Open access is often misunderstood in institutional meetings, and myths about quality, APCs, CC BY and preprints quietly drive bad policy.
Myth 1: “Open access means low quality”
Quality comes from peer review and editorial standards, not the paywall. Weak journals exist in both OA and subscription models.
Access and quality are decided at different moments. Whether an article is open describes what happens after publication: who can read it, under what licence, at whose cost. Quality is decided before publication, by who edits the journal, who is asked to review, what those reviewers are asked to do, and above all what the editor is willing to reject. Removing a paywall touches none of that. A journal can do all of it well and be free to read, or do all of it badly and cost a library a great deal every year.
What the myth usually tracks is age, not access. Many open access titles are younger than the subscription journals a discipline grew up with, so they have no long citation record, no accumulated reputation among senior colleagues, and no place in an inherited departmental hierarchy. That is a real limitation and worth saying out loud. But it is a statement about one journal's track record, and it applies just as much to a new subscription title.
What to look at instead. For any journal, ask five things that are answerable from the journal's own site in about ten minutes. Who sits on the editorial board, and can you find those people at the institutions listed beside their names? What review model is used, and how many reviewers see a manuscript? Who makes the final decision? Is there a published corrections and retraction policy? And can the indexing claims on the site be confirmed at the index's own site, rather than only asserted on the journal's? A journal that answers all five plainly is being transparent about its process. One that answers none of them is a problem whether or not it is open.
What goes wrong. The usual damage is a rubric. Someone writes "peer-reviewed subscription journals" or "established international journals" into promotion criteria as shorthand for quality, and years later a candidate who published in a strong open access venue — sometimes because a funder required it — is quietly marked down by a panel that is faithfully following the text. If any of your institutional documents use the access model as a proxy for quality, that is the sentence to fix first, and fixing it costs nothing.
Myth 2: “All OA journals are predatory”
Predatory behavior is about fake review and deception — not the OA label. Use a checklist: Identify a Predatory Journal
Deception is the whole of the definition. A predatory journal takes a fee, or takes a manuscript, for a service it does not provide — most often peer review that is faked, skipped, or reduced to a formatting check dressed up as a referee report. The open access label is incidental to that. Deceptive practice exists on the subscription side too: hijacked titles that clone a legitimate journal's name and identifiers, proceedings volumes assembled from anything submitted, and outlets that exist to sell a line on a CV.
Screen the title, not the model. Run the checklist above against anything unfamiliar. The signals that do the most work in practice are these: a solicitation email that flatters your last paper and promises review within days; a fee that surfaces only after acceptance; editorial board members listed without their knowledge, which you can test by emailing one of them and asking; identifiers that do not resolve when you paste them; a scope that spans unrelated disciplines in one title; a contact address that turns out to be a mailbox service; and no visible corrections or retractions in a journal claiming a very large back catalogue.
What goes wrong. Offices try to solve this with a list. A blocklist is out of date the moment a title renames itself, and it says nothing at all about the journal an author is actually asking about, which will usually appear on neither list. An allowlist is worse, because it silently excludes legitimate regional, society, and non-English journals that nobody ever got round to assessing. Keep a list if it saves time, but the policy has to be the checklist, taught to authors, so that a researcher facing an unknown title late on a Friday can reach a defensible answer without you.
If an author has already published in one. Do not open with discipline. Establish the facts first: was there any review, what was paid, what was signed, is the article still online, and does the agreement permit withdrawal. Then decide what should be recorded and how the author should list the work from now on. Most of these cases are inexperience rather than misconduct, and the way you handle the first one decides whether anyone tells you about the second.
Myth 3: “OA always means authors pay”
Business models vary. Some OA is institutional/society funded; waivers exist for many APCs.
Name the routes, because "business models vary" is not usable advice. There are journals with no author charge at all, funded by a society, a university, a library consortium, or a ministry. There are subscription journals with an optional open route. There are agreements negotiated by your own library or consortium that cover fees for corresponding authors at member institutions — money already committed whether or not your authors use it. There are publication lines inside grants. And there are publisher waiver and discount schemes, commonly tied to the corresponding author's country or to demonstrated need.
