumeration. Where out-stations existed, one day was allocated to enumerating the out-station. Any available household was enumerated in out-stations within this time constraint. Within each selected dwelling in remote areas, up to one Aboriginal and/or Torres Strait Islander adult (aged 15 years and over) and up to one Aboriginal and/or Torres Strait Islander child (aged 0–14 years) was randomly selected to participate in the survey. In non-remote areas, up to two adults and two children were randomly selected to participate in the survey.
In non-community areas, dwellings were selected using a stratified multistage area sample. Mesh Block level information within SA1s was used to assist in targeting Aboriginal and Torres Strait Islander people. A sample of SA1s was selected with the likelihood of selection based on the number of Aboriginal and Torres Strait Islander households recorded in the area for the 2011 Census. Some areas with small numbers of Aboriginal and Torres Strait Islander people according to Census data were excluded from the frame to improve the sampling efficiency. This approach significantly reduced screening effort in areas of low Aboriginal and Torres Strait Islander density, such as major capital cities. In SA1s with a large expected population of Aboriginal and Torres Strait Islander persons (based on Census 2011 data), a systematic sample of dwellings (using a skip) was screened rather than every dwelling.
For each randomly selected dwelling, one usual resident aged 18 years or over, was asked whether anyone in the household was of Aboriginal or Torres Strait Islander origin. This screening question was used to identify Aboriginal and Torres Strait Islander households, from which the sampling process of persons was undertaken. If a dwelling contained one or more usual residents of Aboriginal or Torres Strait Islander origin, random selection for participation in the survey occurred as follows:
In non-remote areas:
In remote areas:
The original sample allocation for the non-community component for each state/territory appear in the following table:
Table 2.2.1. Sample allocation(a), Non-community
Weighting, benchmarking and estimation
Weighting is the process of adjusting results from a sample survey to infer results for the total in-scope population. To do this, a 'weight' is allocated to each sample unit corresponding to the level at which population statistics are produced. For example, at person or household level. The weight can be considered an indication of how many population units are represented by the sample unit. For the 2014–15 NATSISS, separate person and household weights were developed.
The first step in calculating weights for each person or household is to assign an initial weight, which is equal to the inverse of the probability of being selected in the survey. For example, if the probability of being selected in the survey was 1 in 45, then the person would have an initial weight of 45 (that is, they would represent 45 people).
The person and household weights were separately calibrated to independent estimates of the population of interest, referred to as 'benchmarks'. Weights calibrated against population benchmarks ensure that the survey estimates conform to the independently estimated distributions of the population rather than to the distribution within the sample itself. Calibration to population benchmarks helps to compensate for over- or under-enumeration of particular categories which may occur due to either the random nature of sampling, non-response, non-identification or various other undercoverage factors. This process can reduce the sampling error of estimates and may reduce the level of undercoverage bias.
A standard approach in ABS household surveys is to calibrate to population benchmarks by state, part of state, age and sex. In Aboriginal and Torres Strait Islander surveys, benchmarking against remoteness areas is also usual. In terms of the effectiveness of 'correcting' for potential undercoverage bias, it is assumed that within the weighting classes the characteristics of the covered population (ie survey respondents) are similar to the uncovered population (ie people who were not surveyed), as determined by the benchmarking strategy. Where this assumption does not hold, biased estimates may result.
For this survey, person weights were simultaneously calibrated to the following population benchmarks:
The 'state' population benchmarks consist of the six states and two territories:
'Remoteness area' was categorised into five benchmark groupings, defined in the table below.
'Age group' was benchmarked against the following groupings, of persons aged:
'Sex' was benchmarked against two categories:
'Remote community status' was benchmarked against two categories:
'Remote community' included all persons enumerated in dwellings in discrete Aboriginal and Torres Strait Islander communities in remote and very remote areas of Australia. All other people were considered to be in the category 'not a remote community', including those in non-remote discrete Aboriginal and Torres Strait Islander communities. As there were no communities selected in Tasmania in the community sample, both benchmarks in Tasmania were collapsed into 'not in a remote community'.
'Torres Strait Islander status' consisted of two categories:
'Torres Strait Region' consisted of three categories:
'Adult status' consisted of two categories:
The survey was benchmarked to the estimated Aboriginal and Torres Strait Islander resident population living in private dwellings at 31 December 2014. This estimated population is based on projections of the latest Estimated Resident Aboriginal and Torres Strait Islander Population (ERP) at 30 June 2011. More information on the calculation of projections is provided in Estimates and Projections, Aboriginal and Torres Strait Islander Australians, 2001 to 2026 (cat. no. 3238.0). As people in non-private dwellings (eg hotels) are excluded from the scope of the survey, they have also been excluded from the survey benchmarks. Therefore, the 2014–15 NATSISS estimates do not (and are not intended to) match estimates for the total resident Aboriginal and Torres Strait Islander population obtained from other sources.
The ABS does not produce Aboriginal and Torres Strait Islander dwelling counts. Therefore, for the 2014–15 NATSISS, household level benchmarks were not available. Instead, the household weights for this survey were derived from the person level benchmarks. This was done by assigning the initial household weight to all Aboriginal and Torres Strait Islander persons in the household. These weights were then calibrated to the person level benchmarks with the constraint that each person in the household must have the same final calibrated weight. The resulting weight was assigned as the final household weight. That is, the weights were calibrated in such a way that the household weights will reproduce the number of known person level benchmarks if all people in a household are included. Therefore, the sum of the household weights will only provide an estimate of the number of Aboriginal and Torres Strait Islander households. This method was then analysed to ensure that person and household level estimates are as consistent as possible.
Estimation is a technique used to produce information about a population of interest, based on a sample of units (ie persons or households) from that population. Each record in the 2014–15 NATSISS contains two weights:
The weights indicate how many population units (ie persons or households) are represented by the sample unit. Replicate weights have also been included—250 person replicate weights and 250 household replicate weights. The purpose of these replicate weights is to enable calculation of the relative standard error (RSE) for each estimate produced from the survey. Further information on replicate weights is provided in the Technical note included in this publication.
Survey estimates of counts of persons are obtained by summing the weights of persons or households with the characteristic of interest. Estimates for means, such as mean age of persons, are obtained by summing the weights of persons in each category (eg individual ages), multiplying by the value for each category, aggregating the results across categories, then dividing by the sum of the weights for all persons.
The majority of estimates contained in this publication are based on benchmarked person weights. However, the survey also contains some household estimates based on benchmarked household level weights.
Perturbation of estimates
To minimise the risk of identifying individuals in aggregate statistics, a technique called perturbation is used to randomly adjust cell values. Perturbation involves a small random adjustment of the statistics and is considered the most satisfactory technique for avoiding the release of information that could identify individual survey respondents while maximising the range of information that can be released. These adjustments have a negligible impact on the underlying pattern of the statistics. After perturbation, a given published cell value will be consistent across all tables. However, adding up cell values to derive a total will not necessarily give the same result as the published totals.
All data presented in the NATSISS summary publication have had perturbation applied to estimates. The introduction of perturbation in publications ensures that published statistics are consistent with statistics released via services such as TableBuilder. It should be noted that this method of perturbation has not been applied to the Expanded CURF, and that statistics from the CURF may not be consistent with published or TableBuilder statistics. For details of perturbation applied to the Expanded CURF, see Using the Expanded CURF in the NATSISS microdata publication.
These documents will be presented in a new window.