Data sharing in cross-border research is governed by personal data protection law, ethics approvals and funder conditions on both sides, and these differ by country and by institution. Check the current rules with your research office before relying on any arrangement described here.
Most collaborations settle authorship early. Very few settle data — and data questions arrive later, when the material already exists and both teams have reasons to want control of it.
Why it surfaces late
Three reasons:
At the start it feels premature. Nobody wants to open a negotiation about ownership before there is anything to own.
The two sides often assume different defaults. One side assumes data belongs to whoever collected it; the other assumes it belongs to the project. Both assumptions are common, and neither is universal.
Funder requirements arrive mid-project. A mandate to deposit data openly can land after collection has already happened under a different understanding.
What to settle before collection
Six questions. They take one meeting and a one-page note:
Who holds the data and where is it stored? Physical and legal custody, not just who has a copy.
Who may use it for what? The joint paper is the easy part. The real question is what happens with secondary analyses — can either side use the data for a further paper without the other, and if so, on what terms?
Will it be published, in what form, and when?
Who is named when it is published? Datasets are cited; being listed as a data author matters to early-career researchers in particular.
What happens if the collaboration ends early?
What restrictions come from the ethics approval and the consent forms? This one constrains all the others — you cannot agree to something the participants were not told about.
The consent question comes first
This is the constraint people discover too late.
If participants consented to a specific study and were not told their data might be deposited openly or reused for other purposes, then depositing it openly is not available to you, whatever the two institutions agree.
The practical implication is upstream: the consent form and the ethics application should be drafted with the data plan already decided. Adding reuse language at the start costs nothing; adding it afterwards usually means going back to participants, and often means the reuse simply cannot happen.
This also applies to anonymisation. Whether a dataset can be shared at all usually depends on whether identifiable elements can be removed without destroying its value — a question worth answering at design stage, not at deposit stage.
Embargo periods
A common and workable compromise: the data is deposited but access opens after a set period, typically after the main papers are published.
Three points to agree explicitly:
How long — and tie it to an event (publication of the primary paper) with a hard backstop date, not to the event alone. Without a backstop, an embargo pending a paper that never gets submitted lasts forever.
Who can request early access and who decides.
What happens at the end — full open access, or access on request with a stated review process.
Choosing where to deposit
Four criteria:
Persistence. A repository that issues a permanent identifier and has an institutional commitment to keeping the record — not a lab server or a personal cloud folder.
Access controls that match your consent. If the data needs managed access, the repository must support it.
Discipline norms. Many fields have an established repository; using it means your data is where people in the field look.
Acceptability to both funders. Check both sides' requirements — they may differ, and satisfying one does not automatically satisfy the other.
When partners disagree
The most common disagreement: one side is under pressure to publish the data openly and the other wants to keep it for further analysis.
Both positions are reasonable. A partner who collected data over three years in difficult field conditions and plans a series of papers from it has a legitimate interest; a partner whose funder mandates deposit has no discretion.
Three ways this gets resolved in practice:
Deposit a defined subset — the data underlying the published paper, not the whole collection.
Deposit under managed access with the collecting team involved in reviewing requests.
Embargo with a firm end date, long enough for the planned papers.
What does not work is leaving it unresolved and hoping the question goes away. It returns at the point of publication, when a journal asks for a data availability statement and the two sides have to answer in one sentence.
The data availability statement
This short statement in the paper is where all the above becomes visible. It should say what is available, where, under what conditions, and who to contact.
Two things to avoid: saying data is available on request when nobody has agreed to handle those requests, and saying data is openly deposited when the deposit has not actually happened. Both are checkable, and both damage the collaboration's credibility when checked.
One document
None of this needs a formal contract for most projects. A one-page note — agreed by email, dated, held by both sides — covering the six questions above is enough, and it is enough precisely because it exists before anyone has a stake in a particular answer.
Write it in the first month. It is the cheapest insurance in a collaboration, and the only time it can be written without either side defending a position.
Why do data questions surface late in collaborations?
Because raising them early feels premature, the two sides often assume different defaults about ownership, and funder mandates can arrive mid-project.
What constrains every other data decision?
The ethics approval and the consent participants gave — you cannot agree to a use that participants were not told about, so the consent form should be drafted with the data plan already decided.
How should an embargo be defined?
Tied to an event such as publication of the primary paper, but with a hard backstop date — without one, an embargo pending a paper that is never submitted lasts forever.
How do partners resolve a disagreement about open deposit?
Deposit a defined subset, deposit under managed access with the collecting team reviewing requests, or embargo with a firm end date.
