|Page tools: Print Page Print All RSS Search this Product|
LABOUR FORCE SURVEY
The “other territories” of Australia, namely Jervis Bay, Christmas Island, Cocos (Keeling) Island, and Norfolk Island after the 2016 Census, are included in the estimated resident population of Australia, but excluded from household survey collection procedures and population benchmarks.
The “external territories” of Australia, namely Territory of Ashmore and Cartier Islands, Coral Sea Islands Territory, Australian Antarctic Territory, and Territory of Heard and McDonald Islands, are not included in the estimated resident population, household survey collection procedures or the population benchmarks.
Coverage rules are applied to ensure that each person is associated with only one household, and hence has only one chance of selection. The chance of a person being enumerated at two separate households in the one survey is considered to be negligible. Persons who are away from their usual residence for six weeks or less at the time of interview are enumerated at their usual residence (relevant information may be obtained from other usual residents present at the time of the survey).
The majority of LFS information is obtained by specially trained interviewers, using face-to-face and telephone interview collection methods, from the occupants of selected households. Interviews are generally conducted during the two weeks beginning on the Sunday between the 5th and the 11th of each month. Most information obtained relates to the week before the interview (referred to as the reference week). Selected households remain in the survey for eight consecutive months. Information about each household member in scope of the LFS is collected from one adult using the 'Any Responsible Adult' methodology (described in Chapter 16).
Prior to August 1996, all interviews were conducted face-to-face at selected households. Over the period August 1996 to February 1997, the ABS introduced telephone interviewing. The first interview is generally conducted in person (face-to-face), whilst subsequent interviews are conducted by telephone if this is acceptable to the respondent. Telephone interviewing has been shown to provide data of a quality comparable to that obtained from personal interviews, but requires less interviewer travel time, and hence lowers the costs of the survey.
From October 2003 to August 2004, computer assisted interviewing was progressively introduced for the LFS. Under computer assisted interviewing, interviewers record responses directly onto an electronic questionnaire in a notebook computer, rather than using the traditional 'pen and paper' method.
Online self-completion of the LFS was introduced in December 2012, with one survey group initially being offered the option of self-completing the survey online in place of a face-to-face or telephone interview. The online self-completion offer was expanded to 50% of private households in each new incoming survey group from May 2013 to August 2013. Between September 2013 and April 2014, the ABS expanded the offer of online self-completion from 50% to 100% of private households in each incoming survey group. Interviewer collection (both face-to-face and via telephone) continues to be available for those respondents where online self-completion is inappropriate for operational, technological or personal reasons. Online self-completion has been shown to provide data of a quality comparable to that obtained from interviewer-administered modes, and is the most cost-effective mode of enumeration.
Other collection methods are used in special circumstances. A paper self-enumeration form may be used where it is not possible for a computer assisted interview to take place - for instance, where contact cannot be made with the occupants of selected households or when a respondent refuses to be interviewed but will complete a form. A customised form is also used in very remote Indigenous communities.
Interviewer workloads are completed and returned for processing according to a strict timetable. Interviewers are required to make a number of attempts to contact a household before recording a non-contact (non-response).
The current LFS questionnaire, implemented in July 2014, is available from Appendix 2 or Information Paper: Questionnaires Used in the Labour Force Survey, July 2014 (cat. no. 6232.0), as a PDF file on the Downloads tab.
A multi-stage probability sample design is used. The sample is drawn from the Monthly Population Survey (MPS) Master Sample and has three components: a sample of private households, a sample of discrete Aboriginal and Torres Strait Islander communities, and a sample of non-private dwellings (i.e. hotels, motels, hospitals, retirement villages, etc.). The final stage selection unit is the household.
SIZE AND ALLOCATION
The LFS is designed to provide reliable estimates of the key labour force statistics for both the whole of Australia and each state and territory. The design also yields estimates for a number of broad regions within states.
Between February 1964 and February 1972, all households in Australia had the same probability of selection in the LFS (1 in 100), regardless of state or territory. From May 1972, different sampling fractions applied for each state and territory. The table below shows the sampling fractions for the LFS from the 1971 post-Census design through to the 1911 design.
