Publishing Ethics Workshop: 1-Hour Grad Student Outline

You do not need a semester to prevent the most common integrity failures: this 60-minute outline gives graduate schools one focused, ready-to-run hour.

Guide · 10 min read · Directive Publications Editorial Team

Workshop goal

Students leave able to: choose safer journals, define authorship, and respond to review ethically.

Put that on the first slide and leave it there, because it is also your test for what to cut. Three verbs, three observable behaviours: a student can apply a checklist to a journal they have never seen, can say who on their own project qualifies as an author and who does not, and can turn a hostile reviewer comment into a reply an editor would accept. If a topic you want to add does not move one of those three, it belongs in a different session.

The material that crowds this hour out is predictable. Definitions of fabrication, falsification and plagiarism belong in the compliance training your institution already delivers. A tour of metrics and rankings turns into an argument and settles nothing. A walkthrough of one publisher's submission interface will be out of date before these students use it. Leave all three out, and say so in the first minute, so the room knows what it is getting and what it still has to get elsewhere.

60-minute outline

Why this matters (0–5). Make two points and move on. The record is permanent, and the decisions that damage it are made early, usually by someone junior, usually under a deadline, usually without anyone senior watching. A paper published in an outlet the field does not read cannot be moved afterwards. An author list settled at proof stage is a dispute, not a decision. Do not open with a famous misconduct case: students file those under things other people do, and the rest of the hour is about ordinary choices they will make this year.

Predatory journal live checklist demo (5–20). Do it live, on screen, working the checklist out loud, so the room watches criteria being applied rather than hearing them listed. Use an archived copy of the page rather than a live link, so the session does not depend on a site being up and you are not sending thirty people to an active solicitation. Teach the checks, not the verdict. An indexing claim gets verified in the database that supposedly does the indexing, not accepted from a logo. A board member gets looked up to see whether they appear to know they are on the board. A promised turnaround of days gets weighed against what real peer review takes. The fee gets located before submission rather than discovered after acceptance. Then show one ambiguous case — a small, legitimate, unglamorous journal with a plain website and no marketing — because a room trained only on obvious fakes has learned to pattern-match on bad design, and the expensive mistakes are rarely the ugly ones. If you have a solicitation email in your own spam folder, read it aloud. Everyone in the room already has one.

Authorship case (20–30). One case, six or seven lines, handed out on paper. A shape that works: a doctoral student runs the experiments and writes the first draft; a postdoc designs the analysis and rewrites the methods; the head of unit is added because that is the custom in the group and has not read it; a statistician who chose and ran the models appears only in the acknowledgements; a visiting student who collected a third of the data left a year ago and nobody has contacted her. Ask the room to build the byline before you give them anything. Then hand out the widely used criteria — a substantial contribution to the work, involvement in drafting or revising it, approval of the final version, and accountability for it — and have them do it again. The gap between the two answers is the lesson, and it lands harder than any definition read from a slide. Close with the practical half: the conversation belongs at the point the analysis plan is fixed, not the week of submission, and a contribution note kept in the project folder from the first experiment turns it into a reading rather than a negotiation.

Similarity report myths (30–40). The first myth to kill is that the percentage is a verdict. A report can show a high number made entirely of references, a methods paragraph the field words one way, and properly quoted material, and be fine. It can show a low number containing one paragraph lifted from somebody's thesis, and be serious. So the rule students leave with is that nobody reads the number without opening the matched sources underneath it. Second myth: the tool detects plagiarism. It detects text overlap against whatever sits in its corpus. It cannot tell whether the overlap was cited, whether the source is the author's own earlier work, or whether an idea was taken while the words were not. A person decides all three. Third myth: a clean report means a clean manuscript, which quietly teaches students to edit for the number, swapping synonyms until a match disappears. That produces worse writing and the same act. If you can get a real report on screen, show a self-match against the author's own preprint, because it is the case that alarms students most and matters least.

