Scholarly Publishing

“Publishing,” on CASRAI, means everything that happens to a piece of research between the point an author is ready to share it and the point the scholarly record treats it as settled — and, where the record later needs correcting, everything after that too. That spans preprints, peer review (before and after formal publication), open access licensing and cost models, how journals and researchers get evaluated, and the standards that keep a published record’s identifiers, corrections, and provenance machine-readable.

This hub collects CASRAI’s guides, dictionary terms, and standards references on scholarly publishing in one place. It links out to more detailed content wherever that exists, and is explicit below about where coverage is still thin — this page won’t claim depth it doesn’t have.

Preprints and early dissemination

A preprint is a version of a manuscript shared publicly before (or independent of) formal peer review, typically on a discipline-specific server such as arXiv, bioRxiv, medRxiv, or a regional/national repository. arXiv itself moved from a Cornell University-hosted service to an independent nonprofit on 1 July 2026; bioRxiv and medRxiv’s governance moved from Cold Spring Harbor Laboratory to the openRxiv nonprofit in March 2025 — both changes worth knowing if you’re citing “who runs” a given server. cOAlition S‘s current (2026–2030) strategy treats preprint deposit as encouraged rather than a strict funder mandate, a shift from earlier framing.

For the practical question of which server to use for a given manuscript — licensing, DOI and versioning behaviour, ORCID auto-population, journal compatibility, and where clinical/health preprints (medRxiv) differ — see the full guide: How to Choose a Preprint Server.

Peer review, before and after publication

Peer review is not one fixed process. Journals vary in whether reviewer and author identities are disclosed to each other (see peer review (in assessment) and equitable peer review for how CASRAI’s Dictionary frames the variants), and some journals — following models like eLife’s publish-review-curate approach, which moved to publishing first and attaching reviews openly afterward in 2023 — separate the decision to publish from the evaluation of the work entirely.

Review doesn’t stop at acceptance. Concerns raised after publication go through their own editorial pathway — from an informal comment through to a correction, expression of concern, or retraction — and Crossref’s Crossmark service plus the update-to metadata field are the actual mechanism by which that status change propagates to anyone holding the original DOI. The full pathway, including where COPE (Committee on Publication Ethics) guidance and independent watchdogs like Retraction Watch and PubPeer fit in, is covered in Post-Publication Peer Review: From Concern to Correction.

Open access: models, embargoes, and cost

“Open access” describes several distinct arrangements, not one policy: gold (published open, typically via an article processing charge), green (a self-archived author accepted manuscript in a repository, often after an embargo), diamond (open with no author-facing fee), and born open access (open from the moment of publication, no embargo involved). The article processing charge (APC) that gold open access usually involves raises its own equity questions — see author fee equity and APC reimbursement (funder) for how funders and institutions handle it.

Publisher and journal self-archiving policies — which version can go where, and after how long an embargo — are searchable via Open Policy Finder (the service most people still know by its former name, Sherpa Romeo, which it rebranded from in November 2024).

How research and journals get evaluated

Two metrics dominate casual conversation about publishing prestige, and both have real limitations worth knowing before citing either. The Journal Impact Factor (JIF), devised by Eugene Garfield and published annually by Clarivate via Journal Citation Reports since 1975, measures a journal’s average citations, not any individual article’s quality — and preprint servers, by design, don’t have one at all, since they aren’t journals. The h-index, defined by J. E. Hirsch in a 2005 PNAS paper, measures an individual researcher’s output against citation count, with well-documented distortions across career stage and discipline.

Because the JIF is calculated per journal and per subject category, the only answerable version of “what is this journal’s impact factor” names a specific title and checks it against the current Journal Citation Reports release. CASRAI keeps per-journal reference entries for the titles researchers ask about most often, and reading a few of them side by side makes the category effect hard to miss: the Cancer Research impact factor and the Cancer Discovery impact factor sit in the same Oncology category and are genuinely comparable with each other; the mBio impact factor belongs to a microbiology category with a different citation baseline entirely; the Current Biology impact factor is a broad-scope life-sciences figure; and the AI and Ethics impact factor comes from a young journal in an emerging field, where a two-year citation window has had little time to settle. None of those four numbers means much set against the others — which is the same point DORA makes one level further down, about individual articles.

Two real reform efforts push back on over-reliance on both: the San Francisco Declaration on Research Assessment (DORA), developed in 2012, asks institutions and funders to stop using the Journal Impact Factor as a proxy for individual researcher quality (see DORA signatory for what signing actually commits an organisation to); and the Coalition for Advancing Research Assessment (CoARA), whose Agreement on Reforming Research Assessment was published 20 July 2022 and which formally launched as a coalition on 1 December 2022, extends that reform agenda more broadly across how research outputs and activities are recognised (see CoARA Agreement).

Systematic reviews, preregistration, and registered reports

Where a publication reports a synthesis of existing evidence rather than new primary research, reporting quality is usually judged against the PRISMA 2020 statement (Page et al., BMJ, 2021, replacing the 2009 version) — a 27-item checklist and flow diagram for transparently reporting a systematic review, including its living variant for evidence that’s updated on an ongoing basis. Separately, a registered report is a publication format where the study protocol is peer-reviewed and provisionally accepted before results exist, addressing publication bias toward positive findings directly rather than through a reporting checklist after the fact.

Where publishing connects to identifiers and CRIS systems

None of the above works as a machine-readable record without persistent identifiers underneath it: a DOI (issued via Crossref for most published articles and DataCite for many datasets and preprints), an ORCID iD linking a specific person to a specific work, and — on the contribution side rather than the identity side — the CRediT taxonomy that CASRAI originated and now stewards jointly with NISO as ANSI/NISO Z39.104-2022. See CASRAI’s CRediT overview for the contributor-role side of this, which sits in its own cluster but touches nearly everything on this page.

What’s covered here so far

Honestly: two full guides exist in this cluster today — preprint server selection and post-publication peer review — plus the CASRAI Dictionary terms linked throughout this page, which cover open access models, research-assessment reform, systematic-review methodology, and publication-integrity vocabulary in operational-definition depth even where a full narrative guide doesn’t exist yet. Open access models and cost, peer review types as a topic in their own right, journal/author metrics, and a comparison of bibliographic databases (Scopus, Web of Science, OpenAlex) are identified as real, evidenced gaps and are the priority queue for this cluster’s next guides — this page will be updated as they publish, rather than describing them as already covered.

Frequently asked questions

Is a preprint a publication?

Not in the traditional sense of a peer-reviewed journal article, but most funders and many journals now recognise it as a citable, dated research output in its own right — the distinction matters most for authorship priority and for whether a specific venue will consider a manuscript that’s already been posted. See the preprint dictionary entry for the operational definition CASRAI uses.

Does open access mean free to publish, not just free to read?

No — that’s specifically what the gold/green/diamond distinction is for. Diamond open access is free to both read and publish in; gold is free to read but usually funded by an author-facing APC; green is a self-archived copy of a manuscript that may itself have been published behind a paywall.

How does DORA differ from CoARA?

DORA (2012) is a declaration institutions and individuals sign, focused specifically on moving away from the Journal Impact Factor as a proxy for research quality. CoARA (2022) is a broader coalition with formal governance and a wider reform agenda covering how research outputs and activities of all kinds are recognised, not journal metrics alone. Signing one doesn’t require signing the other, though many organisations have done both.