|Page tools: Print Page Print All|
Recent activity in the Census Data Enhancement Project
All the linkage work conducted by Analytical Services Branch for four quality studies as part of the Census Data Enhancement project was completed in late October 2007. Since then, the team has been writing reports of what was done, preparing a data linking manual and performing a suite of analyses using the linked data sets. All this work is now coming to fruition with presentations at conferences and the first papers of results being released.
Glenys Bishop and Tenniel Guiver from Analytical Services Branch attended the Second National Symposium on Data Linkage Research held in Adelaide in October 2008. Their presentations on 'Sampling based clerical review methods in probabilistic data linking' and 'Determining the quality of longitudinally linked Census data' were well received for their rigorous and comprehensive approach. Sybille McKeown also spoke about some of the challenges in the Indigenous Mortality Quality Study. Glenys Bishop presented an overview of the Census Data Enhancement project at a Record Linkage Workshop conducted in conjunction with the International Association of Cancer Registries Annual Scientific Meeting in November. Glenys Bishop, Tenniel Guiver and Jeff Wright gave talks about methodological issues of data linking at the Australian Statistical Conference in July.
Two papers using results of the Indigenous Mortality Quality Study were released in November. A discussion paper on the assessment of methods for developing life tables for Aboriginal and Torres Strait Islander Australians used the results of the study to obtain a direct method for estimating Indigenous deaths. An information paper outlined the findings of the study, particularly with respect to assessing the undercoverage of Indigenous deaths in death registration records and identifying factors that may be contributing to undercoverage of Indigenous deaths in death registrations.
James Chipperfield presented a paper to the November Methodology Advisory Committee meeting on generalised linear models with probabilistically linked data. In this piece of work, James and other team members devised a way of implementing a method to adjust regression coefficients when fitting models to inexactly linked data. They were able to show their method worked but that missing links altogether had a much larger effect on the model than incorrect linkages. This paper will be published as a research paper in the near future.
For further information please contact Glenys Bishop on (02) 6252 5140 or firstname.lastname@example.org
These documents will be presented in a new window.