6602.0 - Microdata: Longitudinal Labour Force, Australia, 2008-10 Quality Declaration 
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 10/12/2012  First Issue
   Page tools: Print Print Page Print all pages in this productPrint All  
Contents >> File Structure >> Longitudinal structure


FILE STRUCTURE

<< Previous Section: Survey Methodology


LONGITUDINAL STRUCTURE

Households selected for the Labour Force Survey (LFS) are interviewed each month for eight months, with one-eighth of the sample being replaced each month. This rotating structure allows for the creation of longitudinal linkages into 8 month panels, where the responses from households across an 8 month period can be linked together. Panels that reach the end of their 8 month period are replaced with new panels who generally have similar characteristics.

The diagram below conceptually represents the longitudinal structure for a shortened illustrative period of January 2008 to December 2008. For each month, a complete cross-section of data is provided with all 8 rotation groups included (i.e. columns with squares of all 8 colours).

Image: Conceptual diagram of longitudinal structure where monthly cross-sections of data are aligned into longitudinal panels.

Starting in January, there is a 'red' panel that is answering the Labour Force Survey for the last time after starting in June 2007. The previous 7 monthly responses are not included, as they lie outside of the selected period.

In February, this group is dropped from the LFS and replaced by a new 'red' group, which answers the LFS for the period February to September 2008. In this case, the full 8 month longitudinal panel is included and linked via common and unique identifiers for each person across the months.

In October, this group is again replaced by a new group whose panel spans October 2008 to May 2009. In this illustrative example where the CURF was only for January 2008 to December 2008, the data from January to May 2009 would not be included as they fall outside of the selected period, however in the actual CURF they are included.

It's important to note that not all respondents will be available across the full 8 months. People who move to a new dwelling, are away from their usual dwelling for more than 6 weeks, or become out of scope (such as joining the permanent Australian defence force) do not provide responses for the whole selected 8 month period (see Number of months captured in survey [ABSNMTH]). Furthermore, records obtained from special dwellings (see Residence Status [URSTATC]) are not linked longitudinally, as they are obtained from dwellings such as hotels, hospitals and prisons, which may or may not be occupied by different people from month-to-month.

When analysing data longitudinally, people with incomplete records and those from special dwellings are not included, and these people are more likely to be not in the labour force or unemployed. People who provide responses for the full 8 month period are more likely to be stably employed, so longitudinal analysis is likely to have a significant bias towards employment, so take care when interpreting results. For more information, refer to Population count [POPCNTC].

It is expected that future releases of the LLFS CURF will extend both forwards into 2011 and backwards into 2007. This will mean that the partially complete panels in this release that extend beyond December 2010 and before January 2008 will be included in future releases of the LLFS CURF.

Note: While the above diagram conceptually illustrates the longitudinal structure, the data in the LLFS CURF is actually presented as a 'long' vertical file, as below, rather than a 'wide' horizontal file, as above. To convert the LLFS CURF into the 'wide' horizontal format, refer to Longitudinal analysis.

Image: Diagram of actual CURF structure as a 'long' file of stacked cross-sections



Previous PageNext Page