1504.0 - Methodological News, Sep 2014  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 25/09/2014   
   Page tools: Print Print Page Print all pages in this productPrint All

A More Efficient Sample Design Process for REACS 2013/14

As part of Methodology and Data Management Division's ‘Flagship Program’ for 14/15, one theme is focussed on 'Harnessing Productivity'. Within this theme there is a key project called 'Survey Productivity Strategies' which has a key element focussing on 'more efficient sample design and collection operations for identified business and household surveys'.

As part of the work on this flagship project, Business Survey Methodology (BSM) have recently reviewed the design process for the Rural Environment and Agricultural Commodities Survey (REACS). The original intent of this review was to make the process undertaken for the design more efficient, with the hope that a more efficient process would potentially lead to a decrease in sample as more time could be devoted to analysing the design rather than simply running the process.

The main focus of the review was reducing the large number of design constraints being used in the design. With a large number of commodities being designed at national, state and Natural Resource Management (NRM) level, the number of design constraints quickly reaches over 1,000. This complicates the design in terms of analysing design results to improve the design as well as increases computing time to run through different options.

The review concentrated on whether designing at national level would naturally achieve reasonable results at the state and NRM level with the logic that in order to achieve the national Relative Standard Error (RSE), sample would need to be devoted to the particular states and regions which contributed the most to the national estimate, thereby also achieving reasonable RSEs for these important states and regions.

The results suggested that the number of constraints could be cut to just over 100, with these being mainly at the national level, with some broad level information (e.g. total area of holding) for all states and regions. A small number of commodities were added into the design at the request of the Rural Environment and Agriculture Statistics Branch (REASB) where there were some concerns regarding state or regional RSEs.

This more efficient design process enabled the REACS 13/14 design to be conducted in a much shorter timeframe. This also allowed BSM to improve the design to a point where sample size savings were achieved with a reduction in sample size from 36,531 for REACS 12/13 to 34,694 for REACS 13/14. Following completion of processing of the 13/14 survey, an evaluation will be undertaken to determine whether productivity savings have been achieved and the scope for further savings in the future.


Further Information
For more information, please contact Brett Frazer (07 3222 6028, brett.frazer@abs.gov.au)

The ABS Privacy Policy outlines how the ABS will handle any personal information that you provide to us.