Preservation decisions

Data producer support home page

Selecting data for long-term access and future reuse

Preservation decisions determine which outputs will be kept for long-term access and reuse and which files can be securely disposed of at the end of a project. This stage is about moving from a large collection of working materials to curated data that has clear long-term value.

These decisions should be guided by funder requirements, institutional policies, ethical approvals and the principle that shared data should allow others to understand, verify and build on the research.

For social science research, preservation usually involves retaining curated data collections rather than only small publication-specific extracts.

 

Appraisal is the process of identifying which data and supporting materials should be preserved.

Under the UKRI Research Data Policy, grant holders are expected to share data that have long-term reuse value. In social science research, this normally means preserving the curated data generated by a project, not just selected variables, tables or outputs used in a single publication. Note the policy is expected to be published by the end of 2026.

In practice, preserved collections commonly include:

  • Project-generated data about people, organisations, economic activity, places or social processes.
  • Derived versions of data that are needed to interpret or reuse the collection.
  • Documentation describing how the data were created and processed.
  • Code and workflows required to reproduce results.

Many projects also use data obtained from external sources, such as administrative records, official statistics or platform and digital trace data. Where reuse and redistribution are permitted under licence or agreement, these data should be included in preserved collections. When original data cannot be shared, data producers should still preserve and share well-documented code, processing steps and metadata to support transparency and reproducibility.

Social science projects increasingly combine data from multiple sources, for example by linking surveys with administrative records or merging datasets.

When working with linked data, preservation decisions should consider:

  • Legal and contractual conditions attached to each data source.
  • Whether the full linked dataset can be shared.
  • Whether de-linked, derived or aggregated versions can be preserved.
  • How linkage methods and quality measures will be documented.

If full sharing is not possible, these limitations should be clearly documented and reflected in deposit metadata and access conditions.

Final disposal helps reduce security risk, supports compliance with data protection obligations and avoids unnecessary storage of sensitive or restricted materials. Disposal should follow institutional guidance and secure erasure standards.

Before deleting local copies of project data, data producers should confirm that the deposit has been successfully completed and that preserved versions are safely stored and accessible within the repository.

This includes checking that:

  • all intended files were deposited
  • metadata and documentation are complete
  • access conditions have been correctly applied
  • any embargo or controlled access arrangements are in place
  • repository validation or review processes have been completed.

Only after the preserved version is confirmed should local working copies be considered for secure disposal, in line with institutional retention and security policies. Where appropriate, disposal actions should be recorded so that audit trails and project records remain complete.