internet resources

Allison Langham-Putrow and Amy Riegelman Discovery and scholarly communication aspects of Sources for online information

reprints are not a 21st-century inven- periods on some platforms. As explained by Ption, but there has been a considerable Cameron Neylon et al., a in phys- uptick in the creation of preprint repositories ics means something different than a work- and the attention to the role that preprints ing paper means in economics.1 Whereas play in . Note, the definitions the physics community cites preprints and of preprint have evolved over time, lack considers preprints as establishing priority, consensus, and vary in terminology used a working paper operates under a different across disciplines. For the purposes of this standing as a work in progress. It is important article, we will be referring to a preprint as a to acknowledge that language has evolved version of a that is self-archived oddly as scholarly communication has tran- and shared publicly before publication in a sitioned awkwardly from print to electronic. scholarly journal. The status of a preprint The purpose of preprints is to increase can exist at several points on the scholarly the speed at which research results are dis- continuum including, but not limited to, an seminated. They are not a way to bypass author’s early draft, a submitted manuscript peer review—they bypass delays resulting under review, or an accepted manuscript. from the peer review process.2 They have Preprints are classified as grey literature clear benefits to the authors, as preprints al- and green . Prior to electronic low authors to stake a claim in their research communications, early versions of manu- by putting a “time-stamp” on their ideas.3 scripts were circulated by authors via mail- The National Institutes of Health (NIH) cites ing services and informal sharing among both of these reasons for sharing preprints colleagues. After the creation of Los Alamos in a March 2017 Notice allowing NIH-funded National Laboratory’s (LANL’s) arXiv in 1991, researchers to cite preprints as products of preprints were commonly referred to as e- NIH funding and cite them in further grant prints. Confusing matters, the term e-prints applications. It identifies additional benefits encompasses preprints as well as post-prints of sharing preprints: the ability to obtain made publically available on a server. feedback and offset publication bias.4 For the purposes of this paper, the authors are not treating preprints as synonymous with Allison Langham-Putrow is scholarly communications working papers. This has been disputed, but librarian, email: [email protected], and Amy important distinctions exist as working papers Riegelman is a social sciences librarian, email: aspringe@ umn.edu, at the University of Minnesota are not necessarily green open access due to privileged access, paywalls, and/or embargo © 2019 Allison Langham-Putrow and Amy Riegelman

C&RL News October 2019 506 Some criticisms of preprints include the by Cornell University, the Simons Founda- fear of being scooped, plagiarism, and indi- tion, and contributions from 233 member viduals making medical decisions and other institutions. Access: https://arxiv.org. life choices based on unrefereed (potentially • bioRxiv. Launched in 2013, this repos- unverified) research findings. A counterpoint itory contains research from the life sciences to the latter is that flawed research has and and is maintained by Cold Spring Harbor Lab- will likely continue to be published by highly oratory (CHSL). reputable peer review journals.5 Being penal- bioRxiv facilitates ized by journal publishers for sharing data direct transfer of or results before submitting a manuscript is submissions to a common concern regarding preprints. The participating journals. This means authors “Ingelfinger rule”6 is often cited as a reason are able to submit manuscripts to participat- to not share a preprint; however, publishers ing journals by transmitting files as well as are changing their stances to encourage the metadata, from the bioRxiv platform. They re- sharing of preprints.7 fer to this as “B2J.” Access: www.biorxiv.org. Preprint sharing, particularly in health- • ChemRxiv. The scope of this preprint related research, is not without its critics, who repository is chemistry and related fields. It worry of negative consequences of waiting is a collaborative initiative including the fol- to review research until after publication. lowing groups: the American Chemical So- However, as Amy Harmon reported, “. . . ciety, the German Chemical Society, and the some #ASAPbio [a non-profit organization Royal Society of Chemistry. Access: https:// that promotes sharing of preprints as quickly chemrxiv.org. and openly as possible] advocates argue that • MedRxiv. Launched in June 2019, since the rise of the Internet, biologists have this health sciences preprint server is a col- been abdicating their duty to the public— laborative effort between CSHL, BMJ, and which pays for most academic research—by Yale. Submissions are screened to deter po- not sharing results as quickly and openly as tential risk to possible.”8 public health. In this article, readers will be informed Access: me- about where to locate preprints, the gen- drxiv.org. eral infrastructure of repositories, preprint platforms for journal clubs, and where to OSF preprints locate journal and publisher policies regard- Several partner repositories are hosted on ing preprints. the Open Science Framework (OSF) pre- print platform. The partner repositories vary Servers, repositories, and platforms by discipline, region, governing structure, This list does not include institutional reposito- and screening process. The repositories and ries, which often host self-archived preprints. their disciplines are listed below. • arXiv. Initially the server was hosted at LANL and was limited to physics research AfricArXiv, literature from African scientists but now includes several other disciplines, AgriXiv, agriculture and allied sciences including mathematics, computer science, Arabixiv, Arabic open science repository quantitative biology, quantitative finance, BodoArXiv, medieval studies statistics, electrical engineering and systems EarthArXiv, earth sciences science, and economics. arXiv was created EcoEvoRxiv, ecology, evolution, and by Paul Ginsparg in 1991, and in 2001 Gin- conservation sparg and the repository moved to Cornell ECSarXiv, electrochemistry and solid University. arXiv is considered a trailblazer state science and technology for electronic preprints. It is currently funded EdArXiv, education research

