A key point that led off—and was reiterated several times throughout—“Strategies and Partnerships: Tailoring Data Services for Your Institutional Needs,” the ACRL (Association of College and Research Libraries) President’s Program at the recent American Library Association (ALA) Annual Conference in Orlando, FL, was the importance of establishing a common understanding of what exactly “data services” means. The term is a catch-all for a diverse set of activities; using it without defining its scope can become problematic for everyone involved. In light of the term’s far-reaching definitions, the session brought together panelists from a range of organizations. Yasmeen Shorish, data services coordinator at James Madison University, VA; Sara Bowman, project manager at the Center for Open Science (COS), Charlottesville, VA; and Kristin Partlo, reference and instruction librarian for social science and data, Carleton College, MN, discussed research data activities, realistic methods of meeting data needs, and examples of data partnerships at work. Throughout the discussion, participants were encouraged to fill out a worksheet exploring current data services at their institution, opportunities for support, and potential partners and resources.

DATA SERVICES IN PRACTICE

Finding and using data, Shorish explained in her overview, is often related to information literacy instruction, but is also generated in scientific academic domains such as genetics or atmospheric science—“known subject areas where students are expected to go in and use raw data as part of their curricular work or individual research.” With data to be found in subject and institutional repositories and within organizations, there is no single interface for data discovery, although initiatives like SHARE (SHared Access Research Ecosystem) are working to help aggregate search practices and institutions are beginning to build data search into existing reference practices. When it comes to managing and depositing data, noted Shorish, as more activities on campus are centered around content creation it’s important for libraries to ask what services are in place or in development to support responsible management of these outputs. While it’s not the library’s job to discern whether content is what she terms “Big-S Scholarship,” it is critical that institutions evaluate their ability to incorporate it into existing roles. While some of the process is beyond the library’s territory, mechanisms available through organizations such as EZID and DataCite to mint digital object identifiers (DOIs) and archival resource keys (ARKs) for data citation can help streamline the workflow. Data as a scholarly product is an emergent area of engagement, and one that’s highly variable across institutions. “What kind of support should the library be providing in statistical analysis and data visualization?” Shorish asked—especially as these require expertise that librarians may not have developed, or software and technical skills that are more often found elsewhere on campus. A strategic and proactive approach is critical, and, she said, and the library is an appropriate home for these services, as long as the necessary institutional support is in place: “While we in this room may all be acutely aware of changes within libraries, of which data services is just one facet, our external constituents may not—and communicating changing and expanding roles is critical to the sustainability of these services.” In addition to institutional buy-in, data activities require instructional support—across the library, through workshops or seminars, or even via for-credit courses or online learning modules. Shorish noted a wide range of designated librarian roles for working with data services, from the all-purpose “data management librarian” to a recently instituted role at New York University, “librarian for research data management and reproducibility.” Shorish noted that it’s important to be explicit about an institution’s expectations. "Data services has the potential, like most of our other services that we offer, to grow with attention—that is, it can follow the 'if you build it they will come' [model],” she said. “You want to be cognizant of those trends, and either meet the growing demand or set your scope and communicate those limits to the community that you're working with."

FITTING SERVICES TO INSTITUTIONS

“Reality can strike when our aspirations don’t match reality, and we try or want to take on too much,” Partlo cautioned. “In any institution of any size, it’s necessary to be strategic and pragmatic about crafting data services to the institution itself, not some ideal.” To find the perfect fit of data services, the library should look at a combination of institutional needs, available assets, and existing trends on campus. One of the easiest ways to gauge needs is to identify gaps in services, she advised, such as helping students find data sources for capstone research projects, or helping faculty find sources for teaching quantitative literacy across the curriculum. What’s important, she stressed, is looking beyond the needs that initially present themselves. “I’m in the social sciences, and have had to push myself and my colleagues to reach out to the [hard] scientists who are not knocking on my social science–labeled door for help.” Potential partnerships can be found in the library, across campus, and beyond, she noted, from helping data literacy librarians weave more sources into their instruction to creating a community of practice with data visualization practitioners on campus to developing a copyright information pipeline. With all the available options, however, establishing a role and then standing your ground can be a challenge. “Data librarians are the most specialized of generalists,” Partlo noted. “Don’t let a massive list of needs keep you from diving wholeheartedly into that seemingly small project that you can do exceptionally well and then showcase.” Always be on the lookout for sources of energy and momentum, she added. “What are the projects, questions, ideas, strategic directions that currently animate your campus?”—perhaps a strategic plan or new curriculum. Campus partners can include a number of unlikely suspects: the art librarian who becomes a metadata evangelist, the academic technologist who team-teaches, or the grants officer grateful for additional help with a data management plan. Opportunities for self-education around data services have also exploded over the past few years, and there are a number of conferences, listservs, blogs, and professional organizations that provide networking and professional development on the subject.

COS STEPS UP

Among the organizations helping institutions streamline data services, COS is a three-year-old nonprofit working to capture information across the entire research workflow. COS approaches its mission through technology, training, and incentives via three teams: an infrastructure team of software developers, comprising two-thirds of the staff, is working to build free and open source tools to help researchers manage their workflow; a metascience team explores what makes science reproducible and develops best practices guidelines; and a community team works with stakeholders—librarians, funders, and publishers—to move toward more incentives for data sharing. Its flagship product is the open source Open Science Framework (OSF) platform for managing scholarly research, collaboration, documentation, and archiving in one place. Among other features, the OSF interface includes a Wiki for team collaboration, a recent activity widget, an analytics page, and version controls for merging workflows—including the ability to split a project into components that can be made public or kept private. Every file, user, and project component is given a persistent, unique citable identifier—OSF mints DOIs and ARKs as well—and a built-in registration feature creates a frozen copy of a project at any point in its life cycle for archival purposes. “We’re not going to try to reinvent the wheel,” noted Bowman, so OSF connects to platforms such as Google Drive, Mendeley, Dropbox, and GitHub, enabling users to see all their information at once. “Everyone is using all of these great tools but they're all over the place,” Bowman said, “so if we can just connect to them and have a central hub, one place for your researchers to go to do their work from study inception all the way through publishing and then ultimately sharing their data, this is our goal." A version of OSF for institutions, in partnership with the University of Notre Dame, is currently in the beta phase. In addition to its work with OSF, COS provides free statistical and methodological consulting, provides online resources and training videos, and partners with a number of other organizations to do training at academic libraries to help researchers, staff, students, post-docs, and faculty engaged in quantitative research learn skills to improve its reproducibility. It also works in partnership with funders and journal editors on incentives “to give researchers a little bit of a nudge toward more openness,” including issuing digital badges to acknowledge open practices. Another incentive, COS’s $1,000,000 Preregistration Challenge, promises to give 1,000 researchers $1,000 each to preregister their research and then publish the results, which is “like a clinical trial situation,” said Bowman. “You register what you’re going to do, how you’re going to do it, how you’re going to collect…[and] analyze your data, before you collect any of [it]…. We think it’s a great idea. The problem is convincing anyone outside of clinical sciences to try this.” What can academic librarians do to increase openness and reproducibility at their institutions? The session’s takeaway was clear: Connect your services, meet researchers where they are, help them find immediate benefit, and find strong partners such as COS. All links mentioned in the session, as well as additional resources, are available on the OSF site.