Disputes about shared data are almost never about the data itself. They are about who assumed they were the custodian — and both sides usually assumed it was them, for reasons that felt obvious at the time.
Decide before the first transfer
- Who holds the master copy and where it physically resides.
- Who may access it, and through what process.
- What each party may publish independently.
- What happens to the data when the project ends.
- Whether either party may share it onward, and on whose approval.
Legal and regulatory checks
- Whether personal data may cross the border at all, and under what conditions.
- Whether anonymisation is sufficient under both jurisdictions.
- Whether the consent given covers international transfer and future reuse.
- Whether any category of data requires specific national approval.
Practical arrangements that work
- Mirror the dataset so both institutions hold a full copy.
- Version and document it so both sides interpret it identically.
- Agree a shared codebook before analysis begins.
- Use access logs so nobody has to reconstruct who did what.
The publication question
- Agree an embargo period during which neither publishes alone.
- Agree how a party proposes an independent analysis.
- Agree what happens if one side wants to publish and the other does not.
- This last case will occur and is easiest to settle before it does.
Long-term stewardship
- Decide who deposits the data in a repository, and when.
- Ensure the local partner has full access after the project ends.
- Data held only by the funded partner effectively leaves the country that produced it.
- Agree how future requests from third parties will be handled.
One thing worth remembering
Mirror the dataset so that the local institution holds a complete copy from the start, not a copy handed over at the end.
Projects end untidily — funding stops, people move, relationships cool. A partner who already holds the full data does not have to ask anyone for it later, and that single arrangement removes the most common source of long-term grievance in international research.
Câu hỏi thường gặp
What are data sharing disputes actually about?
Not the data itself but who assumed they were the custodian — and both sides usually assumed it was them for reasons that felt obvious.
What should be decided before the first transfer?
Who holds the master copy and where it resides, who may access it and how, what each party may publish independently, what happens at project end, and whether either may share it onward.
What legal checks are needed?
Whether personal data may cross the border and under what conditions, whether anonymisation suffices in both jurisdictions, whether consent covers international transfer and reuse, and whether national approval is required.
What practical arrangements work?
Mirror the dataset so both institutions hold a full copy, version and document it identically, agree a shared codebook before analysis, and use access logs.
Why should the local institution hold a complete copy from the start?
Because projects end untidily and a partner who already holds the data does not have to ask anyone for it later — removing the most common source of long-term grievance.
