ABSTRACT

There are three longitudinal census studies in the UK: the Office for National Statistics Longitudinal Study, the Scottish Longitudinal Study and the Northern Ireland Longitudinal Study. These studies, which have been running for different lengths of time, vary in their sample size and are constructed on the basis of individuals’ birth dates. All three studies include linked administrative data (or permit such data to be linked for specific project extracts), with considerable variation in the range of such data. The data are detailed and thus potentially disclosive. Access to the data is therefore restricted in a number of ways: all researchers must individually have approved status, and each project must also be approved. Once the researcher has acquired these approvals, there are close restrictions on the use of the data: access is either within a secure facility to which researchers must travel or via remote submission of a processing script to a support officer, who can return results only if they satisfy certain criteria including minimum cell counts. Additional release criteria then apply to the publication of any results. The three studies are separate, and cannot be analysed together as a UK-level dataset. However, a number of resources have been developed to aid cross-study analysis, including an integrated data dictionary, and a process by which models can be securely and iteratively run across more than one study.