# Authorship in International Collaborations By the end of this guide you will be able to decide who earns a byline on a multinational paper, run an authorship meeting at project kickoff, write a contribution statement in the CRediT format, record author order and equal-credit decisions in writing, and route disputes before they harden. Keep five things on hand: the authorship policy page of your target journal, your funder's collaboration guidance, the 14-role CRediT taxonomy (Contributor Roles Taxonomy, the standard vocabulary for describing who did what), a shared spreadsheet or repository file, and an ORCID iD (a 16-digit researcher identifier) for every contributor. International collaboration now drives discovery across the life sciences. Parkinson's disease research, for example, has moved from single-investigator studies to global consortia that produce most major advances [1]. Author lists grew with that shift. The 100 most cited papers in global neurosurgery, all published between 2000 and 2023, carry 811 co-authors from 61 countries [2]. Every additional border adds a language, a funding system, and a set of local norms, and each one can blur who deserves credit. This guide turns authorship ethics in international collaborations into a process you can run, not a debate you improvise. ## Why Authorship Gets Harder Across Borders Authorship problems are old. Distance makes them new in three ways. **Norms travel badly.** In some systems the head of the lab expects a senior position on every paper the lab produces. In others, alphabetical order is the tradition. A review of international Parkinson's research flags language barriers, limited interdisciplinary communication, and the need to respect diverse cultural contexts among the standing obstacles to cross-border work [1]. A rule that feels obvious in one country can feel presumptuous in another, and teams rarely discover the mismatch until a draft circulates. **Power rarely balances.** Collaborations between high-income countries (HICs) and low- and middle-income countries (LMICs) tend to concentrate decisions about authorship, data, and intellectual property on one side. A content analysis of 21 funding documents issued between 2017 and 2021 by the National Institutes of Health, the European Commission, and the Wellcome Trust found that all three funders addressed data sharing in general terms, yet none provided guidance specific to international collaborations, including the secondary use of data without prior consent from the LMIC researchers who generated it [3]. The same study cataloged authorship and intellectual property ownership as recurring equity concerns in these partnerships [3]. Production mirrors the imbalance. In a systematic review of hospital artificial intelligence research, 83 percent of the 673 included studies originated in high-income countries [4]. **Paper trails are thin.** Trainees rotate, grants end, and verbal promises about authorship evaporate across time zones. Partnership research treats written agreements and conflict management as core practices for keeping collaborations healthy, not as bureaucracy [5]. The rest of this guide builds that paperwork in about an hour. ## Authorship Criteria You Can Defend Most biomedical journals follow the recommendations of the International Committee of Medical Journal Editors (ICMJE), which define an author as someone who meets all four of the following tests: 1. Substantial contribution to the conception or design of the work, or to the acquisition, analysis, or interpretation of data. 2. Drafting the work or revising it critically for important intellectual content. 3. Final approval of the version to be published. 4. Accountability for all aspects of the work, including the confidence to answer questions about its accuracy and integrity. All four are required. A biostatistician who ran the analysis but never reads the draft fails the second and third tests. A department chair who secured the space but touched nothing else fails the first. Neither belongs in the author list, though both can be acknowledged, with their consent, for the actual contribution. Contributorship gives you a vocabulary for the middle ground. CRediT defines 14 roles: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing the original draft, and reviewing or editing the manuscript. Every author should hold at least one role, and people who hold roles but fail the four criteria belong in the acknowledgment section. Three failure modes have names. Gift authorship adds someone who did not earn the byline, usually a senior figure. Ghost authorship omits someone who did the writing, often an uncredited professional writer. Guest authorship is gift authorship with a thin pretext. Editors treat all three as misrepresentation of who did the work, not as etiquette. Large consortia solve a scaling problem. When hundreds or thousands of people qualify, the journal prints the group name as the author and lists members in a supplement or a public file. Members should confirm how the consortium's publication policy handles individual citation before assuming the paper counts toward their own record. | Decision | Main options | Best fit | |--|--|--| | Standard for the byline | The four ICMJE-style criteria, journal policy, funder policy | Read the target journal's policy first, then apply the four criteria to every candidate | | Recording who did what | CRediT statement, shared contribution log, share calculation | CRediT for the published paper, the log for internal fairness | | Author order | Contribution-based sequence, negotiated first and senior positions, alphabetical | Sequence in most biomedical fields, alphabetical in mathematics and some large physics teams | | Equal credit | Co-first footnote, co-last footnote, group author name | Footnotes for genuinely matched contributions, a group name when individual listing is impractical | | Dispute path | Corresponding author