Table 19.1: LFS Sample – Sampling Fractions
From 2018, the LFS sample is selected using information from the Address Register. The new sample has been designed with the aim of achieving similar levels of sampling error as the target levels established for the 2011 sample design.
Under the current sample design, the sampling fractions yield a sample size of approximately 26,000 households each month. This results in approximately 50,000 persons responding to the survey, covering about 1 in 312 (or 0.32%) of the civilian population aged 15 years and over. For further information, refer to Information Paper: Labour Force Survey Sample Design, May 2013 (cat. no. 6269.0).
There was a 24% reduction in the LFS sample size for the period July 2008 to August 2009, relative to the June 2008 sample size. The sample reduction was reinstated progressively between September and December 2009, with December 2009 estimates being the first produced under the fully reinstated sample. Extra care should be taken when using estimates from this period due to increased sampling error.
Since the monthly LFS commenced in 1978, one-eighth of the sample has been replaced each month. The sample can be thought of as comprising eight sub-samples (or rotation groups), with each sub-sample remaining in the survey for eight months. A new rotation group is introduced each month to replace an outgoing rotation group. This replacement sample generally comes from the same geographic area as the outgoing one.
Figure 19.1: Sample rotation
Sample rotation enables reliable measures of monthly change in labour force statistics to be compiled, as seven-eighths of the sample from any month is retained for the following survey. At the same time, the sample rotation procedure ensures that no household is retained in the sample for more than eight months, and that the sample reflects changes over time in the household population (such as construction of new households).
The component of the sample that is common from one month to the next makes it possible to match the characteristics of most of the persons in those households: this group is referred to as the 'matched sample'. The availability of this matched sample permits the production of estimates of 'gross flows' - the number of persons who change labour force status between successive months.
Until 2018, the LFS sample was selected every five years after each Census of Population and Housing to ensure that the survey continued to accurately reflect the socio-demographic distribution of the Australian population. From 2018, the sample frame using the Address Register will be updated every three years.
The estimation method used in the LFS is composite estimation, which was introduced in May 2007, with data revised back to April 2001. In January 2014, composite estimation was applied to all estimates from July 1991 as part of the historical revision to labour force data to reflect revised population benchmarks based on the 2011 Census. Composite estimation combines data collected in the previous six months with current month's data to produce the current month's estimates, thereby exploiting the high correlation between overlapping samples across months in the LFS. The composite estimation process combines the seven months of data by applying different factors according to length of time in the survey. These factors sum to unity for the current month, and once they are applied, the data are weighted to align with current month population benchmarks. For more information on composite estimation, see Information Paper: Forthcoming Changes to Labour Force Statistics, 2007 (cat. no. 6292.0).
Prior to the introduction of composite estimation, the estimation method used in the LFS was generalised regression, which only used the current month's data.
Estimates of the number of persons employed, unemployed and not in the labour force are calculated in such a way as to add up to independent estimates of the civilian population aged 15 years and over (population benchmarks) for age groups, sex and regions. There are two sets of benchmarks used in the LFS. The first set of benchmarks is classified by state or territory of usual residence, part of state of usual residence (capital city, rest of state), age and sex. The second set is classified by statistical region of usual residence and sex (known as 'regional benchmarks'). The use of regional benchmarks improves the quality of estimates for LFS regions, with negligible impact on estimates at national, state and territory levels.
Since the most recently released Estimated Resident Population (ERP) estimates lag the current time period for LFS estimates by 6-8 months, the population benchmarks are initially derived as short-term projections of the most recent ERP estimates. These projections are based on Census of Population and Housing data, adjusted for under-enumeration and updated for births, deaths, interstate migration, and net permanent and long-term migration. The short-term projections are based on the historical pattern of each population component - births, deaths, interstate migration and net overseas migration.
Prior to July 2010, the LFS population benchmarks were only revised every five years following the release of the final population estimates from the Census of Population and Housing. Benchmark revisions that incorporated Net Overseas Migration revisions to ERP were released in July 2010 (for the period July 2006 to June 2010) and November 2012 (for the period July 2008 to October 2012). For more information, refer to the article ‘Rebenchmarking of Labour Force Series’ in the November 2012 issue of Labour Force, Australia (cat. no. 6202.0).
