Socio-Economic Indexes for Areas (SEIFA), Australia methodology

This publication outlines how SEIFA is constructed and provides guidance on how it should be used and understood.

See SEIFA for general information on SEIFA 2021, and to access the data. 

For detailed technical specifications refer to the SEIFA Technical Paper.

SEIFA is a collection of four indexes, each summarising a different aspect of the socio-economic conditions in an area using different Census data:

• the Index of Relative Socio-economic Advantage and Disadvantage (IRSAD) focuses on both advantage and disadvantage
• the Index of Relative Socio-economic Disadvantage (IRSD) focuses on relative socio-economic disadvantage
• the Index of Education and Occupation (IEO) focuses on relative Education and Occupation advantage and disadvantage
• the Index of Economic Resources (IER) focuses on Economic advantage and disadvantage.

The same area may score differently for each index due to their constituent variables.

For SEIFA purposes, relative socio-economic advantage and disadvantage is defined as people's access to material and social resources, and their ability to participate in society. The dimensions are constrained by what is collected in the Census; they include income, education, employment, occupation, housing, and family structure.

SEIFA is designed and constructed as an area-based measure. Area-based deciles are calculated by dividing the areas, ordered by disadvantage, into ten equally sized groups. Decile one contains the most disadvantaged areas and decile ten contains the most advantaged areas.

The SEIFA scores are initially calculated by Statistical Area Level 1 (SA1).

Scores for higher level geographic areas, such as Statistical Area Level 2 (SA2), Local Government Areas (LGA), Suburbs and Localities (SAL), and Postal Areas (POA) are calculated using population-weighted averages of the SA1 scores. This ensures that the higher-level geographic area scores reflect the size of the population within each contributing SA1.

For a detailed discussion on SEIFA’s construction, refer to the SEIFA Technical Paper.

Variables that relate to the concepts of advantage and disadvantage that SEIFA is trying to capture are included in the initial variable list. The main constraint on the initial variable list is that the variables can only be derived from Census data.

The indexes are calculated from 2021 Census data, with variables calculated as proportions at the SA1 level. For example, the proportion of unemployed people is calculated as the number of unemployed people divided by the total number of people in the labour force in each SA1.

Due to non-response and low population counts, not every SA1 can be given a meaningful index score. ABS maximises the number of areas that receive a score, while maintaining confidentiality and an acceptable level of quality. SA1s do not receive an index score if:

• the SA1 has no usual addresses
• the usual resident population is less than or equal to Ten
• the area is classified as offshore or migratory (refer to special purpose codes)
• there are fewer than six employed people
• there are fewer than six classifiable occupied private dwellings
• the proportion of people in private dwellings is less than or equal to 20 per cent
• the denominator of a variable in the index is less than six.

The indexes are a weighted combination of Census derived variables. The method used to determine the weights and final variable composition is known as Principal Component Analysis (PCA). The weights describe the relative importance of each variable. If the weight determined by PCA for a variable is too low, the variable would not meaningfully contribute to the index, so is excluded.

The variable weights are derived from a consistent, data-driven method and are not determined subjectively.

Once the variable weights have been finalised, they are used together with the variable proportions to create index scores for SA1s across Australia.

The indexes are checked to ensure they are measuring the desired concept and that the results make intuitive sense. Validation includes:

• investigating the correlations between the four indexes
• comparing SEIFA 2021 rankings with 2016 rankings
• identifying which variables are the drivers of change between SEIFA 2016 and 2021
• peer review of the results.

SEIFA 2021 uses edition 3 of the Australian Statistical Geography Standard (ASGS).

• SEIFA 2021 largely uses the same method as SEIFA 2016.
• Where available, SEIFA 2021 uses the same candidate variables as SEIFA 2016.
• The Dwelling Internet Connection data item was not collected in the 2021 Census, so the NONET variable could not be used.
• Occupation variables are based on the Australian and New Zealand Standard Classification of Occupations (ANZSCO).
• Variables using cut-off values as part of their specification, such as high and low income (INC_HIGH and INC_LOW), were updated.
• The only processing change was to start from confidentialised (instead of unconfidentialised) area level counts. Using confidentialised area level counts did not have a significant impact on the results, and appropriate comparisons to 2016 results are still valid.

The indexes rank areas according to their relative socio-economic conditions and are useful in understanding the distribution of socio-economic conditions across different areas.

It is recommended that index rankings or quantiles (deciles and percentiles) are used for analysis, rather than index scores.

The indexes reflect the socio-economic characteristics of an area, rather than of individuals.

They are calculated at the SA1 level and reflect SA1 characteristics.

Because people within an SA1 are not identical, the index scores for an SA1 do not directly reflect the relative advantage or disadvantage of an individual residing in that SA1. For example, it is possible for a relatively advantaged person to live in an area with a low score or for a relatively disadvantaged person to live in an area with a high score.

Refer to SEIFA: Getting a handle on individual diversity within areas for more information.

