1504.0 - Methodological News, Jun 2013  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 28/06/2013   
   Page tools: Print Print Page Print all pages in this productPrint All

Family Tree Method

Data from the Australian Census of Population and Housing are typically supplied to users in the form of flat files wherein each record pertains to a single individual. Dwelling, family and person identifiers are provided so that users may reorganise the data hierarchically, and this arrangement appears to satisfy the requirements of most users. However, many important research questions require more detailed information on family interrelationships. Examples include the analysis of fertility rates, social mobility across generations, patterns of intermarriage and ethnic identification, and social inclusion/exclusion of households.

On every Australian Census household form, a reference person is identified and other household respondents are asked to define their relationship to this person. Based upon this reported direct relationship and responses to other Census questions, it is often possible to deduce information on the relationships between other family household occupants – in particular spousal, parent-child and grandparent-grandchild relationships. An advantage of the Australian Census is that the coding of relationship in household is based on the nuclear family concept, i.e. each separately identified couple relationship and/or parent-child relationship forms the basis of a family.

Due to confidentiality reasons, Australian data users are only granted access to confidentialised person-level sample Census files, e.g. 1% sample file via CD-ROM and 5% only through Remote Access Data Laboratory (RADL). For analyses requiring family interrelationships, non-ABS users have to make a special requests to the ABS for producing customised data, which requires significant resources for complex programming.

Drawing on international experience from the Integrated Public Use Microdata Series (IPUMS) project, Analytical Services Unit have been developing an extension of the ‘own-children’ method to identify family interrelationships. With this method, the Census de-identified data can be re-organised in a manner such that each record is appended with demographic and socioeconomic information from parents, grandparents, and/or spouse, where applicable. This method, known as the ‘Family Tree’ method, will allow users to utilise family interrelationships in their analysis. The re-organised data can be easily handled using a standard statistical package. The name of this method should not be confused with genealogy, which is an entirely different thing.

Analytical Services have recently used the method to look at intermarriage (inter-partnering), mixed parentage and inter-generational transmission of Indigenous status to help explain intercensal changes in the count of Aboriginal and Torres Strait Islander people. They are also looking into how it can be used to improve match rates in data linking exercises. Children have limited information to differentiate each other, while parents’ information can be used as additional linking variables to distinguish between children with very similar characteristics.


Further Information
For more information, please contact Guangyu Zhang (02 6252 5791, guangyu.zhang@abs.gov.au)