QUALITY DECLARATION


INSTITUTIONAL ENVIRONMENT

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, see ABS Institutional Environment.

In April 2012, the ABS became an accredited Integrating Authority under the Commonwealth data integration interim arrangements. A copy of the accreditation claims made by the ABS, which have been verified by an independent auditor, is available on data.gov.au. The ABS only undertakes data integration for statistical and research purposes and where there is a strong public benefit in doing so.

The Australian Census Longitudinal Dataset (ACLD) is released in TableBuilder and as a microdata product in the DataLab. Microdata files are released in accordance with the conditions specified in the Statistics Determination section of the Census and Statistics Act 1905. This ensures that confidentiality is maintained whilst enabling micro level data to be released. More information on the confidentiality practices associated with TableBuilder can be found in TableBuilder, User Guide (cat. no. 1406.0.55.005) on the Confidentiality page. To protect confidentiality of data within the DataLab, users are supervised at all times and must not bring mobile phones, cameras, USB keys, laptops, palm pilots or similar transmission or storage devices into the secure location. All outputs produced by users in DataLab are manually cleared for release after the session.


RELEVANCE

Data for the Census of Population and Housing used in this product were collected on 8 August 2006, 9 August 2011 and 9 August 2016. The scope of the Census is all persons enumerated in Australia on Census night. The Census covers all areas in Australia and includes persons living in both private and non-private dwellings but excludes:

• diplomatic personnel of overseas governments and their families
• Australian residents overseas on Census Night and

The ACLD is built upon a 5% sample of records taken from a particular Census that is then linked to following Censuses. There are currently two samples, 2006 and 2011, with each being representative of the Australian population at the time of the Census collection.

Overseas visitors are excluded from the 2006 and 2011 ACLD Panel samples. Visitors within Australia to private and non-private dwellings on Census Night are included

The Census collects information on demographics, income, labour force, unpaid work, dwelling characteristics and family and household relationships.

For more information, see How Australia Takes a Census, 2006 (cat. no. 2903.0), How Australia Takes a Census, 2011 (cat. no. 2903.0), Census of Population and Housing: Understanding the Census and Census Data, Australia, 2016 (cat. no. 2900.0), and the 2006, 2011 and 2016 issues of the Census Dictionary (cat. no. 2901.0).


TIMELINESS

The Census of Population and Housing is conducted every five years. For further information see the publications How Australia Takes a Census, 2006 (cat. no. 2903.0), How Australia Takes a Census, 2011 (cat. no. 2903.0) and Census of Population and Housing: Understanding the Census and Census Data, Australia, 2016 (cat. no. 2900.0).

The first wave of Census data for the ACLD was from 2006, the second wave was from 2011, and the third wave was from 2016.

Microdata from the 2006-11 ACLD was first made available in December 2013. The 2011-16 data was available from February 2018, and the 2006-11-16 data available from March 2019.


ACCURACY

The ACLD is a random 5% sample of persons enumerated in Australia on either Census Night, 2006 or Census Night, 2011 which has been linked using statistical techniques to records from successive Censuses. False links can occur during the linkage process as even when a record pair matches on all or most linking fields, it may not actually belong to the same individual. The nature of the process used for the ACLD linkage means that while the methodology is designed to ensure links obtained are to a high degree of accuracy, some false links may be present within the ACLD dataset. There is an estimated 5-10% false link rate in the original linkage of the 2006-2011 ACLD, an estimated 5% false link rate in the re-link of the 2006-2011 ACLD and an estimated 1% false link rate in the 2011-2016 linkages.

Sampling error occurs because only a small proportion of the total population is used to produce estimates that represent the whole population. Sampling error refers to the fact that for a given sample size, each sample will produce different results, which will usually not be equal to the population value. There are two common ways of reducing sampling error - increasing sample size and/or utilising an appropriate selection method (for example, multi-stage sampling would be appropriate for household surveys). Given the large sample size for the ACLD (1 in 20 persons), and simple random selection, sampling error is minimal.

The ACLD sample was weighted to an estimate of the population that was resident in Australia on Census Night for the relevant linkage periods. For example, the linkage of the 2011 Panel to the 2016 Census is weighted to an estimate of the population that was resident in Australia at both the 2011 and 2016 Censuses. The weights adjust for missed links and Census undercount.

Information on methodology, linkage quality and weighting can be found in Information Paper: Australian Census Longitudinal Dataset, Methodology and Quality Assessment, ACLD (cat. no. 2080.5). Steps are taken to confidentialise the data made available on TableBuilder in such a way as to maximise the usefulness of the content while maintaining the confidentiality of respondents selected in the ACLD sample. As a result it may not be possible to exactly reconcile all the statistics produced from the microdata with other published statistics. Further information about the steps taken to confidentialise the microdata can be found in TableBuilder, User Guide (cat. no. 1406.0.55.005) on the Confidentiality page.