October 2019 507 C&RL News engriXiv, engineering existing preprint servers. Access: http://z.umn.edu FocUS Archive, focused ultrasound /PreprintRepositories. research Frenxiv, literature from French scholars Discovery INA-Rxiv, literature from Indonesian scholars The following are a few resources research- indiarxiv, literature from Indian scholars, ers can use to find preprints. maintained by Open Access India • Europe PMC. Europe PMC began LawArXiv, legal scholarship indexing preprints in 2018.10 Preprints are LIS Scholarship Archive, library and index using Crossref searches on specific information science DOI prefixes. It currently indexes preprints MarXiv, ocean and marine-climate sci- from bioRxiv, PeerJ Preprints, ChemRxiv, ences and F1000Research. Preprints are labeled MediArXiv, media, film, and communica- as such and crosslinked to the peer-re- tion studies viewed versions once they are published. MetaArXiv, research transparency and Access: https://europepmc.org/. reproducibility Google Scholar. Google Scholar results MindRxiv, the mind and contemplative may include links to preprints, although there practices are some idiosyncrasies with how preprints NutriXiv, nutritional sciences are linked. OSF provides several reasons why PaleoXiv, paleontology a preprint may not appear in Google Scholar PsyArXiv, psychological sciences results, including issues with author names.11 SocArXiv, social sciences And although Google Scholar’s policy 12 is to SportRxiv, sport, exercise, performance, set the primary link to the publisher version, and health there are cases in which the record of an ThesisCommons, theses article links only to the preprint, even after the article has been published in a journal. Access all OSF partner repositories: This has been a known issue caused by a lag https://osf.io/preprints/. in time between publication in a journal and Google Scholar updating its record.13 Check • PeerJ Preprints. Launched in 2013, the preprint server record to determine if a this multidisciplinary preprint service is final version has been published. Access: complementary to PeerJ, which is an open https://scholar.google.com/. access megajournal. On September 3, 2019, • search.bioPreprint. The University of PeerJ announced its plan to stop accepting Pittsburgh Health Science Library Service preprints as of September 30, 2019. The ar- released this preprint search tool in 2016, chive will remain accessible.9 Access: https:// introducing it through a preprint posted peerj.com/preprints/. on bioRxiv.14 search.bioPreprint conducts • Preprints.org. Launched in 2016, this a federated search of the Quantitative Bi- platform is multidisciplinary and maintained ology section of arXiv, bioRxiv, F1000Re- by Multidisci- search, PeerJ Preprints, preprints.org, and plinary Digital Wellcome Open Research. It allows results Publishing Insti- to be filtered by topic or source, but it tute. Access: https://preprints.org. also uses IBM Watson Explorer to offer a • Research preprints: Server list. The list of “remix” feature that reclusters search re- preprint repositories in this column is not meant to sults to identify secondary topics. search. be comprehensive but instead to give readers an bioPreprint also offers a bookmarklet that idea of some of the options. Alternatively Martyn allows the user to highlight a word or Rittman from Preprints.org maintains a spread- phrase on any webpage and search on it. sheet with timely updates of characteristics of Access: https://www.hsls.pitt.edu/preprint.

