The preliminary estimates in this release are from the September quarter of the 2020-21 Survey of Income and Housing (SIH). Final estimates from all survey quarters will be released from mid-2022, in Household Income and Wealth, Australia and Housing Occupancy and Costs, Australia.

# Household financial resources methodology

## How the data is collected

### Scope

The scope of the survey includes:

- all usual residents in Australia aged 15 years and over living in private dwellings
- both urban and rural areas in all states and territories, except for very remote areas of Australia and discrete Aboriginal and Torres Strait Islander communities.

The survey excludes the following:

- visitors to private dwellings
- overseas visitors who have not been working or studying in Australia for 12 months or more, or do not intend to do so
- members of non-Australian defence forces stationed in Australia and their dependants
- non-Australian diplomats, diplomatic staff and members of their households
- people who usually live in non-private dwellings, such as hotels, motels, hostels, hospitals, nursing homes and short-stay caravan parks
- households in very remote areas
- households in discrete Aboriginal and Torres Strait Islander communities.

The exclusion of very remote areas is unlikely to impact on national estimates, and will only have a minor impact on any aggregate estimates that are produced for individual states and territories, except the Northern Territory where the excluded population accounts for around 21% of the population.

### Sample design

The September quarter SIH was enumerated from 2 July to 19 September 2020. The enumeration period broadly aligns with pension indexation rates. Dwellings were selected at random using a multistage area sample of private dwellings.

The 2020-21 sample was designed to:

- be representative of income patterns across the financial year
- produce national estimates that have a relative standard error (RSE) no greater than 5% for key income, wealth and housing cost indicators
- produce reliable state and territory estimates and capital city / rest of state estimates for key indicators (detailed estimates should be used with caution, especially for Tasmania, the Northern Territory and the Australian Capital Territory).

Quarterly estimates do not support all of these design targets due to the smaller sample size of each quarter.

### Final sample

The final sample for the September quarter was 3,669 households and 6,971 persons (aged 15 years and over).

### Collection method

Selected households completed the survey online or by telephone. One adult (aged 18 years and over) acted as the household's representative and answered questions about the household's financial situation (for example rent, rates, and loan payments) on behalf of the whole household. Each person, aged 15 years and over, then completed a personal interview, answering questions about their education, employment, income, and wealth.

### Questionnaire

The 2020-21 SIH questionnaire will be available in the survey's Concepts, Sources, and Methods release, which will be due for release in mid-2022. The 2017-18 SIH questionnaire, in the Survey of Income and Housing, User Guide, Australia, 2017-18 (cat. no. 6553.0), aligns closely with the 2020-21 questionnaire. New and modified questions to support monitoring of the COVID-19 pandemic were included in 2020-21.

## How the data is processed

Preliminary estimates involved similar but less extensive coding and editing checks than are performed on annual estimates. This includes treatment of statistical outliers and removal of households that did not provide sufficient principal information (see Survey of Income and Housing, User Guide, Australia, 2017-18 (cat. no. 6553.0)).

Some more intensive processing tasks were not performed in producing preliminary estimates. This has resulted in more sample being excluded from preliminary estimates, when compared to annual estimates. Survey weights (see below) have been calibrated to compensate for this additional loss.

The preliminary estimation method targeted key data items featured in this release and will be subject to revision when final estimates are produced.

### Estimation methods

Every person (including children aged 0 to 14 years) and every household in the sample were assigned a weight. Initial weights were based on probability of being selected in the sample. The person and household level weights were calibrated to align with independent estimates of the in-scope population, referred to as benchmarks.

The estimates were benchmarked to the estimated resident population living in private dwellings in non-very remote areas of Australia for each quarter, excluding persons living in discrete Aboriginal and Torres Strait Islander communities. The estimates, therefore do not match estimates of the total Australian resident population (which include persons living in very remote areas and non-private dwellings, such as hotels) obtained from other sources.

The benchmarks used additional information about the population to ensure that:

- people and households in the sample represented people and households that were similar to them
- the survey estimates reflected the distribution of the target population, not the sample.

The benchmarks used in the calibration of the final weights, for each quarter and state or territory, can be categorised into two groups:

Number of persons:

- by sex, in five-year age groups up to 80+ years for all states (excluding TAS, NT and ACT where age groups were collapsed further)
- by labour force status (except NT and ACT)
- by 2016 SEIFA Index for Relative Socioeconomic Disadvantage decile of household (state level).

Number of households:

- by household composition (number of adults (1, 2 or 3+) and whether or not the household contains children) (except NT and ACT which only uses whether or not the household contains children).

## Comparability

### Previous collections

The concepts and standards used for preliminary estimates are unchanged from previous SIH collections. Table 1 presents data from the September quarter 2019 and September quarter 2020 preliminary estimates.

### Australian System of National Accounts

Preliminary estimates have a similar level of coherence with the Australian System of National Accounts (ASNA) to that observed in previous SIH cycles.

In comparing SIH and ASNA, adjustments are necessary for each data source to account for scope and measurement differences. As the SIH and ASNA estimates of household income and wealth have been developed for different purposes, there are a number of differences in the resulting estimates. These are covered in further detail in the Survey of Income and Housing, User Guide, Australia, 2017-18 (cat. no. 6553.0) alongside previous cycle comparisons.

## Concepts and definitions

The concepts and definitions relating to income, wealth and housing statistics included in this release are described in Survey of Income and Housing, User Guide, Australia, 2017-18 (cat. no. 6553.0).

## Accuracy

### Reliability of estimates

Two types of error are possible in estimates based on a sample survey:

- non-sampling error
- sampling error.