COHERENCE

A small percentage of linked records have inconsistent data, such as a different country of birth at the two time points or an age inconsistency of more than one year. Inconsistencies may be due to:

• false link - the record pair does not belong to the same individual
• reporting error - information for the same individual was reported differently in 2011 and in 2016
• processing error - the value of a data item was inaccurately assigned or imputed during processing.

ACLD microdata contains a large number of data items and in some cases the level of detail has been collapsed from that described in the Census Dictionary. For more information on the level of detail provided, please see the associated Data Items list.

While the 2011 and 2016 Censuses had predominantly the same questions and were processed in a similar way, there were some differences between them.

For example, a number of changes were made to how industry of employment information was collected for the 2016 Census. The ABS advises this data is not directly comparable to the previous Census Industry of employment data, and should not be used to measure longitudinal transitions between industries from 2011 to 2016. For further information refer to Industry of Employment (INDP) in Census of Population and Housing: Understanding the Census and Census Data, Australia, 2016 (cat. no. 2900.0).

Notable data items that are different between Census years are personal, family and household income. Income was collected in ranges and these ranges are different in different Census years. The ACLD does not include an adjustment to income data for inflation.

Some data items were derived differently between Censuses. In these instances, to aid comparability, the 2006 and 2011 variables were re-derived to make them consistent with the 2016 derivation.

For more information on the differences between the 2006, 2011 and 2016 Censuses see What's New for 2011? and What's New for 2016?

Estimates derived from the ACLD may differ to those derived from other sources. This is due to a range of factors including:

• Collection methodology
  The ACLD is derived from Census data that is self-reported by households across Australia on Census night. This will differ from other ABS collections which may rely on different collection methodologies (e.g. trained interviewers, administrative sources). In addition, the way survey questions are phrased and the answer options available for a given question may affect the information provided by respondents.

• Reference period
  The reference periods for the ACLD are the Census nights of each year. Other collections may use different reference periods.

• Sampling design
  The ACLD uses a 5% sample of Census data as its base population. This will differ from other collections that may collect information from the entire population of Australia (e.g. the Census) or from a sample of dwellings (e.g. Labour Force Survey).

• Sampling and non-sampling error
  While every effort is made to minimise error, each collection will have some level of error. Survey collections are subject to some level of sampling error, as they are based on information obtained from a sample of dwellings or businesses. The Census is not subject to this type of error, but is subject to some level of undercount. The ACLD is constructed using a sample of records from the Census, and is therefore subject to a level of sampling error of its own.

• Scope and coverage
  The ACLD weights benchmark the linked records to the longitudinal population that was in scope of both or all three Censuses. This will be different to cross-sectional estimates which may be benchmarked to a point-in-time population, such as the Estimated Resident Population.

• Linkage error
  The ACLD is subject to linkage error, as records from one Census are linked to corresponding records from the subsequent Census. While every effort is made to minimise false links, they can occur. Linkage error will not be apparent in other collections which are not produced through data integration.
  

For these reasons, while the results from the ACLD are considered to be broadly representative of the Australian population, they are not strictly comparable with statistics derived from other collections.

For detailed information about the different methodologies for each collection, refer to the Explanatory Notes within each release.

For detailed information regarding the differences between the Census and Labour Force collections, refer to The 2016 Census and the Labour Force Survey in Census of Population and Housing: Understanding the Census and Census Data, Australia, 2016 (cat. no. 2900.0).

For detailed information regarding Census data, including changes to Census questions and data quality statements for each Census data item, refer to Understanding the data in Census of Population and Housing: Understanding the Census and Census Data, Australia, 2016 (cat. no. 2900.0).


INTERPRETABILITY

This publication should be referred to when using the microdata. It contains information on the Methodology, File Structure, Using the ACLD in TableBuilder, The ACLD in the DataLab, Conditions of Use and the Data Items list.

Detailed information on methodology, linkage quality and weighting can be found in Information Paper: Australian Census Longitudinal Dataset, Methodology and Quality Assessment, ACLD (cat. no. 2080.5). The ABS publishes extensive information on Census Data Quality.


ACCESSIBILITY

The Australian Census Longitudinal Datasets, 2006-11, 2011-16 and 2006-11-16 can be accessed through TableBuilder and the DataLab.

These microdata products are available to approved users. Users wishing to access the microdata should read the How to apply for Microdata web page, before applying for access by emailing microdata.access@abs.gov.au. Users should also familiarise themselves with information available via the Microdata Entry Page.

Any questions regarding access to microdata can be forwarded to microdata.access@abs.gov.au or phone (02) 6252 7714.