2033.0.55.001 - Census of Population and Housing: Socio-Economic Indexes for Areas (SEIFA), Australia, 2011 Quality Declaration 
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 28/03/2013   
   Page tools: Print Print Page Print all pages in this productPrint All

QUALITY DECLARATION - SUMMARY

INSTITUTIONAL ENVIRONMENT


The Socio-economic Indexes for Areas (SEIFA) were created using data collected in the Census of Population and Housing (Census), which is conducted every five years. Refer to the Census Quality Declaration for more information on the Census.

For information on the institutional environment of the Australian Bureau of Statistics (ABS), including the legislative obligations of the ABS, financing and governance arrangements, and mechanisms for scrutiny of ABS operations, please see ABS Institutional Environment.


RELEVANCE

SEIFA 2011 is a suite of four summary measures (indexes) that were created using information from the 2011 Census. The four indexes in SEIFA 2011 are:

  • Index of Relative Socio-economic Disadvantage (IRSD)
  • Index of Relative Socio-economic Advantage and Disadvantage (IRSAD)
  • Index of Economic Resources (IER)
  • Index of Education and Occupation (IEO).

Each index summarises a different aspect of the socio-economic conditions in an area, and were therefore created using a different set of information. For a list of variables used in each index, refer to the IRSD, IRSAD, IER and IEO variable lists. For more detail on the information collected in the Census, refer to "Census Dictionary, 2011 (cat. no. 2901.0)".

A SEIFA index lists every geographic area in Australia, along with a set of numbers that describe how relatively advantaged or disadvantaged each area is compared to the other areas. This set is comprised of numbers that can be used for comparisons with other areas in Australia : a score, a rank, a percentile and a decile. There is also a set of numbers that can be used for comparisons with other areas within the same state/territory (created from the score) : a State-rank, a State-percentile, a State-decile. Note that all of these numbers are specific to an index, so that the numbers from one index cannot be directly compared to those of a different index, or to a previous release.

A lower number means that an area is relatively more disadvantaged than other areas. A higher number for the IRSD means that an area is relatively less disadvantaged than other areas; a high number for the remaining indexes means that an area is relatively more advantaged than other areas.

SEIFA indexes were created for six types of geography: Statistical Area Level 1 (SA1), Statistical Area Level 2 (SA2), Postal Area (POA), Statistical Local Area (SLA), State Suburb (SSC), and Local Government Area (LGA). The indexes were constructed from Census data at the SA1 level; scores for larger areas were created by taking a population-weighted average of the SA1s within each larger area.

A SEIFA score is a summary measure of the usual residents in an area. Because areas are diverse, a score may not reflect the socio-economic conditions of a person or household within the area. We recommend using SEIFA at the smallest geographic area possible for your analysis, preferably the SA1 level. SEIFA indexes were not created for very large geographies. Additional spreadsheets have been created to show the distribution of SA1 scores in large geographies: Statistical Area Level 2 (SA2), Statistical Area Level 3 (SA3), Statistical Area Level 4 (SA4), Local Government Area (LGA), Statistical Local Area (SLA), State Suburb (SSC), Postal Area (POA), Commonwealth Electoral Division (CED), and State Electoral Division (SED).

Some individual geographic areas were excluded from SEIFA for various reasons, such as low population or high non-response to certain Census questions.

While consistency (with past releases) in the construction of SEIFA is important, changes were made where important or necessary. SEIFA may also be impacted by external changes, such as changes made to standards, classifications and geographies.

Due to the complexity of the SEIFA product, we strongly recommend users refer to the supporting documentation described in the Accessibility section below.


TIMELINESS

SEIFA is created using information from the Census, which is conducted every five years. The 2011 Census was conducted on 9 August 2011. The SEIFA indexes are created after the second release of Census data. SEIFA 2011 was released on 28th March 2013.


ACCURACY

It is important to be aware that SEIFA is a general measure of relative socio-economic advantage and disadvantage, and to use SEIFA appropriately. Deciles are appropriate for most uses, particularly as scores may imply a level of accuracy that is not provided by a general measure of relative socio-economic advantage and disadvantage like SEIFA. Ranks and scores are appropriate only for more technical use.

SEIFA is a complex product and there are many issues in terms of accuracy.

SEIFA is created using Census data. Refer to Census documentation for further information on data-item questions, standards and classifications used, non-response codes and imputation methods for Census data.

The ABS broadly defines relative socio-economic advantage and disadvantage in terms of people's access to material and social resources, and their ability to participate in society. For SEIFA, four different indexes are created to try to capture different aspects of this complex concept. Each index is a summary of a different subset of variables.

SEIFA is an area-level measure and is not intended to reflect individual people or households. A SEIFA score is an average; a summary measure of the people and households within an area, and may not necessarily represent the individual status of a person or household within the area.

A SEIFA score is a relative measure, and cannot be used to say that an area "is disadvantaged", only that it is disadvantaged relative to other areas in Australia.

After variables were selected based on the concept of relative socio-economic advantage and disadvantage, the SEIFA indexes were created using a method called Principal Component Analysis. Under this method, the weights for each variable were determined by the complex interaction between the variables themselves, and were not predetermined.

An index like SEIFA is reliant on the information selected and methodology used to create the index. Census information is carefully selected and specified. The indexes are created using this information, then examined using a range of methods. The relationship between the indexes themselves, and with other measures, were analysed. The sensitivity of the indexes to outliers and variables was also analysed. The SEIFA indexes as a whole were found to be generally robust. However, individual areas, groups of areas, or sub-populations may be more sensitive. Refer to the supporting documentation for more detail on validation.

