The sample for the AWRS was primarily drawn from the Dun & Bradstreet register of Australian businesses. This sample frame is generated from records of enterprises seeking credit approval.
For this reason, it was found to have some deficiencies in coverage of public sector, non-government and non-profit organisations as these types of enterprises do not typically seek credit. The primary sample frame was therefore supplemented by a frame generated from information that was sourced to map the national system employer population.
In order to avoid an over-concentration of these supplemented records relative to those from the primary sample frame, the Dun & Bradstreet stratum-level sampling fractions (sample provided/population available) were applied to each stratum of these additional sample records to determine the number of records to be selected from the supplementary frame.
Over the duration of the AWRS fieldwork a total of almost 33 000 records were sourced from the primary frame and 1 640 records via the supplementary frame.
In total, ORC International attempted to invite 26 578 enterprises to participate in the AWRS. Further information about the fieldwork outcomes is available in the (Provisional) AWRS Fieldwork Report.
The employee participants were recruited to the AWRS through their employer.
Two rules were adopted by the AWRS:
- All employees from enterprises with 20 or fewer staff were invited to participate.
- A random selection of employees from enterprises with more than 20 employees was invited to participate.
The selection process was based on the ABS EEH survey protocols to encourage the generation of a random sample of employees for the AWRS.
Participating enterprises with more than 20 employees were provided with an Excel file that automatically generated the random selection of 20 employees.
The enterprise representative was instructed to enter payroll records for up to 1000 current employees into the Excel file, including the name and any relevant identifying information of employees. Once all records were entered, a macro generated the list of 20 employees that were to be invited to participate in the study by carrying across the information about employees into a separate sheet of the Excel file. The macro was designed to take account of the total number of records entered into the file.
An example of how the macro was designed to randomly select 20 records where 400 records were entered: select the 18th record and then select every 18th record until 20 records have been selected.
The employee selection process adopted for the AWRS whereby a maximum of 20 employees from any enterprise could be selected meant that fewer employees from larger employers were invited to participate in the AWRS relative to the number of employees employed by large employers in the population.
This approach was primarily taken to minimise the burden on smaller enterprises so that they were not required to undertake the random selection process. As a result a disproportionate number of employees working for smaller enterprises were invited to participate in the AWRS.