SAMPLE DESIGN
The GSS was designed to provide reliable estimates at the national level and for each State and Territory. The sample was therefore spread across the states and territories in order to produce estimates that have a relative standard error (RSE) of no greater than 10% for characteristics that are relatively common in the national population, say that at least 10% of the population would possess.
For the 2010 cycle, the standard sampling methodology was adapted in order to provide better estimates of people experiencing disadvantage. Census 2006 data were used to identify areas with higher concentrations of people 3 or more of the following characteristics:
1. Lives with no car at dwelling
2. Indigenous
3. Renting from public authority or community group
4. Part of one-parent family with dependent offspring only
5. Lives in private dwelling with two or more families
6. Household income less than $20,600
7. Unemployed
8. Does not speak English well
9. Does not have a non-school or year 12 qualification.
These areas, once updated for migration and population growth since 2006, were then given a higher probability of selection in the survey. Selections of small geographic areas (meshblocks) were used in urban areas, while Census Collection Districts were used in rural areas. Households were then randomly selected from each area selected to participate in the survey.
Given this sampling strategy, some areas received a smaller initial weight to reflect their higher probability of selection. Conversely, areas with a lower probability of selection received a larger initial weight. It is important to note that because of the change in sample design, a wider spread of weights is to be expected. This situation may result in greater overlap of weights across States and Territories.
The initial sample for the survey consisted of approximately 19,576 private dwellings. This number was reduced to approximately 17,158 dwellings due to the loss of households which had no residents in scope for the survey and where dwellings proved to be vacant, under construction or derelict. Of the eligible dwellings, 87.6% responded fully (or adequately) which yielded a total sample from the survey of 15,028 dwellings.
Print Page
Print All