C&RL News October 2019 508 • SHARE and OSF Preprints. SHARE is oratively write reviews of preprints and issue a collaboration between the Center for Open a DOI to their review, allowing the review Science and the Association of Research Li- to be cited and indexed by search engines. braries that was established to make research, PREreview is working with bioRxiv to make particularly that hosted in institutional reposi- reviews discoverable from the preprint server. tories, widely Access: https://www.prereview.org/. discoverable • PubPeer. PubPeer presents itself as by improving “The online journal club.” Users can provide the standardization of metadata. SHARE reviews on any type of manuscript (includ- is hosted by COS and is what powers the ing preprints) that has a DOI or arXiv ID. It search functionality of OSF Preprints. Access: allows for anonymous commenting, which https://share.osf.io/discover, https://osf.io they suggest has led to the exposure of “un- /preprints/. expected levels of research misconduct.”17 Access: https://pubpeer.com/. Infrastructure • Protocol Journal and publisher perspectives for Metadata Harvesting (OAI-PMH) and Preprint policies vary by publisher and jour- the Santa Fe Convention. OAI-PMH de- nal. veloped as a result of a 1999 meeting to ad- • Committee on Publication Ethics dress issues with interoperability of preprint (COPE). COPE provided a number of rec- servers and other digital repositories, which ommendations in a Discussion Document resulted in an interface/protocol called the in March 2018:18 journal editors will need to Santa Fe Convention.15 OAI was founded the determine whether to consider for review following year in recognition of the fact that work that has been posted to a preprint the same issues of interoperability affected platform and decide which, if any, versions libraries, museums, and publishers. Version of the article the authors may post during 2.0, the current version, was released in 2002. peer review or after article acceptance. These Crossref, which has allowed registration of decisions have implications on issues, such preprints since 2016, makes its metadata as anonymizing peer review. COPE recom- available through OAI-PMH. Access: https:// mends that publishers have clear policies on www.openarchives.org/pmh. copyright and license requirements and notes • ASAPbio: Surveying the landscape that preprint platforms should provide a clear of products and services for sharing pre- description of the license applied to pre- prints. ASAPbio maintains a list of preprint prints. Access: https://publicationethics.org server products and services available to /resources/discussion-documents/preprints. server operators. The list identifies the prod- • International Committee of Medical uct or service, the owner or developer, and Journal Editors (ICMJE). ICMJE addresses a list of known users. Open source products preprints in their recommendations documen- identified include OSF, ePrints, and a preprint tation. The recommendations state that au- platform that the is thors should not submit the same manuscript developing for SciELO. FigShare and Atypon16 to two different journals, but that journals can have partnered with a number of publish- consider publishing articles that were posted ers and scholarly societies. Access: https:// as pre- asapbio.org/preprint-products. prints. The Platforms for preprint journal clubs pre- • PREreview (Post, Read, and Engage print server must clearly identify the work with preprint reviews). PREreview is as a preprint that has not been peer reviewed hosted by Authorea. Journal clubs can collab- and include conflict of interest statements.