Check them in that order. Is the journal free to publish in? If not, does the library have an agreement covering it? If not, does the grant carry a publication allowance? If not, does the publisher operate a waiver policy, and who applies to it? Two mechanics get missed constantly. The entitlement normally attaches to the corresponding author, so which co-author submits can decide whether a fee exists at all. And waivers are normally requested at submission, not on receipt of an invoice: asking before an editorial decision is an ordinary administrative request, while asking after acceptance is a far weaker position and at some journals is simply too late.
What goes wrong. An office concludes that open access is unaffordable and steers everyone toward closed venues, sometimes into breach of the funder policy attached to the very grant paying the salaries. Or the fund exists but is first-come, so it is exhausted well before the year ends, and the people who then pay personally are the ones least able to.
The mirror-image myth. Some staff assume that because a fee was paid, acceptance was bought. Keep two questions apart. Is the fee covered and proportionate? And does payment influence the editorial decision? The first is a budget question. The second is the predatory question from the previous section, and a journal where the honest answer is yes is not an expensive journal — it is a fake one.
Myth 4: “CC BY gives away the work carelessly”
CC BY requires attribution and supports reuse with credit — aligned with many funder policies. See Directive’s OA framing at Directive Publications.
Read what the licence asks of the reuser. CC BY lets anyone copy, redistribute, and adapt the work, including for commercial purposes, on the condition that the original authors are credited and changes are indicated. It does not transfer authorship. It does not let anyone imply your endorsement. And it does not make plagiarism acceptable: an uncredited reuse of a CC BY article breaches the licence and remains an integrity matter your office can act on. Under many open routes the author keeps copyright and grants a licence, which is more control than the outright copyright transfer the older model asked for and that most people signed without reading.
The real question is which licence, not whether to license. The variants differ in ways that decide practical outcomes: whether a translation can be published, whether the article can go into a course pack or a mined corpus, whether a figure can be reused by a commercial body. Some funders require the most permissive option and will not accept a non-commercial or no-derivatives variant. Authors often pick the restrictive variant precisely because it sounds safer, which is exactly how a compliance problem gets created by someone trying to be careful.
What goes wrong. A blanket internal warning against CC BY puts compliant authors in breach of their own mandate. The other failure is timing. The licence gets chosen at acceptance, in the same rushed hour as proofs, payment, and a co-author on leave, by whoever happens to be handling the email. Decide the licence when you choose the journal, check it against the funder's requirement at that point, and put the licence choice on your submission checklist with the corresponding author's name written next to it.
Myth 5: “Once posted as preprint, peer review is optional”
Preprints are not versions of record. Peer-reviewed VoR remains the scholarly standard for most evaluation contexts.
A preprint is a timestamp and an invitation, not a verdict. Posting establishes a date, makes the work readable immediately, and invites comment from people no editor would have thought to ask. Under some funder policies it also helps with open access compliance. What it does not do is tell a reader that anyone qualified has assessed the claims. Note that this myth runs the opposite way to the first four: it over-trusts openness rather than under-trusting it, and it tends to be repeated by the people most enthusiastic about reform.
The mechanics that prevent most of the trouble. When the peer-reviewed article publishes, update the preprint record with the DOI of the published version, so that anyone landing on the preprint is told a reviewed version exists. Label versions honestly wherever the work is listed: preprint, accepted manuscript, and version of record are three different objects, and a CV that blurs them reads as inflating a record even when nothing was intended. Cite the version of record where one exists, and cite a preprint only when it is genuinely the source, saying so in the reference.
What goes wrong. A communications team treats a posting as a finding and issues an announcement no reviewer has seen. An assessment panel cannot tell what has been reviewed, so it either discounts everything unpublished or credits everything equally. Two records with slightly different numbers circulate for years because nobody linked them. If your researchers work on anything clinical or policy-relevant, agree with your communications colleagues in advance what may be announced from an unreviewed posting — before the first time it matters, not during it.
