Open Book: Institutional intent vs. repository reality

Institutional repositories have not evolved in the way they were first intended, writes Alvin Hutchinson

The creation of institutional repositories (IRs) which became common among research organisations during the first decade of the 21st century was in part a response to excessive cost increases of scientific journal subscriptions. A common aspiration was that as a key part of the open access (OA) movement, institutions would use repositories to take back ownership of their published output and provide access to those unable to afford these exorbitant subscriptions. And because journal subscriptions are mostly handled by libraries, the campus library was often a central sponsor of the IR.

It is clear that IRs have not evolved as first intended. Since the widespread proliferation of IRs at American universities, the participation by researchers in contributing scholarly content has been notably weak. A common experience was that after planning and policy creation; installation of a software platform; and the creation of user accounts, few scholars took the time to contribute to their IR. The reluctance by scholars to deposit their works had several reported causes. Among those most cited: unfamiliarity with rights and permissions; difficulty navigating and learning the repository software; and a much stronger affinity with one’s discipline than institution. Whatever the explanation for low deposit rates, the assumption of many early planners did not come to fruition and many IRs were left largely barren of the material they expected to be a core part of their content.

Where even marginal success was achieved in recruiting content, IRs frequently relied on mediated deposit whereby library or other staff took on the task of metadata creation, rights and version assessment and uploading files supplied by scholar-authors. Some institutions have had the staff resources to offer this service but where it wasn’t feasible, repositories grew to become primarily a home for locally produced content like electronic theses and dissertations, special collections such as gray literature, and university-published or unpublished material. This and other trends have led to the suggestion that research libraries have to reevaluate their approach and vision for IRs.

A lot has happened in the OA realm since the proliferation of IRs. OA journals certainly seem to have gone mainstream with many of the major commercial and nonprofit science journal publishers launching either dedicated OA journals or offering a hybrid option for existing titles. The early distrust and misunderstandings that scholars once had for OA journals (for example that they weren’t peer-reviewed) have largely vanished. Additionally, transformative agreements and other library negotiations with commercial publishers to foster open access have added weight to the idea that OA journals are here to stay. Although it is not exactly clear whether these developments represent a gain (a 'win' if you prefer) for the OA movement, it is clear that the weight of the movement has shifted away from locally archived papers by the parent institution (green OA) to content hosted by professional publisher or other communal platforms representing scholarship across organisations.

No doubt IRs have helped countless researchers find scholarly articles that would otherwise have been locked behind paywalls. But in managing repository content there are difficult choices to make, including the allocation of scarce (human) resources. Creating metadata, verifying permissions and versions and uploading content on behalf of scholars themselves requires staff with a variety of skill levels. Given the current state of budgets in higher education and in particular research libraries, any and all cost saving measures should be considered. And one of the easiest in my opinion is to stop collecting articles published in OA journals where it is otherwise available from a trusted source.

It makes sense for repository managers to link to content published in open access journals rather than collecting, describing and storing them in a local repository. Due to the costs of capturing, evaluating and managing this digital material, the process should be evaluated very carefully for any return on the investment of staff time if mediated deposit is required. 

Undoubtedly green open access has helped put enough pressure over time on publishers to see the writing on the wall and get into the OA journal game. But the widespread participation by major publishers in the OA journal space will likely shift IR priorities away from the redundant collection of this content. While the ideal of a scholar-author entering and uploading their OA journal articles to their home IR might be desirable, where mediated deposit is required, keeping a local copy of these articles should be done sparingly. Archiving articles from popular OA journals which participate in professionally recognized archival programs may be an admirable effort to backstop the OA movement, but in most cases it is not an optimal use of resources.

Of course this decision has to be weighed against the estimated longevity of individual OA journals which is more art than science. A 2021 article published in the Journal of the Association for Information Science and Technology and reported in Science, Nature and other outlets found that dozens of OA journals had vanished from the web over the years and their contents were essentially lost. Several months later, a thread on the Association of College and Research Libraries Scholarly Communication discussion group, (“OA publications and institutional repositories” July 9, 2021) asked IR managers whether they collected copies of articles which were published in open access journals or simply linked to them. Although the discussion did not mention the earlier news of the disappearance of OA journals, it was certainly timely.

A significant portion of mainstream OA journals have established plans for the longevity of their content including participation in cooperative archiving programs such as CLOCKSS and Portico. (Information on archiving plans and practices for many OA titles is collected in the Directory of Open Access Journals.) Despite recent reports of the disappearance of OA journals, these institutional commitments to established archiving programs – although not a guarantee of permanence – should be enough to allow many repository collections managers to concentrate their limited resources on other content where discovery, access and longevity are less certain. 

One might argue that just as the OA author-pays model has an inherent free-rider problem, the same situation could exist with centralized archiving solutions for OA journals – the benefits are enjoyed by many but the costs are borne by very few. And while it is possible that these mainstream archiving programs that include OA journals could one day face financial hardship or even insolvency, the same could be said of just about any organisation, nonprofit or commercial. 

Institutions which have enacted OA mandates or which perform text-mining and other machine operations against a full corpus of organisation’s research output may have good reason to collect papers published in accepted OA journals. But for purposes of researchers having access to individual papers, leave the open access articles to the open access journals – at least from the established publishers. Many are covered by consortial and cooperative archiving efforts and collecting them locally in our IRs is increasingly a luxury most repository managers cannot afford.

Alvin Hutchinson is an Information Services Librarian at Smithsonian Institution