### Non-sampling error

Non-sampling error is caused by factors other than those related to sample selection. It is any factor that results in the data values not accurately reflecting the true value of the population.

It can occur at any stage throughout the survey process. Examples include:

- selected people that do not respond (e.g. refusals, non-contact)
- questions being misunderstood
- responses being incorrectly recorded
- errors in coding or processing the survey data.

### Sampling error

Sampling error is the expected difference that can occur between the published estimates and the value that would have been produced if the whole population had been surveyed. Sampling error is the result of random variation and can be estimated using measures of variance in the data.

### Standard error

One measure of sampling error is the standard error (SE). There are about two chances in three that an estimate will differ by less than one SE from the figure that would have been obtained if the whole population had been included. There are about 19 chances in 20 that an estimate will differ by less than two SEs.

The relative standard error (RSE) is a useful measure of sampling error. It is the SE expressed as a percentage of the estimate:

\(R S E \%=\left(\frac{S E}{e s t i m a t e}\right) \times 100\)

Only estimates with RSEs less than 25% are considered reliable for most purposes. Estimates with larger RSEs, between 25% and less than 50% have been included in the publication, but are flagged to indicate they are subject to high SEs. These should be used with caution. Estimates with RSEs of 50% or more have also been flagged and are considered unreliable for most purposes. RSEs for these estimates are not published.

Another measure of sampling error is the Margin of Error (MOE). This describes the distance from the population value that the sample estimate is likely to be within and is particularly useful to understand the accuracy of proportion estimates. It is specified at a given level of confidence. Confidence levels typically used are 90%, 95% and 99%.

For example, at the 95% confidence level, the MOE indicates that there are about 19 chances in 20 that the estimate will differ by less than the specified MOE from the population value (the figure obtained if the whole population had been enumerated). The 95% MOE is calculated as 1.96 multiplied by the SE:

\({MOE} = S E \times 1.96\)

The RSE can also be used to directly calculate a 95% MOE by:

\({MOE}(y) \approx \frac{R S E(y) \times y}{100} \times 1.96\)

The MOEs in this publication are calculated at the 95% confidence level. This can easily be converted to a 90% confidence level by multiplying the MOE by:

\( {1.615\over 1.96}\)

or to a 99% confidence level by multiplying the MOE by:

\( {2.576\over 1.96}\)

Depending on how the estimate is to be used, an MOE of greater than 10% may be considered too large to inform decisions. For example, a proportion of 15% with an MOE of plus or minus 11% would mean the estimate could be anything from 4% to 26%. It is important to consider this range when using the estimates to make assertions about the population.

### Confidence intervals

A confidence interval expresses the sampling error as a range in which the population value is expected to lie at a given level of confidence. A confidence interval is calculated by taking the estimate plus or minus the MOE of that estimate. In other terms, the 95% confidence interval is the estimate +/- MOE.

### Calculating measures of error

Proportions or percentages formed from the ratio of two estimates are also subject to sampling errors. The size of the error depends on the accuracy of both the numerator and the denominator. A formula to approximate the RSE of a proportion is given below. This formula is only valid when the numerator (x) is a subset of the denominator (y):

\({RSE}\left(\frac{x}{y}\right) \approx\sqrt{[R S E(x)]^{2}-[R S E(y)]^{2}}\)

When calculating measures of error, it may be useful to convert RSE or MOE to SE. This allows the use of standard formulas involving the SE. The SE can be obtained from RSE or MOE using the following formulas:

\(SE = \frac{{RSE\% \times estimate}}{{100}}\)

\(SE = \frac{{MOE}}{{1.96}}\)

### Comparison of estimates

The difference between two survey estimates (counts or percentages) can also be calculated from published estimates. Such an estimate is also subject to sampling error. The sampling error of the difference between two estimates depends on their SEs and the relationship (correlation) between them. An approximate SE of the difference between two estimates (x - y) may be calculated by the following formula:

\(S E(x-y) \approx \sqrt{[S E(x)]^{2}+[S E(y)]^{2}}\)

While this formula will only be exact for differences between unrelated characteristics or sub-populations, it provides a reasonable approximation for the differences likely to be of interest in this publication.

When comparing estimates between surveys or between populations within a survey, it is useful to determine whether apparent differences are 'real' differences or simply the product of differences between the survey samples.

One way to examine this is to determine whether the difference between the estimates is statistically significant. This is done by calculating the standard error of the difference between two estimates (x and y) and using that to calculate the test statistic using the formula below:

\(\left(\frac{|x-y|}{S E(x-y)}\right)\)

where

\(S E(y)\approx\ \frac{R S E(y) \times y}{100}\)

If the value of the statistic is greater than 1.96, we can say there is good evidence of a statistically significant difference at 95% confidence levels between the two populations with respect to that characteristic. Otherwise, it cannot be stated with confidence that there is a real difference between the populations.

## How the data is released

### Release strategy

### Publications

This release includes preliminary quarterly indicators for the September 2020 quarter of SIH.

Estimates for the full financial year from this survey will be released within:

Preliminary estimates from the remaining quarters of 2020-21 of the SIH will be published in future quarterly household financial resources releases throughout 2021. Detailed publications and microdata for the full 2020-21 financial year will also be released from mid-2022

### Microdata

For users who would like to undertake more detailed analysis, a variety of microdata products from the 2020-21 SIH will be available from late 2022.

Information about current microdata services is available from the Microdata Entry Page.

### Confidentiality

The Census and Statistics Act 1905 authorises the ABS to collect statistical information, and requires that information is not published in a way that could identify a particular person or organisation. The ABS must make sure that information about individual respondents cannot be derived from published data.