Background with rationale An overarching concern influencing models of data linkage for public good is the maintenance of personal privacy. This concern is at times so strong that it prevents or slows the progress of achieving worthwhile linked administrative datasets across allied government departments, and even between distinct units within a single department. Where linkage has succeeded it has generally produced data sets that, by design, are difficult or impossible to re-identify, therefore meeting the requirement to guard privacy at the costs of the resulting data’s value to government decision makers.
Main Aim The main aim of this paper is to convey criteria to inform data linkage policy and practice in government that maintains a central role for privacy, but which can better deliver on the promise of high value data for policy.
Methods/Approach This paper is informed by the Tassie Kids project, a longitudinal linked administrative data study using an embedded researcher model underway in Tasmania, Australia. Among other outcomes, the project was designed to assist allied government agencies to identify key policy leverage points across multiple services. Using the Tassie Kids project as a case study this paper asks why allied departments don’t routinely link administrative data. Several important linked administrative data design principles are drawn from discussion of this question.
Results The paper explains the practice implications of these design principles relevant to policy analysis and information management units in government.
Conclusion The paper concludes with the suggestion that high value linked administrative data is data that maximises its representation of the dynamic mechanisms that affect the outcomes desired by government, while simultaneously minimising the data’s distance from its point of origin.
Managing AUTOTYP Data: Design Principles and Implementation
