Tuesday, January 05, 2010

Policy Forum on Public Access to Federally Funded Research: Management

Following is my response to the Policy Forum on Public Access to Federally Funded Research: Management (U.S.) (Phase 3).

Thanks once again for an opportunity to participate in this important consultation.

Question: Compliance. What features does a public access policy need to ensure compliance? Should this vary across agencies?

Response:
There should be a simple basic policy covering all agencies: mandate requirement of open access to published results of funded research, with an embargo period of no more than 6 months. Full libre open access – no embargo, re-use rights for results – should be encouraged. Beyond the basics, implementation details might best be left to the individual agency. For the NIH, for example, the PubMedCentral database initiative is well established, in the U.S., the U.K., and Canada, with discussions to extend this initiative further on an international basis. It makes sense, then, for the NIH policy to specify deposit in PubMedCentral. In addition to providing access and an easy means for monitoring compliance, this stipulation allows the NIH to carry forward the tradition of the U.S. National Library of Medicine as the world’s premiere library steward for medical literature, ensuring the preservation of the medical literature.

This centralized, government-led database approach may or may not work best for other agencies. Academic disciplines vary a great deal, and interdisciplinarity is increasingly common. The approach that I would recommend is working with existing strengths of academics and their organizations (mostly universities) and disciplines. There are only a very few highly successful, centralized subject databases similar to PubMedCentral. Where these exist, my suggestion is to work cooperatively with the existing initiative. For example, in physics, the arXiv, http://arxiv.org/was created by and for physicists who use it heavily; arXiv has established cooperative relationships with many publishers, allowing researchers to begin the submission process for publication through arXiv, and publishers also host arXiv mirror sites; for physics materials, arXiv is the obvious locus of deposit. For physics articles, then, I would recommend collaborating with and supporting arXiv, whose main site is Cornell University Library. For economics, collaboration with the distributed RePEC http://repec.org/ is recommended. In education, the U.S. Educational Resources Information Coalition (ERIC) http://www.eric.ed.gov/ is the first place to start. For library and information studies, my suggestion is to collaborate with and support the global volunteer-based collaboration, E-LIS, the Open Archive for Library and Information Studies http://eprints.rclis.org/ [disclosure: I am volunteer member of the E-LIS Governance Team]. This is not a comprehensive list of disciplinary approaches; apologies for any omission.

The vast majority of academic literature, however, does not fall into any of the disciplinary repository approaches. For most U.S. government-funded literature, therefore, what I would recommend is deposit in an institutional repository. Universities and their libraries are more than ready to cooperate in this manner, which has advantages for the U.S. university system; more on this later.

Question: Evaluation. How should an agency determine whether a public access policy is successful? What measures could agencies use to gauge whether there is increased return on federal investment gained by expanded access?

Response: Evaluation is an area where I would suggest agency-specific approaches. For example, I would argue that the NIH leadership in open access, going back to providing free access to the PubMed index, has been key to the development of the practice of evidence-based medicine. The basic idea of evidence-based medicine is that after doctors and other health care practitioners graduate, they should continue to keep up their knowledge through reading, and make use of the research literature in diagnosing and treating illness. It would make sense to hypothesize that the NIH Public Access Policy will further advance evidence-based medicine; the tools and research methods of evidence-based medicine, then will be useful in evaluation. Other positives from the NIH Public Access Policy include speeding up advances in medical knowledge, and technology transfer (for example, facilitating the discovery of new drugs). These potentials would suggest different methodological approaches.

In education, I would be very surprised if there was not a serious gap in access to the research literature for educators and parents. School libraries anywhere are lucky to have sufficient resources for the kids themselves; most would see professional development resources for teachers as a luxury. In this case, simple before and after access measurements for access to research would be suitable. Surveys of educators to determine awareness of potential access would be beneficial; there may be a need for an educational campaign, as those who are not used to having access will not automatically think to look for it once it becomes available.

For technology transfer to the private sector, economic studies similar to those conducted by Houghton et al. http://www.knowledge-exchange.info/Default.aspx?ID=316 would be most suitable.

Question: Roles. How might a public private partnership promote robust management of a public access policy? Are there examples already in use that may serve as models? What is the best role for the Federal government?

Response: The key partnership that I would recommend is with the researchers’ home organizations, generally universities. University libraries have well-established services and infrastructure that can greatly assist management. For example, the vast majority of university libraries have functioning institutional repositories. Virtually all university libraries participate in consortial groupings that could quickly find means to collaborate to ensure that all researchers have a repository available. University libraries have a robust technological infrastructure, reflecting their current leadership roles in providing access to (and increasingly, creation of) electronic information, and knowledgeable professional staff with established liaison relationships across campuses, across the nation. Librarians also have long-established business relationships with scholarly publishers. Even where the locus of deposit is a central repository (as with the NIH policy), librarians have expertise in assisting researchers with questions surrounding author rights retention to ensure compliance, and with deposit. The collaborative nature of the library profession means that it would be easy for any agency to work with the university library institutional repositories to ensure adherence to metadata standards to facilitate virtual repositories that will meet the specific needs of the agency, e.g. to showcase the work of the agency. [Disclosure: I am a professional librarian].

Working cooperatively with universities and their libraries has advantages for the U.S. university system as a whole. A full university IR will showcase the work of the university. Conceptually, there are strong reasons to believe that the IR will be a key evaluative component for universities in the very near future. For example, showcasing the work of a creative research department through the IR will only help to attract the best and brightest graduate students, and usage statistics for the IR will increase the university’s web presence, an emerging evaluative criteria for universities.

When work is deposited in a central repository such as PubMedCentral, it is ideal if it is cross-deposited in an institutional repository as well. This will facilitate preservation and assured ongoing access; the more copies of an item, the more likely it is to be preserved, and preservation is a key function for the academic library.

Once again, thank you for the opportunity to participate, and I hope that these comments are helpful to you.