From August 2015, the LFS population benchmarks have been revised quarterly (in February, May, August and November), with estimates revised for the previous 22 months to ensure that labour force series maintain coherence with the latest ERP estimates. This process ensures that the labour force population benchmarks are updated with the most recent ERP information available. For more information, refer to the article ‘Rebenchmarking Labour Force Estimates’ in the February 2015 issues of Labour Force, Australia (cat. no. 6202.0).
In addition, ERP rebasing is undertaken every five years to incorporate additional information from the latest available Census data. From December 2017, labour force estimates have been compiled using population benchmarks based 2016 Census-based population estimates.
TIME SERIES ESTIMATES
Both seasonally adjusted and trend estimates are produced for key series from this survey.
Many monthly series have been seasonally adjusted from February 1978, and are available in a range of products. Quarterly historical series (from August 1966) are available in Labour Force Historical Timeseries (cat. no. 6204.0.55.001). Concurrent seasonal adjustment was introduced from the December 2003 survey, replacing the annual forward factor method. At the same time, other improvements were made to the seasonal adjustment methodology to better handle the moving January interview start date, and the proximity of Easter to the April survey period.
Concurrent seasonal adjustment uses original data up to and including the current month (or quarter for quarterly series) to estimate seasonal factors for the current and all previous months (or quarters). Seasonally adjusted estimates from this method are usually closer on average to their final values, as any change in seasonality is observed sooner. The seasonal factors are further reviewed annually, to take account of each additional year’s original data. Revisions under this method are more frequent (every month for a monthly series), although the degree of revision is generally smaller than with the forward factor method of adjustment (where revisions are only made annually). For more information on concurrent seasonal adjustment, see Information Paper: Forthcoming Changes to Labour Force Statistics, 2003 (cat. no. 6292.0). For information on seasonally adjusted series, see Labour Force Survey Standard Products and Data Item Guide (cat. no. 6103.0) or Appendix 3.
Trend estimates are available for many series dating back to February 1978. Trend estimates are produced using a centred 13-term Henderson moving average of the seasonally adjusted series for monthly estimates, and a centred 7-term Henderson moving average for quarterly estimates (e.g. employment by industry). Centred symmetric moving averages cannot be used to directly estimate smoothed series values all the way to the end of the series, since there are insufficient observations available for the moving average calculations. The ABS uses non-symmetric moving averages to determine estimates of trend at the current end of the series. Revisions of trend estimates occur as data become available for later periods - these revisions are mainly because of the non-symmetric moving averages at the end of the series, but also because of concurrent seasonal adjustment. For further information, refer to Information Paper: A Guide to Interpreting Time Series - Monitoring Trends, 2003 (cat. no. 1349.0) and Labour Force Survey Standard Products and Data Item Guide (cat. no. 6103.0).
Trend estimates are derived by removing irregular influences from the seasonally adjusted estimates. As they do not include systematic, calendar related influences or irregular influences, trend estimates are the best measure of the underlying behaviour of the series, and the labour market.
ESTIMATES OF GROSS FLOWS
As a high proportion of the private households selected each month remain in the sample for the following month, it is possible to match the characteristics of most of the persons in those households from one month to the next. This makes it possible to record any changes in the labour force status of these persons, and hence to produce estimates of 'gross flows'. Gross flows provide information on the transition of individuals between the different labour force status classifications in successive months.
Prior to the July 2015 issue of Labour Force, Australia (cat. no. 6202.0), gross flow estimates related only to those persons in private households for whom information was obtained in two successive surveys. The procedures used to select persons in non-private households preclude the possibility of matching such persons who may be included in successive surveys. Also, the mobility of the population and non-response in either or both surveys means that a proportion of persons in private households who are included in the sample in successive months cannot be matched.
Overall, those who can be matched (in the private household sample) from one month to the previous month represent about 80% of all persons in the survey. About two-thirds of the remaining (unmatched) 20% are likely to have characteristics similar to those in the matched group, but the characteristics of the other third are likely to be somewhat different. The expansion factors (weights) used in calculating the estimates are those applying to the second of each pair of months. The estimates are not adjusted to account for the unmatched sample component.
