Survey weights are calibrated against population benchmarks to ensure that the survey estimates conform to the independently estimated distribution of the population, rather than the distribution within the sample itself.
When calibrating the weights, the survey sample is grouped into categories based on the following characteristics:
- State or territory
- Capital city or rest of state
- Sex
- Age
- Employed full-time, part-time, unemployed or not in the labour force.
The Labour Force Survey estimates are calculated in such a way as to sum to the independent estimates of the civilian population aged 15 years and over (population benchmarks). These population benchmarks are updated quarterly based on Estimated Resident Population (ERP) data. See Labour Force, Australia, Methodology for more information.
From August 2015, Labour Force estimates have been compiled using population benchmarks based on the most recently available release of ERP data, continually revised on a quarterly basis.
The B&I benchmarks were based on a 12-month average of the LFS estimates for the June to July financial year, as reported in the April 2023 issue of Labour Force, Australia. This approach is used to remove the seasonality from the employed, unemployed and not in the labour force benchmarks and to improve coherence between the two publications.
Estimates from previous surveys back to 2014-15 have also been revised using this method, with benchmarks based on the same population series (as at April 2023). These estimates were calibrated to population benchmarks based on revisions to ERP that incorporated the results of the 2021 Census (introduced to LFS in the November 2022 issue).
Quarterly benchmarks
The benchmarks for quarterly estimation were based on the same method as for the full financial year sample, using the corresponding 3-month average of the LFS estimates for each quarter. This shorter time period does not account for seasonality in the same way as using a 12-month average, so it is expected that some seasonality remains in the estimates.
The weights for the reduced quarterly sample were calibrated based on grouping the sample using broader characteristics than the full financial year sample, notably excluding State and territory from the calibration. This means that State and territory estimates are not available on a quarterly basis.
The LFS estimates used in benchmarking the quarterly estimates were also taken from the August 2023 issue of Labour Force, Australia.
Microdata in DataLab
The financial year and quarterly weights (and replicate weights used for calculating relative standard errors) provided on the most recent B&I microdata file in ABS DataLab have both been based on benchmarks using the 12-month and 3-month averages of the LFS estimates as reported in the April 2023 issue of Labour Force, Australia. Both sets of weights back to the 2014-15 financial year and September 2014 quarter have been aligned to include the revisions to ERP that incorporate the results of the 2021 Census.
Microdata for June quarter 2023 and the complete 2022-23 financial year are not yet available in the B&I microdata file in ABS DataLab and are scheduled for release on 27 November 2023.