Integrated Data Enabling Statistics and Analysis (IDEAS) for UK Statistics

Abstract

IntroductionThe Office for National Statistics (ONS) is transforming the way it produces statistics and how it facilitates policy and academic research. Legislative change will enable access to administrative and commercial data to underpin this transformation but only through an innovative record linkage will the full benefits of transformation be realised.
 Objectives and ApproachGreater access to non-survey data brings with it huge opportunities for both social and economic statistics in the UK - significant improvements in timeliness and geographic fidelity. There are also risks in working with large, complex datasets which have not been compiled for statistical analysis. ONS are developing the Integrated Data Enabling Analysis and Statistics (IDEAS) initiative to provide ONS staff to linked datasets, without constructing a large database of everything. The approach involves routinely linking data to three indices (person, address, business) and assigning unique reference numbers. Data are then de-identified but can be re-joined for defined projects.
 ResultsThe session will set out case studies of linkage work that either have already been taken forward using IDEAS or those which are planned. Particular emphasis will be given to how IDEAS can facilitate research across a range of statistical research topics by linking on person, address and business.
 By making anonymised linked data available in a safe setting the IDEAS framework also has the potential to facilitate research across government and across academic research. ONS have ambitious plans to use this approach to draw together expertise from across both sectors to be able to realise the benefits of linked where knowledge is shared on data quality, methods and previous research undertaken.
 Conclusion/ImplicationsThe IDEAS framework is as step change in the production of official statistics. With appropriate legal, ethical and security controls, linked data within IDEAS can be repeated and refined for different applications. The framework enables collaboration across disciplines and has the potential to be using across government and academia.