The paper has survived its second round of review when a colleague emails: “I helped shape the proposal at the start. Why isn't my name on it?” The lab head wants to add the department chair “to give it weight”. The person who ran every analysis has been moved to the acknowledgements. Authorship disputes are rarely technical. They happen because nobody agreed anything at the start.
The good news is that publishing has a reasonably clear set of criteria. Knowing them gives you a basis for a frank conversation with people above and below you.
The four ICMJE criteria: all four, not one
The International Committee of Medical Journal Editors (ICMJE) sets out four criteria, now widely borrowed beyond medicine. A person qualifies as an author when they meet all four:
- Substantial contribution to the conception or design of the work, or to the acquisition, analysis or interpretation of data.
- Drafting the work or revising it critically for important intellectual content.
- Final approval of the version to be published.
- Agreement to be accountable for all aspects of the work, ensuring questions of accuracy or integrity are investigated and resolved.
The fourth criterion is the one people forget and the one that matters most: authorship means accountability. If the paper is retracted over flawed data, every co-author's name is attached to that retraction. That is a good reason to decline authorship on a paper whose results section you cannot follow.
Contributions that earn an acknowledgement instead
In the spirit of ICMJE, the following on their own do not justify authorship:
- Securing funding or general supervision of a research group.
- Administrative support, language editing or formatting.
- Providing samples, equipment or an existing dataset without taking part in interpretation.
- Commenting once on a proposal.
These people belong in the acknowledgements, with a note on what they did. Ask them first: being acknowledged can imply endorsement of the conclusions.
Gift, ghost and coerced authorship
| Type | What it is | Why it harms |
|---|---|---|
| Gift or honorary | Adding a senior or well-known name with no real contribution | Dilutes accountability; the added person may carry errors that are not theirs |
| Ghost | The real writer is not listed, often when a sponsor drafts the paper | Hides influence and conflicts of interest |
| Coerced | A superior demands a slot as a condition of continued support | Abuse of a power relationship |
| Reciprocal | “I'll add you to this one if you add me to yours” | Inflates both records |
Where promotion depends on publication counts, the pressure to add names is real. The most effective way to defuse it is to move the conversation onto paper: journals increasingly ask for each author's role, and a name that cannot be matched to a role stands out.
Declaring roles with CRediT
CRediT, the Contributor Roles Taxonomy, lists 14 standard roles, including Conceptualization, Methodology, Software, Formal analysis, Investigation, Data curation, Writing – original draft, Writing – review & editing, Supervision and Funding acquisition. Many journals print these statements in the article itself.
The practical benefit cuts both ways. People who build software or curate data, long overlooked, now have a visible role. Someone whose only role is Funding acquisition will find it hard to justify an author slot against the four criteria.
Author order: conventions differ by field
There is no global standard, so know your discipline's habits:
- Biomedical and natural sciences: the first author did the most work; the last author is usually the senior lead; the corresponding author handles communication with the journal.
- Mathematics, theoretical economics, parts of high-energy physics: many groups list alphabetically.
- Social sciences: usually descending order of contribution.
When two people contributed equally, a joint first authorship note is common. Promotion committees generally understand it, but check how your own institution credits a shared first position.
An authorship agreement: twenty minutes at the first meeting
The most effective way to prevent disputes is a written agreement at the start of the project, updated as roles change. A lean version covers:
- Everyone involved and their expected CRediT roles.
- Provisional author order and the rule for changing it (for example, whoever rewrites the results after a change of analysis moves up).
- Who will be the corresponding author.
- What happens if someone leaves the project midway.
- Plans for later papers from the same dataset, to avoid a scramble down the line.
Circulate it by email and ask everyone to confirm. A short dated email carries far more weight than a verbal promise when a disagreement needs resolving.
When a dispute has already started
- Talk directly first, anchored in the four criteria and actual roles rather than a sense of effort.
- If that fails, bring in a neutral party: a head of department, an ethics committee or your institution's research integrity office.
- Journals do not adjudicate internal disputes, but they will pause processing or require every author to sign off on changes to the author list after submission. COPE's flowcharts describe exactly this step.
Your next step: if you have a project that has not been submitted yet, open a shared document, list each person's CRediT roles and ask the team to confirm this week. Every difficult conversation that follows will be shorter.
Câu hỏi thường gặp
Does my PhD supervisor automatically get authorship?
Not automatically. Supervisors qualify when they contribute substantially to design, interpretation and revision, and accept accountability. Many do meet the criteria, but it should be confirmed rather than assumed.
Does editing the English of a paper count as authorship?
Usually not. Language and formatting help deserves thanks but is not critical revision of scientific content. Acknowledge the person instead.
What is the difference between first author and corresponding author?
The first author typically did the most work. The corresponding author communicates with the journal and readers, and may be the first author or the senior lead.
Can I add an author after submission?
Yes, but journals will ask for a reason and written consent from every existing author. The later the change, the more scrutiny it attracts.
What is the CRediT taxonomy?
It is a standard list of 14 contributor roles. Many journals require CRediT statements; where they are optional, using them still improves transparency and reduces disputes.
