Five researchers at three universities are writing a paper together. Two months in, the shared folder holds “draft-final”, “draft-final-2”, “draft-final-Marco-edits” and “final-REAL”. Two people are revising the same Discussion in parallel on different copies. A senior author emails comments on a version three weeks out of date. The paper is submitted four months late, not for lack of ideas but because of process.
Co-writing is a different skill from writing alone. You still need to write well, but you also have to manage documents, time and negotiations between people with different views, rhythms and seniority.
Step 1: appoint a lead writer
Every multi-author paper needs one person with final responsibility for the text, usually the first author. They need not write everything, but they hold the master file, decide when comments conflict, keep the voice consistent and set the schedule. Without this role, the paper drifts towards whoever commented last.
Authorship, author order and individual contributions should be agreed separately, early, and in writing. This article concentrates on the writing process.
Step 2: agree the outline before anyone writes a sentence
The most important meeting is the one that agrees a detailed outline, not the one that reviews a draft. The outline should include:
- The research question and core message in one or two sentences.
- Target journal, word limit and citation style.
- Section headings with a few bullet points and a target length each.
- A list of tables and figures.
- Who writes what, and the deadline for first drafts.
Major disagreements about theory or interpretation should surface here. Discovering them once the draft reaches 8,000 words costs many times more.
Step 3: one master document
| Approach | Suits | Watch out for |
|---|---|---|
| Shared online document (Google Docs, Word online) | Teams comfortable working simultaneously in Word-style files | Your reference manager’s plugin must work on the shared platform |
| Overleaf | LaTeX papers with heavy notation | Co-authors new to LaTeX can use comments rather than edit source |
| Word file circulated, one keeper | Teams where someone cannot work online | Only one person edits the master at a time; strict version naming |
| Git with plain-text files | Computational teams used to version control | Excellent for code and data, less friendly for prose |
If you must pass files around, use names with a date and editor’s initials: “2026-09-24_paper_v05_TA.docx”. Never put the word “final” in a filename.
Step 4: edit directly or comment
- Tracked edits for wording, grammar and trimming.
- Comments for content, argument and structure: “Does this contradict Table 3?” The lead writer decides how to act.
- Do not rewrite a colleague’s whole paragraph without discussion. Offer your rewrite in a comment.
- The lead writer accepts or rejects changes, answers each comment and resolves it. Unresolved comments pile up until nobody can read the document.
Step 5: feedback rounds with deadlines and scope
Every round needs three things: which version, the deadline and the kind of feedback wanted. For example: “Version 06, comments by Friday the 3rd. This round we only need feedback on the argument in the Discussion, not wording.” Agree at the outset that silence by the deadline counts as approval.
A typical schedule for an empirical paper: full draft, content feedback round, revision, wording feedback round, final read and sign-off by all authors. Send each version to everyone at once rather than to one person after another, unless there is a clear reason.
Step 6: one pass to unify the voice
Multi-author papers often show their seams: the Methods reads nothing like the Introduction, and one concept goes by three names. Once content is stable, the lead writer reads the whole paper once purely to:
- Standardise terms and abbreviations (a short team glossary agreed early helps a lot).
- Standardise tense, person (“we” or passive) and number formatting.
- Check cross-references: the questions posed in the Introduction are answered in the Results.
When co-authors disagree
Disagreement about content is normal and often improves the paper. What works:
- Separate disagreements about facts (which can be checked against the data) from disagreements about interpretation.
- For interpretation, consider presenting both readings with the evidence for each.
- Call or meet briefly instead of trading long emails; record the decision in the shared document.
- If agreement is still out of reach, the lead writer decides within the agreed framework. Any author retains the right to withdraw their name if they cannot accept the final content.
Before submission
- Every author has read the final version and approved it in writing; an email is enough.
- Each author has checked their name, affiliation, ORCID, contribution statement, conflicts of interest and funding details.
- The master document is frozen as a fixed version before submission.
For a project already under way: send your co-authors a one-page note with four lines: who leads, where the master file lives, the naming convention, and the next two feedback rounds with deadlines. That single page often saves weeks.
Câu hỏi thường gặp
Who should lead the writing of a multi-author paper?
Usually the first author. They keep the master file, decide when comments conflict, keep the style consistent and manage the schedule, even if they do not write every section.
What tools are best for co-writing a research paper?
Shared online documents such as Google Docs or Word online suit most teams, while Overleaf suits LaTeX papers. The most important thing is having a single master document.
Should I edit directly or comment on a co-author’s draft?
Use tracked edits for wording and comments for content, argument and structure. Avoid rewriting a colleague’s entire paragraph without discussing it first.
What if co-authors do not respond to drafts?
Agree at the start that each round has a deadline and that no response counts as approval. Before submission you still need every author to confirm the final version.
How should disagreements between co-authors be handled?
Separate factual disagreements from interpretive ones, talk directly rather than through long email threads, and record decisions. If consensus fails, the lead writer decides within the agreed framework.