From the July 2015 issue of Labour Force, Australia (cat. no. 6202.0), gross flow estimates have been expanded to include the 'unmatched' part of the sample for completeness (i.e. those who do not have a labour force status for the previous month, or those who do not have a labour force status for the current month). Weights for both the first and second months of each pair were also introduced. Gross flow estimates are available monthly in a data cube in Labour Force, Australia (cat. no. 6202.0) - see GM1 - Labour Force Statistics and Gross Changes (flows) by Sex, State and Age. For more information on the contribution of sample changes to gross flows, refer to the monthly summary ‘Insights from the Original Data’ in Labour Force, Australia (cat. no. 6202.0).
While every effort is made to reduce non-sampling errors to a minimum, any such errors affecting labour force status will tend to accumulate in the gross flow statistics. The estimates are also subject to sampling variability.
For more information on gross flows, see Labour Force Survey Standard Products and Data Item Guide (cat. no. 6103.0) or Appendix 3.
RELIABILITY OF THE ESTIMATES
Estimates from the survey are subject to both sampling and non-sampling error.
Sampling error occurs because a sample, rather than the entire population, is surveyed. The most commonly used measure of the likely difference resulting from not including all households in the survey is given by the standard error. Tables of standard errors of survey estimates are published each month in Labour Force, Australia (cat. no. 6202.0). The standard errors in these tables are mathematically modelled after each sample redesign, using many different estimates from several months of survey responses. Standard errors for other estimates and other movements may be calculated by using the spreadsheet contained in Appendix 3 or Labour Force Survey Standard Errors, Data Cube (cat. no. 6298.0.55.001). Further information about sampling error is available in Chapter 16.
Non-sampling error arises from inaccuracies in collecting, recording and processing the data. Every effort is made to minimise reporting error by the careful design of questionnaires, intensive training and supervision of interviewers, and efficient data processing procedures. Non-sampling error also arises because information cannot be obtained from all persons selected in the survey. The LFS receives a high level of co-operation from individuals in selected households, with response rates averaging around 93%. Further information about non-sampling error is available in Chapter 16.
Data are compiled according to concepts and definitions outlined in Chapters 2 to 8. Estimates are published monthly in Labour Force, Australia (cat. no. 6202.0). More detailed data are published one week later in Labour Force, Australia, Detailed - Electronic Delivery (cat. no. 6291.0.55.001) and in the quarterly release Labour Force, Australia, Detailed, Quarterly (cat. no. 6291.0.55.003). Other detailed estimates may be available on request.
An expanded Confidentialised Unit Record File (CURF) for the years 2008 to 2010 is available via the Remote Access Data Laboratory (RADL). This Longitudinal CURF is created from the monthly Labour Force Surveys along with data collected from labour supplementary surveys and multipurpose household surveys conducted between January 2008 and December 2010. It includes a range of data to enable users to better understand the dynamics of the labour market and transitions between employment, unemployment and not in the labour force. For more information on CURF, see Microdata: Longitudinal Labour Force, Australia, 2008-10 (cat. no. 6602.0).
Historical estimates can be found in Labour Force Historical Timeseries, Australia (cat. no. 6204.0.55.001). Family estimates are currently published in Labour Force Status and Other Characteristics of Families (cat. no. 6224.0.55.001). Estimates on Aboriginal and Torres Strait Islander Australians are currently published in National Aboriginal and Torres Strait Islander Social Survey, 2014-15 (cat. no. 4714.0)
The LFS survey output includes:
Persons in the labour force
Labour force status; unemployment rate; labour underutilisation rate; labour force participation rate; gross flows (changes) in labour force status; volume measures of underutilisation*; and retrenchments in the previous quarter*.
Persons not in the labour force
Seasonally adjusted and trend (i.e. smoothed seasonally adjusted) data are available for selected series including labour force status; unemployment; labour force participation rate; industry of employment; and long term unemployed. Seasonally adjusted data are not available for the Northern Territory or the Australian Capital Territory. See Chapter 16 for further explanation of these terms.
* These data are released for the months of February, May, August and November only.
BACK TO TOP
These documents will be presented in a new window.