The variables that contribute to each index should be considered when deciding which index to use (see sections on each index for lists of variables). For example:

• if you are looking for areas of disadvantage for the purpose of allocating funds or services, the Index of Relative Socio-economic Disadvantage would be the most appropriate
• if you are looking for areas with relatively high populations in unskilled jobs or with relatively low educational qualifications, the Index of Education and Occupation would be most suitable.

The indexes are designed to compare the relative socio-economic characteristics of areas at a point in time, not to compare areas over time. There are several issues that make longitudinal or time series analysis of SEIFA difficult to interpret.

• The constituent indicators and indicator weights for each index are likely to have changed. 
• The geographic boundaries and numbers of relevant small areas may have changed. 
• The distribution of the standardised index values will have changed (e.g., a score of 800 does not represent the same level of disadvantage in different years). 
• There are changes in the way the indicators are defined.

If comparisons over time are made, the use of deciles or percentiles is recommended rather than ranks or scores.

Topics represented in SEIFA are limited to what is collected in the Census.

Measures relating to wealth and infrastructure may provide more information about the relative advantage or disadvantage within an area, but these measures are not collected in the Census.

Long-term health conditions, asked for the first time in the 2021 Census, were not included in SEIFA. This is to allow health researchers to analyse the relationship between health outcomes and socio-economic advantage/disadvantage. Adding health variables to SEIFA would make these relationships less clear.

Other potential topics that could be associated with advantage and disadvantage but are not captured in the Census include crime, and the environment.

Scores

SA1 scores are created by adding together the weighted characteristics of that SA1. The scores for all SA1s are then standardised to a distribution where the average equals 1,000 and the standard deviation is 100.

Scores for larger geographic areas, such as Statistical Area Level 2 (SA2) and Local Government Area (LGA), are calculated using population-weighted averages of the SA1 scores.

Lower scores indicate areas of relative disadvantage compared to areas with higher scores.

Scores are an ordinal measure on an arbitrary scale and do not represent the quantity of advantage or disadvantage (i.e. it’s not accurate to say an area with a score of 1,000 is twice as advantaged as an area with a score of 500).

Rankings and quantiles (deciles and percentiles) should be used in most cases. Scores should be used for more technically complex analysis. For more information see Technical Paper.

Ranks

Every area is ordered from the lowest to highest score, with the area with the lowest score given a rank of one, the area with the second lowest score given a rank of two and so on, up to the area with the highest score which is given the highest rank.

Deciles

Every area is ordered from lowest to highest score, with the lowest 10 per cent of areas given a decile number of one, the next lowest 10 per cent of areas given a decile number of two and so on, up to the highest 10 per cent of areas which are given a decile number of 10. This means that areas are divided up into 10 equal sized groups, based on their score.

Percentiles

Every area is ordered from lowest to highest score, with the lowest one per cent of areas given a percentile number of one, the next lowest one per cent is given a percentile number two and so on, up to the highest one per cent of areas which are given a percentile number of 100. This means that areas are divided up into one hundred equal sized groups, based on their score.

Ranking within state or territory (rank, decile, and percentile)

Every area within a state or territory is ordered from lowest to highest score, then ranks, deciles and percentiles are assigned to each area within that state or territory.

State/territory ranks, deciles and percentiles can only be used to compare areas within a single state or territory.

This section outlines the variables used in the Index of Relative Socio-economic Advantage and Disadvantage (IRSAD).

The following variables are indicators of disadvantage. NOYR12ORHIGHER is the strongest indicator of disadvantage in the index.

The following variables are indicators of advantage. INC_HIGH is the strongest indicator of advantage in the index.

The following variable was initially considered for the index but was excluded due to being highly correlated with other variables.

The following variables were initially considered for the index but were excluded when the analysis showed that they were weak indicators of relative advantage or disadvantage.

For more information on the IRSAD variables, refer to the Technical Paper.

This section outlines the variables used in the Index of Relative Socio-economic Disadvantage (IRSD).

All variables in this index are indicators of disadvantage. INC_LOW is the strongest indicator of disadvantage in the index.

The following variables were initially considered for the index but were excluded when the analysis showed that they were weak indicators of relative disadvantage.

For more information on the IRSD variables, refer to the Technical Paper.

This section outlines the variables used in the Index of Education and Occupation (IEO).

The following variables are indicators of disadvantage. NOYR12ORHIGHER is the strongest indicator of disadvantage in the index.

The following variables are indicators of advantage. OCC_SKILL1 is the strongest indicator of advantage in the index.

The following variable was initially considered for the index but was excluded due to being highly correlated with other variables.

The following variables were initially considered for the index but were excluded when the analysis showed that they were weak indicators of relative advantage or disadvantage.

For more information on the IEO variables, refer to the Technical Paper.

This section outlines the variables used in the Index of Economic Resources (IER).

The following variables are indicators of disadvantage. INC_LOW is the strongest indicator of disadvantage in the index.

The following variables are indicators of advantage. HIGHBED is the strongest indicator of advantage in the index.

The following variable was initially considered for the index but was excluded when the analysis showed that it was a weak indicator of relative disadvantage.

For more information on the variables, refer to the Technical Paper.