As SEIFA is created from the Census rather than a sample survey, measures of sampling error are not relevant, however other diagnostics and supporting information were provided. For the SA1-level indexes, the scores have been standardised to a mean of 1000 and standard deviation of 100 to assist interpretation. The creation of four indexes allows users to test the sensitivity of their analysis to different indexes. Ranks, deciles and percentiles have been provided for users to test the sensitivity of their results to different analytical methodologies. Minimum and maximum SA1 scores were provided for the larger geographic areas: SA2s, POAs, SSCs, SLAs, and LGAs. Areas with too little information were excluded from SEIFA. The proportion of the population in excluded SA1s were provided for SA2s, POAs, SSCs, SLAs, and LGAs. Distribution graphics were provided for SA3s, SA4s, LGAs, CEDs and SEDs, whilst distribution tables were provided for SA2s, SA3s, SA4s, SLAs, POAs, LGAs, SSCs, CEDs and SEDs. In addition, supporting documentation has been provided to assist users to understand and use SEIFA, including recommendations and examples of appropriate use.

Due to the complexity of the SEIFA product, we strongly recommend users refer to the supporting documentation described in the Accessibility section below.


COHERENCE

An index similar to SEIFA was first produced using information from the 1971 Census. Substantial development of the indexes occurred in the late 1980s. SEIFA was first produced in its present form in 1990 using information from the 1986 Census, and consisted of five indexes. For 2001 SEIFA, the variable selection process was revised and two indexes (the Urban and Rural indexes) were replaced with a single index (the Index of Relative Socio-economic Advantage and Disadvantage). The same four indexes have been released for 2011 SEIFA.

While consistency in the construction of SEIFA 2011 was important, changes were made where important or necessary. SEIFA may also have been impacted by external changes, such as changes made to standards, classifications and geographies. The major changes are overviewed below and discussed further in the supporting documentation:
  • New Census data items were considered for inclusion in the indexes. After analysis, one newly included measure related to children living in jobless families. Variables on people living in improvised dwellings, people living in supported accommodation and young people not involved in education or employment were analysed but not included. Also, some existing Census data items not previously included in SEIFA were reconsidered for inclusion in some 2011 SEIFA indexes. For example, a variable relating to dialup internet access was considered in the 2011 Index of Relative Advantage and Disadvantage.
  • Some variables have had some changes to definitions in order to better capture the concept they are trying to measure. These altered variables related to housing tenure and educational attainment.
  • Some SEIFA variables involve a cut-off value to determine relatively 'high' or 'low' values of measures such as rental payments or income. These cut-off values were revised based on deciles/quintiles in the 2011 Census data.
  • The rules for excluding areas were updated to provide a small but positive refinement to the final list of areas receiving a score.
  • SEIFA 2011 is released according to the Australian Statistical Geography Standard (ASGS). This is a change from past versions of SEIFA, which used the Australian Standard Geographical Classification (ASGC). The main implication for SEIFA from this change is that the new base unit of analysis is the Statistical Area Level 1 (SA1), rather than the Census Collection District (CD) used in the past.

Caution is urged if comparing between SEIFA releases, for the above reasons and for the reasons below:
  • SEIFA is based on information measured at a certain point in time. The index weights are based on the complex interactions between this information. Because the world changes, not only does the data change, but the weights also change, resulting in changes to the indexes.
  • The scores are standardised to a mean of 1000 and standard deviation of 100 before release, which means that a score or rank cannot be directly compared with previous releases (e.g. a score of 800 does not represent the same level of disadvantage in different years).
  • A change in the rank of a particular area is not necessarily a reflection of changes within that area, but may be due to other areas becoming more/less disadvantaged.

For these reasons, comparing SEIFA over time can lead to misleading results. If you must compare the SEIFA score of an area over time, we recommend the use of deciles, rather than ranks or scores, particularly as scores may imply a level of accuracy that is not provided by a general measure of relative disadvantage like SEIFA. Scores are appropriate only for more technical use. Further, we recommend that the extreme deciles be compared, because these tend to be more stable over time.

Due to the complexity of the SEIFA product, we strongly recommend users refer to the supporting documentation described in the Accessibility section below.


INTERPRETABILITY

SEIFA numbers are ordinal; that is, they can only be used to compare one area to other areas in Australia. A lower number means that an area is relatively more disadvantaged than other areas. The Index of Relative Socio-economic Disadvantage (IRSD) includes only indicators of disadvantage, and therefore should not be used to identify areas of high advantage. The other three indexes contain indicators of advantage as well as disadvantage.
Interpretation of Deciles: IRSD

Interpretation of Deciles: IRSAD, IER, IEO

Due to the complexity of the SEIFA product, we strongly recommend users refer to the supporting documentation described in the Accessibility section below; in particular, Section 1 (Introduction) and Section 6 (Using and Interpreting SEIFA) of the Technical Paper.


ACCESSIBILITY


SEIFA indexes are available for every Census, from the 1986 Census onwards. The indexes are available free of charge on the ABS website under:

Census of Population and Housing: Socio-Economic Indexes for Areas (SEIFA), Australia (cat. no. 2033.0.55.001)

Further information for SEIFA 2011 is available in the Technical Paper under the 'Downloads' tab.

For all inquiries please refer to the Census Contacts page, or contact the National Information Referral Services on 1300 135 070 from within Australia or +61 2 9268 4909 from overseas. Alternatively, please email client.services@abs.gov.au.