The Longitudinal Impossible Dataset: Helping Users Navigate the ONS Longitudinal Study

Abstract

The Office for National Statistics Longitudinal Study (ONS-LS) follows a 1% sample of the population of England and Wales through each decennial Census, linking Census data with data from birth, death, and cancer registers. As one of the largest datasets of its kind in the UK, the ONS-LS is used for public good research on topics ranging from public health to labour market outcomes. However, access to the data is highly controlled and only possible via secure settings, meaning that researchers must often identify required variables and develop code prior to seeing the data. With thousands of variables available, navigating and exploring the available metadata can be a complex task for users, which also often results in additional administrative burden on user support teams.

This presentation will showcase the Longitudinal Impossible Dataset (LIDS): a customisable artificial dataset intended to familiarise prospective users with the data structures and variable domains represented in the ONS-LS. Areas covered will include conceptualisation, development, deployment and user feedback, sharing insights for practice in secure data user support services for social science research. It will also offer a brief introduction of related initiatives undertaken at the Centre of Longitudinal Study Information and User Support (CeLSIUS).