ONS
The Office for National Statistics (ONS) plays a crucial role in sourcing, linking, and curating public sector data for ADR UK, ensuring that data is accessed by researchers safely and securely. With the support of ADR UK funding, the ONS continues to expand, improve and increase the range of administrative data available in the ONS Secure Research Service – the organisation’s facility for providing secure access to de-identified unpublished data for research, as an accredited processor under the Digital Economy Act 2017. The ONS Secure Research Service is the trusted research environment for ADR England.
This year has been busy and varied for the ONS as they continue to run the Secure Research Service. This provides secure access to de-identified record level data for research in the public good. The ONS continues to improve services for users: reducing project approval times; expanding the remote access scheme; and implementing two factor authentication for user access. Work has continued to ensure that the ONS Secure Research Service Metadata Catalogue, launched in February 2022, is now fully embedded.
Increasing the range of data for public good research
The ONS has made an increasing number of datasets available from other government departments, including Ministry of Justice, Department for Education and Department for Energy Security and Net Zero. New partnerships have been developed with the Bank of England and some charities to support the hosting of their data in the Secure Research Service for wider research use. The Education and Child Health Insights from Linked Data (ECHILD) - England dataset has been made available for researchers to apply to use, in addition to an updated and expanded version of the Longitudinal Education Outcomes dataset that will be made available imminently.
Synthetic data has been a key focus for ADR UK this year, because of its potential to allow users to develop and test their code before applying to use the real data. The ONS has made significant progress in supporting this work. They have created a new synthetic version of the Annual Survey of Hours and Earnings 2020 dataset and ran a successful pilot project to test its value and usefulness.
Tracking the outputs of public good research
The ONS has added to its capability to monitor data use in the Secure Research Service via the introduction of digital object identifiers (DOIs). These are unique digital identifiers that make it easier to track online where digital assets, in this case, datasets, have been cited in the public domain. DOIs are now included in the ONS Secure Research Service Metadata Catalogue.
ONS has successfully brought more new datasets across a range of themes into the Secure Research Service over the last year. We continue to support researchers using the service to enable new analytical insights from the datasets on offer. We are now progressing a new round of exciting ADR UK-funded research-ready data projects that will further increase the range of datasets available to support research in the public good.
Impact case study: Education, social mobility, and earnings outcomes for students receiving free school meals in England
Research carried out by Office for National Statistics analysts estimated social mobility across generations by exploring whether an individual in a low-income household at age 15 was still economically disadvantaged through low earnings by 25 years old. This was measured by using free school meal status and whether an individual’s annual earnings are greater than the Living Wage in the tax year when they turned age 25. Using the Department for Education’s Longitudinal Education Outcomes dataset, the research demonstrates the potential of using linked administrative data to understand labour market outcomes and address complex policy questions, such as social mobility.
The research found that:
-
- at age 25, 23% of free school meal recipients had recorded earnings above the Living Wage, compared to 44% of those who did not receive free school meals
- - 42% of free school meal recipients had earnings below the Living Wage
- - at all levels of qualification, those eligible for free school meals were earning less at age 30 than their peers who had the same highest level of qualification
- - 29% of free school meal recipients had no recorded earnings at age 25 compared to 15% for non-recipients, representing a 14% percentage point difference. No earnings may include being in education or in receipt of benefits.
The Longitudinal Education Outcomes dataset enabled analysis at a level of detail that would not have been possible using previously available data. The dataset provides greater population coverage, greater accuracy, and longer reporting periods than other approaches to understanding labour market outcomes.
Outputs from this research have been widely disseminated by the Department for Education, with organisations such as Teach First using the findings to inform their policy recommendations. The findings were also referenced by the Institute for Fiscal Studies in a report on measuring and understanding English educational inequalities.
Digital object identifiers: Implementation in the ONS Secure Research Service
Monitoring and evaluation are core activities for data services, and keeping track of use of research outputs using ONS data can be a challenging activity. However, there are tools than can help us identify outcomes deriving from the research we support.
Digital object identifiers (DOIs) are just one of these tools. They are unique identifiers, in the form of a permanent web address, that make it easier to track online where digital assets such as datasets have been cited in research, policy and the public domain. For data owners, they help give credit for publishing data and statistics and promote research transparency, while for researchers they support accurate citation and referencing.
In October 2022, the ONS released DOIs for the datasets listed in the ONS Secure Research Service Metadata Catalogue. The team brought together stakeholders from across the ONS to evaluate and discuss business and user needs for DOIs. This is the first time a government agency has used DOIs, and the Secure Research Service pilot has led to a recommendation by the Data Standards Authority for the wider use of DOIs in government.
A tracking tool called Altmetric Badges was added to datasets in March 2023. These badges aim to pick up mentions of the dataset DOI in published research papers, policy reports, the media and social media.
Growing Up in England
The Growing Up in England (GUIE) dataset enables researchers to explore how factors such as geography and household composition shape children’s outcomes. GUIE was featured in the October 2022 ADR UK Research Fellowships funding opportunity, alongside other ADR England flagship datasets. The ONS team collaborated with the ADR UK Strategic Hub on the funding opportunity, including with feasibility checks for applications.
This coincided with the release of an ONS-produced publication on the GUIE dataset, providing researchers with an insight into potential uses for this unique linked data. The ONS also published a user guide detailing the aims of the dataset and some high-level information on its make-up. A data quality manual was also put together, which is available within the Secure Research Service and is focused on lower-level data observations.
The ONS is continuing to enhance the GUIE dataset by setting up a researcher community. This will serve as a hub where researchers can exchange questions and observations regarding the data, and it will be moderated by the GUIE team at the ONS.
Next steps for the GUIE dataset include the addition of English School Census and Vulnerability data, which will bridge gaps in the existing data and increase the coverage.
Sharing code in the ONS Secure Research Service
In May 2023, the ONS launched the Secure Research Service code sharing repository, which allows users to contribute code that could be useful to other researchers. Code can vary from full reproducible pipelines to snippets of code, and might derive from preparatory work for analysis, cleaning data or creating new useful variables.
From mid-2022, ADR UK funding for a dedicated support role within the ONS to initiate and help promote code sharing activity. This included researching and setting up the repository, creating submission and governance workflows, working with early adopter pilot projects, and increasing capability through showing good practice in writing code. This activity included producing guidance on writing good, shareable code, drafting code templates for various statistical software, and running workshops, for example, with ADR UK-funded researchers. The materials are available on the ADR UK Learning Hub.
The code sharing process has now moved to business as usual for the ONS Secure Research Service team. When a researcher wishes to share code, it is checked to make sure it is well-documented and so that personal information is not disclosed. Files are then added to the code sharing drive, where all Secure Research Service analysts can see and use them. Instructions on how to submit code are available in the ONS Secure Research Service.