mediation, institutional research integrity office, journal integrity process | Escalate in that order and document each step | ## Authorship Ethics in International Collaborations, Step by Step Run the seven steps below in order. The first three take a single meeting. ### Step 1. Inventory every contributor and assign CRediT roles List everyone who will touch the project, from the investigator who wrote the grant to the technician who will run the gels. Assign roles from the 14-item CRediT list. A YAML file keeps the record readable by humans and scripts alike. ```yaml project: "Cross-border gut microbiome survey" authors_in_order: - "A. Abebe" - "L. Chen" - "N. Okafor" - "R. Silva" contributions: "A. Abebe": ["Conceptualization", "Investigation", "Writing - original draft"] "L. Chen": ["Methodology", "Formal analysis", "Visualization"] "N. Okafor": ["Data curation", "Investigation"] "R. Silva": ["Supervision", "Funding acquisition"] equal_contribution: [] orcid: {} # paste each contributor's ORCID iD here ``` Save the file in the project repository next to the protocol. Every collaborator can read it, and every later argument starts from a shared text. One parsing note: quote any name that contains a period or colon, because unquoted punctuation in YAML keys triggers errors such as `yaml.scanner.ScannerError: mapping values are not allowed here`. ### Step 2. Test each candidate against the four criteria ![Decision path asking whether a contributor meets the four authorship criteria, with branches to byline, revision offer, or acknowledgment.](/article-images/illustrations/authorship-ethics-in-international-collaborations-four-criteria-triage.webp) Walk the list aloud in the kickoff meeting. For each name, ask one question per criterion. Suppose a field epidemiologist collected half the samples but will not read the manuscript because of clinical commitments. That person fails the drafting and approval tests, so offer a choice: either take on revision duties and qualify, or move to the acknowledgments with a description of the actual contribution. Silent demotion breeds resentment. Explicit offers build trust. ### Step 3. Write a one-page authorship agreement Research on meaningful research partnerships, organized around the Integrated Knowledge Translation (IKT) Guiding Principles, a set of values for researching with intended users rather than on them, scanned the literature across seven practice domains including co-authorship, agreements, and conflict management, and found concrete reported practices in each, such as describing context, governance, and roles in writing [5]. Your agreement does not need legalese. ```text AUTHORSHIP AGREEMENT Project: Cross-border gut microbiome survey Date: 2026-03-02 1. Authorship follows the four criteria used by our target journal. Every listed author must meet all four. 2. Draft order at kickoff: Abebe, Chen, Okafor, Silva. Order will be reviewed when the first full draft exists. 3. Roles follow the CRediT file in the repository (version 1). 4. Any addition or removal of an author requires written notice to all current authors and a recorded reason. 5. Disputes go first to the corresponding author, then to the research integrity office at each signatory's institution. Signatures: Abebe, Chen, Okafor, Silva ``` The corresponding author, the single contact point who handles journal correspondence and signs the integrity statements, should be named deliberately here rather than defaulted to whoever uploads the files. ### Step 4. Track contributions and compute shares as a talking aid ![Worked calculation dividing each author's CRediT role count by the summed role counts across four authors.](/article-images/illustrations/authorship-ethics-in-international-collaborations-credit-share-example.webp) A shared log turns vague memories into rows. Bibliometric studies use fractional counting, splitting the credit for each paper among the units that produced it, to compare collaboration fairly across countries [6]. You can borrow the idea at team scale. ```python roles = { "A. Abebe": ["Conceptualization", "Investigation", "Writing - original draft"], "L. Chen": ["Methodology", "Formal analysis", "Visualization"], "N. Okafor": ["Data curation", "Investigation"], "R. Silva": ["Supervision", "Funding acquisition"], } counts = {name: len(assigned) for name, assigned in roles.items()} total = sum(counts.values()) shares = {name: round(count / total, 2) for name, count in counts.items()} print(counts) print(shares) ``` Expected output: ```text {'A. Abebe': 3, 'L. Chen': 3, 'N. Okafor': 2, 'R. Silva': 2} {'A. Abebe': 0.3, 'L. Chen': 0.3, 'N. Okafor': 0.2, 'R. Silva': 0.2} ``` Here `counts` tallies roles per person, `total` sums them, and `shares` divides and rounds to two decimals. Check the result by confirming the shares sum to roughly 1.0 (rounding may leave you at 0.99 or 1.01) and by cross-checking each row against the YAML file. If you rename an author in one place only, Python raises `KeyError: 'L. Chen'`. Keep a single source file and import it rather than retyping names. Treat the numbers as a conversation starter, not a verdict. One hour of conceptualizing can outweigh ten hours of routine data entry, and the share cannot see that. The point is to surface disagreements while they are still cheap to fix. ### Step 5. Fix the order rule and the equal-credit notation In most biomedical fields the first author made the largest direct contribution and the last author supervised the work, a convention often summarized as sequence determines credit. Mathematics and many large physics collaborations use alphabetical order instead. Pick one rule, write it in the agreement, and use footnotes for ties: a symbol beside two names with the phrase "these authors contributed equally" marks co-first authors, and a parallel note marks co-last supervisors. Record an ORCID iD for every author, since the identifiers distinguish people who share names, which matters when names