Data management planning

Research data management is the organization, storage, preservation and sharing of data collected and used in a research project. It involves the everyday management of research data during the lifetime of a research project, such as using consistent file-naming conventions. It also involves decisions about how data will be preserved and shared after the project is completed, such as depositing the data in a repository for long-term archiving and access.

Where do libraries fit in? Libraries are experts in describing, managing and providing access to research objects. Traditionally, these have been in the form of articles, journals, books, maps and other outputs, but we’re expert in describing and managing data as well, so we can consult with you on research data issues.

Increasingly, funders now are requiring that researchers submit, along with the proposal itself, a plan for managing the data that will be generated as part of the research process, referred to as a data management plan (DMP). While you’re an expert in the subject of your research, you may be less familiar with developing a research data management plan. In addition to funders, some academic journals — Nature and Science among them — are requiring that researchers share their data as a condition of publishing, so it’s likely that data management planning is somewhere in your future. The creation of a data management plan is good practice even if it is not required.

Why DMPs?
Why are funders now requiring data management planning? There are myriad reasons, including:

• Transparency: How did you get the results you got?

• Reproducibility: Can other research replicate your results?

• Impact and visibility: The data, as a research object, can be as valuable an output as the article or book written based on the data.

• Innovation: Other researchers may find uses for the data that you never considered.

What is required?
The first step in developing a DMP is to understand what the funder requires. Sources for funder requirements include:

• Data Management Policies and Guidelines (University of Pittsburgh Health Sciences Library System) — http://hsls.libguides.com/datamanagement/policy.

• DMP Requirements (DMPTool) — https://dmptool.org/guidance.

• Managing Your Data: Funding Agency Guidelines (University of Minnesota Libraries) — https://www.lib.umn.edu/datamanagement/funding.

If your research is funded by a contract rather than a grant, consult with the contracting agency and refer to the terms of your contract to determine whether a data management plan is required.

What does a DMP include?
Most DMPs typically address:

• The types of data to be collected.

• Data format, including data and metadata standards.

• Plans for sharing data, including policies affecting access and reuse of the data by other researchers.

• Plans for archiving and preserving the data.

Where to start?
How do you begin crafting a DMP? Talk to an expert at the library. You also can use an online resource provided by Pitt libraries called DMPTool. This can guide you through the creation of a data management plan that will meet your funder’s requirements. DMPTool is widely used by researchers at institutions around the U.S. and provides templates and step-by-step guidance on creating an effective and compliant DMP. To access go to https://dmptool.org and select the University of Pittsburgh from the drop-down list of institutions.

An important component of data management planning is planning for the sharing of the data. For many disciplines, there are repositories familiar to researchers in the field (for example, in social science disciplines, ICPSR is a notable data archive — https://www.icpsr.umich.edu/icpsrweb/index.jsp).

The University’s own institutional repository D-Scholarship@Pitt (http://d-scholarship.pitt.edu/) offers longterm storage for scholarly output, including datasets.

D-Scholarship@Pitt:
• Accepts nearly any format of file including tar.gz and zip files.

• Assigns your data deposits a digital object identifier (DOI), a permanent and unique identifier for a digital object that is used in citations and will help others to find and cite your data.

• Allows you to add information that provides important context for your data so that others can discover, understand and trust the data files.
• Is best suited for datasets that are inactive (i.e., after the completion of a research project).

• Tracks your work using alternative metrics to help demonstrate your impact and see how others are using your data.

• Can be used to add a catalog-only entry for datasets that you’ve deposited in another data repository.

• Allows you to make data fully public, private or available only to the Pitt community.

If you need help with your data management planning, contact us at www.library.pitt.edu/ask-email.

Jeff Wisniewski is the web services and communications librarian for the University Library System.