Highlights of: National Network of Libraries of Medicine Symposium: Doing It Your Way: Approaches to Research Data Management for Libraries
Rockefeller University, NY, April 28-29, 2014
In late April, the National Network of Libraries of Medicine, Middle Atlantic Region (NN/LM MAR) hosted a two-day symposium on research data management (RDM). The event garnered well over 100 participants from the mid-Atlantic and beyond, professionals both from medical libraries and a variety of other settings who are providing or exploring RDM services.
The initial keynote speaker was Paul Harris, Director, Office of Research Informatics, Vanderbilt University School of Medicine. He encouraged participants to seek out opportunities to develop tools and services immediately useful to their local researchers which also would further the goals of RDM. He profiled several tools that they provide locally, including:
- Project RedCap (Research Electronic Data Capture): This system enables the collection of metadata about active biomedical projects and associated collected data at one’s institution. It was created at Vanderbilt and since has been deployed at other institutions via the RedCap Consortium.
- StarBRITE CMS Researcher Portal: This Vanderbilt-specific platform enables the centralized collection of information about research: news, pilot funding, project information, researcher profiles.
The second keynote speaker was Keith Webster, the Dean of Libraries for Carnegie Mellon University. He provided a general overview of the importance of RDM for academic libraries (in the light of changes in the way that science is done and evolving roles for academic libraries). He then spent time situating the attendees’ work within important trends in the broader, international, professional context. He encouraged participants to develop their skills in this field and to stay aware of the very significant progress and initiatives happening internationally, particularly in Europe and Australia. He noted some key reports and articles to read (listed at the end of this posting).
The final keynote speaker was Jared Lyle, Director of Curation Services of the Inter-university Consortium for Political and Social Research (ICPSR). ICPSR, the large, long-standing social science data archive based at the University of Michigan, has a tradition working with data producers to acquire data and then curate it for optimal re-use by secondary researchers. ICPSR tools highlighted include: the data catalog (which enables discovery of datasets and granular information including variables) and the Bibliography of Data-related Literature (which links ICPSR studies to resulting publications based on the data in the archive). With ICPSR’s history of supporting data re-use, he pointed out that a well-prepared data collection should be complete and self-explanatory. However, researchers (many of whom may have a high willingness to share) rarely have sufficient time, money, or resources to prepare and document their data well for re-use. But he pointed out as well that many professionals in the field are trying to better understand this landscape and develop new services in order to improve the sharing and quality of research data. For example, one Stanford librarian works with local researchers to curate, redistribute, and archive their research data. The Stanford Social Science Data Collection is a type of intermediary repository; staff members work with researchers to capture their datasets, later moving them to a more long-term repository.
University Service Models
In the afternoon, attendees heard from practicing professionals on overviews of their RDM services. Following are highlights of the services of a selection of universities:
University of Minnesota:
The Libraries have a dedicated staff member to RDM, the Research Data Management/Curation Lead, who provides services and coordinates the work of other staff. Their RDM service is overseen by a campus advisory group with members from various stakeholder departments; the Libraries are working with this group to develop a campus-wide referral network. One significant effort of the Libraries is a pilot to have staff actively curate and upload the data associated with 30 researcher projects into their institutional repository (IR). They also have worked with researchers to self-deposit their datasets into the IR, instructing them on practices in realms such as metadata. They use DSpace for their IR and are finding that the newer version (i.e., 4.x) provides more flexible features for research data, including metadata elements beyond Dublin Core.
University of North Carolina:
This university has an RDM service group co-led by two librarians (each of whose primary focus is on other service areas). Staff members provide a range of services in cooperation with other stakeholder departments on campus to whom they reached out over time. For example, they collaboratively conducted a series of information sessions on data management for researchers. The Libraries partnered with campus stakeholders to each teach components on different topics (DMPs, repository options, sensitive research data, data security), including areas of expertise outside of the Libraries. These popular sessions, in addition to being provided in-person, were live streamed (to a large audience) and recorded for later viewing. Looking towards the future, the libraries are in the process of actively reorganizing for improved research lifecycle support.
NYU Health Sciences Libraries:
They have worked at developing RDM services, working in partnerships with staff both inside the library (e.g., subject liaisons) and throughout the university. A core challenge for this institution has been to helping to change perceptions about the scope of a library and demonstrate to researchers the library’s role in RDM services. To that end, they collaborated with various staff members to develop and distribute several quickly-popular YouTube videos on the significance of RDM. These videos are used on their own and as part of library instruction (not only at NYU but, as the symposium illustrated, by many other universities as well):
- Hanson, K, Read, K, & Surkis, A; How to Avoid a Data Management Nightmare
- Hanson, K, Surkis, A, & Yacobucci, K. “Data sharing and management snafu in 3 short acts”
Their dedicated Scientific Data Curation Specialist coordinates services and the work of other staff. She manages a collaborative consulting team, consisting of two groups: 1) a core of staff members (mostly in the Libraries) and 2) additional second-level team members from departments campus-wide whom are call upon as needed (e.g., staff members from IT and legal). In addition, their service is overseen from an upper-level management council (with membership across several university departments).
On the second day, Sherry Lake, Senior Data Consultant, and Andrea Horne Denton, Research and Data Services Manager, of the University of Virginia educated attendees on some key RDM best practices via a case study that they use in their workshops, based on a case from the Digital Curation Profiles Directory. Participants examined the profiled research group’s practices in the realms of: data collection and organization, documentation and metadata, storage and backup, and preservation/sharing/licensing. In doing so, they learned about common issues which researchers might face and how to assist them.
Regarding RDM services, UVA has two different approaches:
- operational: helping to improve researcher efficiency and good organization and documentation practices throughout the life cycle
- sharing: helping researchers to be aware of requirements and plan for downstream data sharing
UVA provides many services similar to other institutions, and like some others does a series of workshops (dubbed “Research Data Management Boot Camp”) with contributing instructors from departments across the university.
Lastly, the presenter shared two lists of resources that she maintains for keeping up-to-date on the field of RDM:
Principles for RDM Work
Over the two days, presentations highlighted various strategies that professionals utilize in providing RDM services:
- Promote curation rather than sharing; the former is more salient for researchers, and must precede the latter.
- A well-prepared data collection should be complete and self-explanatory; help researchers to meet this standard.
- Encourage best practices yet support people where they are. I.e., even if a researcher’s method of sharing data— e.g., storing on one’s hard drive and responding to requests—has significant drawbacks, help them to execute their selected method in an optimal way (i.e., in this example, help them to establish appropriate backups) while at the same time gently share concerns about their method and be available to help them consider other methods when the time is right.
- Continue outreach efforts on a regular basis; people don’t always see ads even if you do a great one-time campaign.
- Once researchers have shared their data, tell them about the ways to track use of their data.
- Develop services based on the hypothesis that researchers will do the right thing (maintain their information securely, track metadata, maintain audit trails, etc.) if provided an easy way to do it with needed tools and services.
- When developing partnerships or services, the technology is the easiest part. Relationships take time to build; be prepared to slow down to work with diverse needs
- Frame one’s services within the data curation lifecycle for staff and stakeholders with whom one communicates or partners.
- In planning collaborative services with senior administrators/department heads, make sure they are communicating plans and expectations down to the PIs.
- Track your work for assessment.
- Stay aware of RDM requirements/regulations around the world, both for professional awareness and given the fact that U.S. researchers likely are collaborating across borders.
In summary, while symposium attendees were largely focused on medical library settings, the lessons learned apply to research and libraries in all disciplinary contexts.
Suggested Reports/Articles to Read