Anonymising data and protecting participants

Protecting the privacy of individuals represented in research data is a core responsibility in social science research. Many data contain information that could directly or indirectly identify individuals, households or small groups. Effective anonymisation enables responsible data sharing while reducing the risk of harm or reidentification.

When does anonymisation apply?

Not all social science data relate to individual people. Some data focus on organisations, institutions, places, economic indicators or fully aggregated statistics. This section applies primarily to research data that include information about living individuals or small groups where there is a realistic risk of identification. Where data do not relate to identifiable people, anonymisation may not be required. However, other forms of risk management, such as contractual restrictions, commercial confidentiality or security classification, may still apply and should be considered as part of responsible data governance.

Anonymisation should not be treated as a single technical step carried out at the end of a project. It is most effective when planned early and applied consistently throughout the data lifecycle, alongside informed consent where appropriate and proportionate access controls.

This section provides guidance on anonymisation concepts, practical techniques, and how privacy protection fits into both active research workflows and data-sharing preparation.

A three-pronged approach to privacy protection

Effective privacy protection in social science research involving people typically relies on a combined approach:

• Transparency through clear communication with participants (e.g. participant information sheets, survey statements, consent forms etc.) about how their data will be used and shared.
• Technical protection through pseudonymisation and anonymisation.
• Governance through access control and usage conditions.

Each component plays a different role.

Participant communication supports ethical transparency by informing participants about how their data may be used and shared. Anonymisation reduces the risk that individuals can be identified in released data. Access controls provide governance mechanisms that restrict who can use data and under what conditions.

These elements work together. Clear communication with participants alone does not remove the need for anonymisation or access controls. Likewise, anonymisation should not be treated as a substitute for clear participant communication or responsible access decisions.

In projects where direct participant communication is not possible, such as those using administrative records, legacy datasets or large-scale digital trace data, enhanced anonymisation and stricter access controls become especially important.