Research Data Policy FAQs

Broadly speaking, the new research data policy is designed to help all research personnel at Duke University adopt best practices for managing research data. The benefits of these practices include attaining confidence that 1) research outputs and results are reproducible, 2) research data and outputs are properly secured and retained, and 3) research impact is increased by the appropriate sharing of data.

Importantly, new data management requirements are being released by various funding agencies. The new policy is designed, in part, to respond to those new requirements and position the Duke research community to meet these coming requirements.

No. Many of our peer institutions are adapting policies to reflect the evolving expectations for managing and sharing data.

Please see the FAQ “Is this type of ownership policy and statement consistent with our peer institutions?” for more information.

Yes. The new policy will govern the conduct of individuals interacting with research data and research outputs, regardless of funding. We believe that these policies will enhance the quality of all types of research currently occurring across Duke, regardless of external requirements.

No. Duke University’s Policy on Intellectual Property Rights, which can be found in Chapter 6 of the Research Policy Manual, governs issues related to intellectual property.

For further comments, questions, and to express interest in joining workgroups focused on policy implementation, email researchdata@duke.edu.

No, Duke University would not prohibit research personnel from working with data obtained through contracts, as long as those contracts are reviewed and signed by an Authorized Organizational Representative (AOR).

In case the existing research data needs to be validated, we recommend that researchers ensure that a DMP is in place that establishes the provenance of research data, verifies that the research data is sufficiently stored, and that they have retained information necessary to allow for others within their scholarly community to reproduce results.

DMPs can vary considerably in the complexity and level of detail depending on the nature of data. Expectations around the level of detail within a DMP may vary based on the best practices of your discipline.

No. At this time, Duke’s policies do not raise the retention threshold above the standards of your discipline and the accepted practices of your organizational unit. The policies were developed in recognition of the important roles research communities play in setting and encouraging best practices for investigators.  Duke encourages investigators to discuss retention standards and best practices with colleagues.  

It must also be noted that there is an ongoing national discussion about research reproducibility and integrity. Even in those disciplines whose best practices are to retain a limited scope of research data, Duke investigators should be mindful of how their community is evolving. We encourage all members of the Duke community to evaluate and assess retention options in the event the standards of their discipline change in the coming years. 

The six-year retention period aligns with the standard set by the Institutional Review Board and the Research Misconduct policy. While data retention is a critical responsibility of all who conduct research, and would ideally be retained in perpetuity, a balance must be struck between the storage and sharing of data and the costs of doing so.

The research data you retain should be enough to validate and/or reproduce your research project per the best practices of your discipline.

Research data may include synthetic data such as those used as part of training an algorithm or program, but not all of it needs to be retained. You should only retain enough to validate the outcomes of a research project. The best practices of your discipline should be used in your determination.

No. Determining the storage location may vary based on the details of the data and phase of research. Researchers should determine the appropriate storage location for their research data and outputs in consultation with institutional research support offices, (such as through the Duke University Libraries or the myRESEARCHnavigators) and according to the standards of their discipline.

Researchers should typically follow the retention periods required by agreements. If the agreements stipulate the retention period is shorter than 6 years, but allows for longer (i.e., they do not have maximum sharing periods or data destruction requirements) then the researcher should adhere to Duke’s 6-year retention period.

It depends on many factors. In cases where any/all of the following apply, the answer is likely yes:

• There are contractual or regulatory obligations over the research data
• The trainees’ research project was funded by a third party/ies,
• Human subjects’ data were involved
• The trainee or mentor plans to use that research data for future research projects
• The research data was disseminated (or is planned to be disseminated) in some form such as publication, research poster presentation or other types of public dissemination

In cases where none of these factors apply, the mentor should ask the trainee if they wish to transfer the research data when they leave the institution (regulatory and contractual obligations around the data permitting). If the trainee declines to request a Data Transfer Agreement for the research data, the mentor is then able to choose whether to retain the trainee’s research data and outputs.

In cases where ones or more of these factors does apply, mentors may archive finalized datasets (except for PII/PHI and export controlled containing data) in the Duke Research Data Repository out of the University Libraries.

Yes. Duke University’s policy on Research Data Ownership is in line with its peer institutions. For references, see the links below to policies regarding research data ownership at peer institutions:

• Stanford University
• Brown University
• Harvard University
• Columbia University
• University of Washington

If Duke personnel (faculty, investigators, staff or trainees) leaves or disengages with Duke, the University retains stewardship and ownership of research data in the form that it exists at the time of departure. This enables the University to respond to any questions or inquiries that may arise regarding data generated or created during their time at Duke, and allows collaborators continued access to research data. To ensure a smooth and compliant transition, personnel are encouraged to consult the relevant Off-boarding checklists for research data. 

If an individual will continue engaging with the research data after they leave or disengage from Duke, a Data Use or a Data Transfer Agreement may be needed, depending upon the regulatory or contractual obligations associated with the research data. If the research data in question was collected from research that was externally sponsored or is classified as sensitive (or of unknown classification), personnel are encouraged to seek assistance from the research contracting office (ORS/ORC).  

In general, Duke is not claiming ownership over data that were collected, generated, or created solely by entities outside of the University. And in cases of collaboration, shared ownership may be articulated in an agreement or contract. Note that such contracts and agreements must be reviewed and signed by an Authorized Organizational Representative (AOR).

Provided the activity was properly characterized as an outside activity from Duke, Duke would not seek ownership of data generated in the course of that activity.  However, you are required to follow Duke University’s Policy on Intellectual Property Rights, which can be found in Chapter 6 of the Research Policy Handbook, which requires disclosure of inventions whether or not on University time or with University facilities.  Also, note that use of Duke resources is not permitted for outside consulting work.  Faculty should refer to the following policy on Faculty Consulting Agreements and note the restrictions around research and intellectual property. 

A public dataset in its original form is not owned by Duke. However, if a researcher conducts any significant modifications to the data set or generates new research data or outputs from these modifications, it may represent a novel creation that would fall under Duke’s ownership.

Ownership of the datasets provided by the subscription service would likely be defined in the agreement between the subscription service and Duke University.

No. These agreements should be executed between the third-party entity and Duke University. As such, individual faculty/staff members cannot sign agreements on behalf of Duke. Only Authorized Organizational Representative (AOR) can sign agreements on behalf of Duke.

The Institutional Review Boards (IRB) and the Duke University Libraries Center for Data and Visualization Sciences can assist the researcher or Principal Investigator in ensuring that a consent form includes appropriate language regarding treatment and sharing of data while guaranteeing confidentiality. Please note that it is common practice for a consent form to mention that data may be viewed by members of the research team, as well as others (Duke officials or sponsors) if required.   

As with any research collaboration, understandings of ownership, stewardship and data use should be addressed and resolved during negotiations over the research agreements executed by all parties, including Duke. To determine the appropriate office to contact per your agreement need please see Navigate contracts and agreements page on myRESEARCHpath.