Who Should Control the Data in Sponsor-Funded Research?

If a sponsor pays for the study, who gets to control the resulting data?

A company funds a university study. The sponsor's contract says that the company owns the database. The university investigators collect much of the information but receive only selected extracts for analysis. Elsewhere, another sponsored study stores the master dataset at an academic coordinating center while giving the sponsor agreed access.

Which arrangement is correct?

There is no universal rule assigning legal ownership of all sponsor-funded research data to either the sponsor or the individual researcher. Ownership, custody, access, stewardship, sharing, and analytical authority are distinct questions governed by contracts, institutional policies, consent terms, funder requirements, privacy obligations, and applicable law. For scientific integrity, however, researchers responsible for published findings need enough access to the underlying evidence to evaluate and stand behind what they report.

Focus on investigator access and scientific authority, not ownership language alone

“Who owns the data?” is often too simple a question

Ownership, custody, access, and control are different

Research teams often use “ownership” loosely to describe several different powers. That can create confusion when negotiating sponsor-funded research.

A sponsor might legally own a database while investigators possess contractual rights to analyze it. A university might retain custody while data are shared through a controlled-access repository. The principal investigator might have broad analytical access without personally owning the data at all.

Researchers often do not personally own institutional research data

A common misconception is that the principal investigator automatically owns all data generated in their laboratory. Institutional policies frequently assign rights or stewardship responsibilities to the university or another research organization rather than to an individual investigator.

The exact arrangement varies by institution, funding mechanism, jurisdiction, and contract. Researchers should therefore consult the policies and agreements governing their project rather than treating possession of a laptop, laboratory notebook, or database password as proof of legal ownership.

Sponsor ownership is not necessarily the central scientific problem

A contract stating that the sponsor owns a database may sound alarming, but the scientific implications depend on the rights attached to that ownership.

Can investigators inspect the complete data relevant to their analyses? Can they query inconsistencies? Can they independently analyze or verify the analysis? Can they retain enough documentation to respond to peer review or post-publication questions? Can they report conclusions that differ from the sponsor's interpretation?

If the answer to those questions is yes, sponsor ownership may coexist with meaningful investigator independence. If the sponsor uses ownership to prevent investigators from seeing or analyzing the evidence, the arrangement becomes much more problematic.

Current ICMJE guidance explicitly addresses author access to data

ICMJE strengthened its recommendations in January 2026 by adding specific guidance on authors' access to data. Current guidance calls for reporting whether authors had access to study data and describing the nature and extent of that access.

ICMJE also advises authors to avoid sponsor agreements that interfere with access to all study data or with their ability to analyze and interpret the data and independently prepare and publish manuscripts.

This is important because authorship carries accountability. A researcher who cannot meaningfully examine the evidence may be poorly positioned to accept responsibility for claims derived from it.

Full access does not necessarily mean unrestricted possession of identifiable data

Calls for investigator access should not be interpreted as requiring every author to receive unrestricted copies of every identifiable participant record.

Human-participant research can involve privacy, consent, confidentiality, data-protection, and security obligations. Access may appropriately be role-based, de-identified, pseudonymized, or provided through secure environments. Particularly sensitive data may require controlled access.

NIH's current guidance on participant privacy explicitly emphasizes that appropriate data sharing must coexist with protection of participant privacy and applicable legal, ethical, and policy requirements.

The relevant standard is therefore sufficient access for legitimate scientific responsibility, not indiscriminate distribution of sensitive information.

The investigators should be able to verify analyses carrying their names

Suppose a sponsor maintains the master database and its statisticians conduct the primary analysis. That arrangement does not automatically make the analysis invalid.

The problem becomes more serious if academic investigators receive only a summary table and cannot examine the underlying data, analytic code, transformations, exclusions, or other information necessary to understand how the result was produced.

This connects directly to sponsor influence over methods and analysis. Researchers need enough access and documentation to meaningfully scrutinize the evidence rather than merely transmit a sponsor-generated result into a manuscript.

Data governance should address what happens when sponsor and investigators disagree

Disagreement is inevitable in some collaborations. Investigators may believe a particular analysis is necessary while the sponsor disagrees. The sponsor may question a data interpretation. Researchers may identify an unfavorable finding the sponsor considers unimportant.

A robust agreement should establish who can conduct additional analyses, what data investigators can access, how disagreements are documented, and whether investigators can report scientifically defensible interpretations.

Without those provisions, whoever controls the database may acquire practical control over the scientific conclusion.

Data sharing introduces obligations beyond the sponsor-investigator relationship

Some funders require prospective planning for data management and sharing. NIH's Data Management and Sharing Policy applies to covered NIH-funded or conducted research generating scientific data and requires submission of a DMS Plan describing how scientific data and accompanying metadata will be managed and shared.

NIH defines scientific data for this purpose as data commonly accepted in the scientific community as sufficient to validate and replicate research findings, regardless of whether those data ultimately support a scholarly publication.

NIH also recognizes that proprietary data and private-sector co-funding can create legitimate limitations. Its current guidance advises applicants to discuss these issues early and avoid agreements that prohibit or unnecessarily restrict data sharing.

These are NIH-specific requirements, not universal rules for every sponsor-funded project. They illustrate why data-sharing obligations should be considered before contractual restrictions are accepted.

Sharing data does not mean ignoring privacy, consent, or proprietary restrictions

Open data can support reproducibility and reuse, but “make everything public” is not a universal solution to sponsor control.

Human-participant consent may limit secondary use. Privacy laws may restrict disclosure. Indigenous or Tribal governance may create additional responsibilities. Commercially licensed materials may impose legitimate restrictions. Security considerations can also affect access.

NIH's DMS framework expressly allows justified limitations on sharing and recognizes legal, ethical, and technical constraints.

The goal is therefore appropriate access and responsible stewardship, not openness without qualification.

Publication rights and data rights should be negotiated together

A contract can promise researchers freedom to publish while simultaneously denying them the data needed to verify what they publish. Conversely, researchers may have extensive data access while the sponsor retains a publication veto.

Neither arrangement provides complete scientific independence.

Data access, analytical authority, and the right to publish sponsor-funded findings should therefore be evaluated together.

Plan for data access after the project ends

Research questions do not disappear on the publication date. Journals may ask questions during peer review. Readers may raise concerns years later. Corrections, secondary analyses, audits, regulatory inquiries, or replication efforts may require access to documentation and data.

The agreement should therefore address retention, archival responsibilities, investigator access after project completion, repository deposition where appropriate, and what happens if the sponsor or research organization changes ownership or ceases operating.

A dataset that can be inspected only while the sponsor finds the collaboration convenient offers weak long-term accountability.

Why legal ownership and scientific control can produce different answers

Now change one condition: investigators receive only sponsor-generated summary tables and are prohibited from accessing the underlying analysis dataset. Legal ownership has not changed, but scientific accountability has. The authors are now being asked to stand behind evidence they cannot meaningfully verify.

Common mistakes about control of sponsor-funded data

Negotiate the powers attached to the data, not merely who “owns” them

When reviewing a sponsored-research agreement, replace the single question “Who owns the data?” with a set of more practical questions about scientific authority and stewardship.

The party storing or legally owning a dataset does not necessarily need exclusive scientific control over it. Good governance separates legitimate stewardship and proprietary interests from the investigators' ability to verify and report the research responsibly.

What should a sponsor-funded data agreement clarify?

Frequently asked questions about control of sponsor-funded data

Data ownership matters less scientifically than what ownership allows someone to control

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.