Peer review response role-play (40–50). Give the room one decision letter and one genuinely sharp comment, the kind that is rude and also correct. Ask for two drafts. The first is the letter they want to send, written fast and thrown away on purpose, because writing it and then not sending it teaches restraint better than being told to be professional. The second is the one that goes: quote the comment, state what you understood it to be asking, say what you changed or why you are not changing it, and point to the section and line numbers where the change now lives. Disagreement is allowed, and it rests on evidence or a stated reason, never on the assertion that something is standard in the field. Have two or three read their second draft aloud. Add the two rules that save months later — the editor's letter outranks the reviewers when they conflict, and volume is not difficulty, since four comments can be two months of work and thirty can be a week.

Resources and Q&A (50–60). Protect this. The questions that matter arrive here and they are always specific: my supervisor wants his collaborator on the paper, I have submitted to a journal I now cannot find in any index, the reviewer wants an experiment I cannot fund. Answer what you can, and when an answer needs a person rather than a policy, say so and name the person. Leave the reading list and the office-hour slot on screen while you take questions.

Who should run it

Pair someone who publishes actively in the students' field with your scholarly communication librarian. The researcher supplies the judgement calls, the real decision letters, and the credibility that comes from admitting to having made these mistakes. The librarian supplies indexing, repositories, licences and the local detail — which fund pays a charge here, which repository a deposit goes to, who to email when something goes wrong on a Friday. Either one alone leaves a gap the room can feel.

Materials to assign

• Predatory journal checklist
• Authorship
• Similarity screening
• Suggesting reviewers
• Hub: Directive Academy

Send them a week ahead with one sentence explaining why: the hour is for cases and practice, and definitions read in advance are definitions nobody has to sit through. Then assume half the room read them. That is enough, provided you have not built the session on the assumption that everyone did. Put the four definitions you actually need on a single handout page, give the room ninety seconds with it before the first exercise, and carry on.

If you add one local document to the list, make it the one that answers where this happens here: your institution's authorship policy, your library's page on deposits and publication charges, or the name and address of whoever handles integrity questions on campus. Institutional detail sticks when it arrives attached to a case the room has just argued about, and slides off when it arrives as a link in an induction pack.

Facilitation tip

Use real anonymized examples from your campus (with care). Abstract advice does not stick.

Anonymising properly means more than deleting a name. Change the field, the numbers and the timeline, and check that the case cannot be reconstructed by anyone in the room who was there — in a department of forty people the details identify someone long before the name does. If the person is still on campus and the case is recognisable, get their permission or do not use it. And keep the case about the decision. A room that is enjoying a scandal has stopped asking what it would have done.

The cases worth using are the ones where reasonable people disagree. A clearly fraudulent example teaches nothing, because everybody already knows the answer and nobody will meet it in that form. Use the borderline ones instead: the colleague who genuinely helped but not enough to be an author, the journal that is real but wrong for the paper, the reviewer who is unpleasant and also right.

If you only have thirty minutes

Keep the journal demo and the authorship case, and reduce the other two to a paragraph each with a link. Those two are where the irreversible damage happens: a paper published in the wrong outlet cannot be moved, and an authorship dispute that surfaces at proof stage can cost a year and a working relationship. Misreading a similarity report or writing a bad response letter is expensive and slow, but it is recoverable, and a student who knows where the guidance lives can usually recover without you.

Follow-up

Offer a 20-minute office hour for students with an active submission. Point them to For Authors.

Announce it in the session and fix a slot within the following week, because the real questions surface only when someone has a submission in front of them and general advice has run out. What arrives is predictable enough to prepare for: a shortlist of three journals with one bad one on it, a supervisor who has added a name, a decision letter nobody has read past the first paragraph, and someone who has already paid a fee they cannot recover.

Keep a note of what comes up. When the same question arrives three times, it belongs in the hour, and something currently in the hour probably does not. Judge the session by what students do afterwards rather than by how they rated it on the way out: shortlists built from the reference lists of their own papers, authorship conversations held while the analysis plan is still being written, revisions returned inside the deadline, and students who bring you a problem in week two rather than an emergency in month six.

More: Directive Academy · For Authors