What research offices should say instead
- Judge journals by transparency and review
- Support authors with policy-aware OA guidance
- Fund fairly, including waivers where needed
Judge journals by transparency and review. Make that operational rather than aspirational. Publish the short list of checks, put it into induction for new staff and postgraduate students, and use the same checks yourself when someone asks you to approve a venue. Then go through your own criteria, forms, and lists and rewrite any line that sorts journals by access model. The test of your guidance is whether an author can apply it without calling you.
Support authors with policy-aware guidance. Policy-aware means you already know, before the author asks, what each of your main funders requires on licence, on repository deposit, and on embargo. Keep one page per major funder, record who last checked it and when, and answer the question the author actually asked — can I publish here, under this grant — rather than forwarding the funder's policy document. Get into the conversation before journal choice rather than at acceptance, because before submission the choice can still change.
Fund fairly, including waivers. Decide in advance who the fund is for and say so publicly. Unfunded early-career researchers, students, and people between grants are the ones who otherwise pay personally or publish somewhere weaker. State what happens when the fund is exhausted, rather than letting people discover it. And tell your authors what the library has already paid for, because the most common waste in this area is paying a fee that an existing agreement would have covered.
The question you will actually be asked
Almost all of this arrives as one email: I have an invitation from a journal I do not know, is it legitimate, and can I use my grant to pay? A reply that closes the whole thing looks roughly like this.
Ask for two things — the journal's homepage and the invitation email itself. Check the editorial board, the description of the review process, the corrections policy, and one indexing claim verified at the index's own site. Note whether the fee is stated before acceptance or only after. Then answer in three parts: what you found, what it means, and what to do next. For example: three of the five board members could not be found at the institutions listed and the fee appears only after acceptance; this is not a title to submit to; if an open route is wanted in this field, here are two journals that pass the checks, and the library agreement covers one of them. Mark plainly which part is fact and which part is your judgement.
That reply takes about fifteen minutes and it teaches the checks by performing them. Do it a dozen times and the questions change shape — people start sending you a journal with their own assessment already attached, which is the outcome you actually want.
None of these five corrections requires a new policy. They require that the words already in your documents describe what you actually mean. Wherever a document says open access, ask whether it means the access model, the licence, the fee, or the strength of review, because those are four separate things, and every myth above comes from treating them as one.
More: Directive Academy
Frequently asked questions
What is a predatory journal?
A predatory journal is one that deceives authors, most often by faking or skipping peer review while presenting itself as a legitimate outlet. Predatory behaviour is defined by that deception, not by whether a journal is open access. Both open access and subscription titles can be weak, so judge each journal on its transparency and review process. A structured checklist is the practical way to screen an unfamiliar title.
Does open access mean lower quality research?
No. Quality comes from peer review and editorial standards, not from the paywall. Weak journals exist in both open access and subscription models, so the business model on its own tells you very little. Assess the editorial board, the review process and how openly the journal describes both.
Do authors always have to pay an APC to publish open access?
No. Open access business models vary, and some journals are funded by institutions or scholarly societies rather than by author charges. Where an article processing charge does apply, waivers exist for many APCs. Research offices should check the waiver policy before assuming a route is unaffordable for an author.
What does a CC BY licence actually allow?
CC BY permits reuse of the work provided the original authors are credited. It is a licence built on attribution, not a surrender of credit, and it aligns with many funder policies that require open licensing. Framing it as giving the work away carelessly misreads what the licence requires of reusers.
What should our research office do next about open access?
Start by judging journals on transparency and review rather than on their access model. Then support authors with policy-aware open access guidance and fund publishing fairly, including waivers where they are needed. Also remind authors that a preprint is not a version of record, because the peer-reviewed version of record remains the standard in most evaluation contexts.