Contents >>
Raw and adjusted linkage rates
RAW AND ADJUSTED LINKAGE RATES
The proportion of total death registrations that were linked to a Census record can be referred to as a raw linkage rate. Whilst it gives an initial estimate of overall linkage results, it does not take into account the number of death registrations for which there was no equivalent Census record.
Although it is impossible to decide on a case by case basis whether or not a Census record exists for a particular death registration, one way to model the overall likelihood is to use the ratio between the actual Census count and Estimated Resident Population (ERP). Under the assumption that people represented in the death registration file were subject to the same likelihood of being counted or missed in the Census as were the general population, this ratio can be applied to the total number of death registrations to provide an estimate of the number who had an equivalent Census record. The result can be referred to as 'expected links'. The ratio between actual and expected links then provides an overall adjusted linkage rate.
To calculate the adjustment factor, the ratio of Census count to ERP by Indigenous status was applied to the death registration data by sex, age group and state/territory, as these variables are known to be critical in Census undercoverage. Due to the small number of Aboriginal and Torres Strait Islander deaths in Tasmania and the ACT, these jurisdictions were combined in the calculation and no separate adjusted linkage rates are available.
After applying the adjustment factor, the number of expected links was estimated at 150,266 (3,189 less than the total number of death registrations available for linking). Of the expected links, there were 2,099 that were for Aboriginal and Torres Strait Islander persons. For more information on the calculation of adjusted linkage rates for the 2011 Death registrations to Census linkage project, see Appendix A: Calculating Adjusted Linkage Rates.
Print Page
Print All