Generations and Gender Survey Australia Wave 1 & Wave 2

Panel

SAMPLING PROCEDURE FOR HILDA WAVE 1

1. Sampling frame 1.1 Type of frame: List of geographic units - 1996 Census districts (CDs). This frame excluded CDs that had zero land area (usually used for homeless people or off-shore/migratory people) or were remote or sparely populated. 1.2 Frame coverage: NA 1.3 Frame size: From 34,422 Census Districts in 1996, CDs defined as remote or sparsely populated were excluded. In addition CDs that had zero land area were also excluded. A sample of 488 CDs were then selected. 1.4 Level of units available: Dwelling

2. Sampling method 2.1 Sampling method type: Multistage 2.2 Sampling stage definition

   • PSU: Census District
   • SSU: Dwellings
   • TSU: Households 2.3 Sampling stage size
   • PSU: 448
   • SSU: 12,252 ( 22-34 dwellings chosen within each census district depending on the expected response and occupancy rates of the area)
   • TSU: 7800 2.4 Unit selection: Census District Selection.The list of 1996 Census Collection Districts (CDs) formed the area-based frame from which 488 CDs were selected.The frame of CDs was stratified by State, and within the five largest States in terms of population, by metropolitan and non-metropolitan regions. The CDs were sampled with probability proportional to their size as measured by the number of dwellings (unoccupied and occupied) recorded in the 1996 Census. To ensure the sample of CDs selected provided good coverage of the CDs in the frame, the CDs were sorted by statistical sub-division and section of State. Within each of these groups, the CDs were sorted into geographical (or serpentine) ordering based on the centroid of the CDs. Using a random start, a systematic selection of CDs was then undertaken by staff at the Melbourne Institute. 2.5 Final stage unit selection: Dwelling selection and Household selection.
     • Dwelling selection: ACNielsen used a specifically trained team of interviewers to visit each selected CD and provide a full listing of the dwellings from which dwellings were selected for the [HILDA] Wave 1 sample. The interviewers followed a predetermined route around the entire CD to list all dwellings they came across. Particular attention was paid to ensure that all dwellings had an equal probability of selection, including granny flats, flats, residential warehouses and battleaxe properties. The actual number of dwellings selected within each CD varied depending on projected variations across CDs in response rates and in occupancy rates. The response rate assumptions were based on ACNielsen's experience with survey response rates within the metropolitan and non-metropolitan areas of each State. The occupancy rates are based on the 1996 Census occupancy rates, with the additional qualification that no more than 10 additional selections could be added to cover the expected number of unoccupied dwellings. Given a targeted average of 16 responding households per CD, this meant the selected sample had to be large enough to generate 23 occupied in-scope dwellings per CD. The average number of selected dwellings was thus 25. The selection of dwellings from those listed occurred as follows: The initial dwelling was selected at random from the list of dwellings in the area. A skip of five in urban areas and two in rural areas was then applied to select the remainder of the dwellings required for the area. This ensured that the cluster of dwellings selected from each CD was sufficiently spread out across the CD while not generating large travel costs. For five of the more remote CDs selected that were extremely large in size, a block selection stage was used so that the entire CD did not have to be listed. For each of these areas, a satellite map which details the buildings in the CD was used to divide the CD into blocks with an approximately equal expected number of dwellings in each. A sample of these blocks was selected for full block listing.
     • Household selection: Where a dwelling contained three or fewer households, all such households were sampled. Where there were four or more households occupying one dwelling, a random sample of three households was obtained. Where there were four or more households occupying one dwelling, all households had to be enumerated at the time of first contact and a random sample of three households obtained. 2.6 Within Household unit selection: All household members aged 15 and over interviewed. 2.7 Stratification: Explicit

AUSTRALIAN GGS WAVE 1 The sampling procedure above outlined, describes the method used to select the initial HILDA sample. However, the GGS Wave 1 did not come from the first wave of HILDA, but rather from Wave 5. The selection of the sample in Wave 5 would have been determined in large part from the initial sample selection used when HILDA first started. However, some people who participated in Wave 5 of HILDA (Wave 1 of GGS), would have been new people who entered after Wave 1 and who wound not have been exposed to the same sample selection as the respondents who had participated since Wave 1. Over time, the Wave 1 household members are followed but the sample is also extended to include: • any children born to or adopted by members of the selected households; and • new household members resulting from changes in the composition of the original households. • a new household member that arrived in Australia for the first time after 2001 (since HILDA Wave 9). Continuing Sample Members (CSMs) include all members of wave 1 households (including children). Any children born to or adopted by CSMs are also classified as CSMs. Further, all new entrants to a household who have a child with a CSM and any recent immigrants to Australia (arriving after 2001 for the Main Sample, 2011 for the Top-Up Sample) are converted to CSM status35. CSMs remain in the sample indefinitely. All other people who share a household with a CSM in wave 2 or later are considered Temporary Sample Members (TSMs). Where the household has moved, split or moved and split, the interviewers and office staff track the CSMs. The CSMs (along with their new household) are then interviewed, where applicable, at their new address or by phone.36 TSMs that split from a household and are no longer part of a household with a CSM are not followed. However, if the TSM is converted to a CSM, then they are followed for interview as any CSM would be.

AUSTRALIAN GGS WAVE 2 The sample is represented by respondents who participated to HILDA Wave 5.