October 2019 509 C&RL News ICMJE places the responsibility for amending 6. E. Marshall, “Franz Ingelfinger’s Legacy the preprint to identify the final, published Shaped Biology Publishing,” Science, vol. version of the article on the authors. Access: 282, no. 5390 (1998): 861. http://www.icmje.org/icmje-recommenda- 7. “Springer Nature Journals Unify Their tions.pdf. Policy to Encourage Preprint Sharing,” Nature • List of academic journals by pre- 569, no. 7756 (2019): 307. print policy. A Wikipedia page lists preprint 8. Amy Harmon, “Handful of Biologists policies for various journals and publishers. Went Rogue and Published Directly to Inter- The information is provided in two tables. net,” New York Times. For publishers with unified policies, the 9. Jason Hoyt, “PeerJ Preprints to stop ac- table lists the policy type, policy text, policy cepting new preprints Sep 30th 2019,” Peer- on preprint licenses (if stated), and a refer- JBlog, September 3, 2019, https://peerj.com ence. Journal entries list the publisher, policy /blog/post/115284881747/peerj-preprints-to type, policy text, policy on preprint licenses, -stop-accepting-new-preprints-sep-30-2019/. and provides a link to a reference. Access: 10. Maria Levchenko, “Preprints in Eu- https://en.wikipedia.org/wiki/List_of_aca- rope PMC: Reducing friction for discoverabil- demic_journals_by_preprint_policy. ity, Europe PMC [blog], July 11, 2018, http:// • SHERPA/RoMEO. This initiative uses blog.europepmc.org/2018/07/preprints.html. color-coding to identify the self-archiving 11. Sara Bowman, “Why isn’t my pre- print showing in Google Scholar? Preprint FAQs,” OSF.io 2019, https://help.osf.io/hc /en-us/articles/360019930493-Preprint-FAQs. policy of journals. In SHERPA/RoMEO, the 12. “Publisher Support.” Google term preprint refers to the version of an article Scholar, https://scholar.google.com/intl/en that has not been peer reviewed and “post- /scholar/publishers.html#policies. print” refers to the version of the article after 13. “The Google Scholar preprint peer review revisions have been completed. bug redux,” Serial Mentor, 2015, https:// Both “green” and “yellow” coded journals al- serialmentor.com/blog/2015/10/8/Google low for preprint sharing. Access: http://www. -Scholar-bug-r. sherpa.ac.uk/romeo/index.php. 14. Carrie L. Iwema, John LaDue, Angela Zack, and Ansuman Chatto- Notes padhyay, “search. bioPreprint: A dis- 1. Cameron Neylon, Damian Pattinson, covery tool for cutting edge, preprint Geoffrey Bilder, and Jennifer Lin, “On the biomedical research articles,” bioRxiv Origin of Nonequivalent States: How We (2016): 052563, https://www.biorxiv.org Can Talk About Preprints,” F1000Research /content/10.1101/052563v1.full. 6 (May 2, 2017): 608. 15. Marshall Breeding. “Understand- 2. Peter Suber, Open Access (Cambridge, ing the protocol for metadata harvesting of MA: MIT Press, 2012). the open archives initiative,” Computers in 3. Ibid., 63. Libraries 22, no. 8 (2002): 24-29. 4. Office of Policy for Extramural Re- 16. Wiley purchased Atypon in 2016. search Administration, “Reporting Preprints 17. PubPeer Foundation, “About Pub- and Other Interim Research Products,” Peer,” PubPeer.com, https://pubpeer.com NOT-OD-17-050, 2017, Grants.nih.gov/grants /static/about (accessed August 25, 2019). /guide/notice-files/NOT-OD-17-050.html (ac- 18. Committee on Publication Eth- cessed August 25, 2019). ics, “Preprints,” Publicationethics.org, 5. “Key MEDLINE Indicators,” https:// https://publicationethics.org/resources www.nlm.nih.gov/bsd/bsd_key.html (ac- /discussion-documents/preprints (accessed cessed August 25, 2019). August 25, 2019).

C&RL News October 2019 510