transliterate from other scripts. ### Step 6. Revisit the list at three checkpoints Review the author list when the first full draft exists, when the journal sends revisions, and before submission. Editors commonly require the corresponding author to confirm that every listed author approved the final version, and some ask why an author joined after the initial submission. A documented reason from your log answers that question in one line. ### Step 7. Escalate disputes in a fixed order Start with the corresponding author mediating between the parties. If that fails, take the disagreement to the research integrity office at each signatory's institution. Journals are the last stop, and most follow published flowcharts for authorship disputes. Skipping straight to the editor weakens your position, because editors expect teams to exhaust internal channels first. Once these seven steps are habits, the two sections below cover the places international teams stumble most often: order and equity. ## Author Order, Equal Credit, and Group Authorship Order carries different weight in different career systems. In some countries hiring committees read the middle of the author list closely. In others only first and corresponding positions register. When teammates span those systems, state the reasoning for the chosen order in the agreement, not just the order itself. Middle authors who feel invisible are the most common quiet grievance in multinational teams. Equal-contribution footnotes work only when the two contributions really are interchangeable. Reviewers and hiring committees increasingly discount reflexive co-first notations, so reserve them for genuinely matched work and say in the footnote what each person led. Group authorship has its own mechanics. The consortium name appears in the byline, the member list lives in a supplement or public repository, and individual members usually cite the paper as the group name followed by their own name, per the consortium's stated policy. Read that policy before the paper is submitted, because retroactive claims are hard to support. ## What Journals and Editors Expect at Submission Submission systems enforce authorship mechanically. Most journals now require an explicit contribution statement, many in CRediT format, and the corresponding author must confirm that every listed author approved the final manuscript. Publishers increasingly require an ORCID iD for the corresponding author at minimum. Some journals ask for justification when the author list runs long, and each journal specifies how group authorships must be formatted. Changing the list after acceptance is the hardest move in publishing: most journals require a written request explaining the change, signed by every existing author, and they treat unexplained swaps as integrity flags. Freeze the list before submission rather than negotiating against the typeset proof. ## Responsible Conduct of Research and Shared Oversight Authorship sits inside responsible conduct of research (RCR), the umbrella term for the professional standards covering [data management](/blog/guides/data-management-basics-principles-processes-and-best-practices), plagiarism, mentorship, and the allocation of credit. Most graduate programs now teach it, and authorship is the module where international teams hit trouble first. A workshop report from the 2022 Global Forum on Bioethics in Research argued that ethical conduct is the responsibility of every stakeholder in the research ecosystem, from funders and governments to institutions, research teams, and [research ethics](/blog/guides/research-ethics-fundamentals-a-practical-guide-for-early-career-scientists) committees, and called for distributed approaches to ethical scrutiny built on collaborative rather than adversarial relationships between researchers and committees [7]. Applied to authorship, that logic moves the load upstream. Instead of treating the byline as a private fight settled at submission, funders write expectations into opportunities, institutions train for them, and teams document decisions as they go [7]. Data stewardship is the companion duty. Because funders have left a gap on data management within international collaborations, including secondary use of data without consent from the researchers who generated it, teams should contract for data ownership and reuse terms before the first sample ships [3]. Authorship disputes often trace back to data disputes, and the cheapest time to settle both is before collection starts. ## Building Equitable Partnerships With Colleagues Abroad The Parkinson's review offers a candid list of what strains international teams: funding inequities, regulatory and data-sharing complexity, difficulties with standardization, unequal resource allocation, insufficient training opportunities, and regional mistrust [1]. The same review points at the fixes that worked: shared databases, harmonized protocols, cross-border funding mechanisms, and partnerships grounded in trust [1]. Funders are one lever. Because the analysis of NIH, European Commission, and Wellcome documents found authorship and capacity development already appearing in funding language, applicants can cite those expectations when negotiating authorship within a consortium [3]. Practice on the ground looks like the spine surgery missions described from Jamaica, where success depended as much on sustained partnership with host institutions and local teams as on logistics, and where training and research opportunities anchored long-term collaboration [8]. Five tips that prevent most inequity disputes: 1. Name site leads as authors when they meet the criteria, and design the work so that they do. 2. Budget time for colleagues writing in a second language, and count substantive revision as qualifying authorship work. 3. Put data ownership, reuse rights, and benefit sharing in the same agreement as the byline. 4. Rotate visible roles, such as presenting at the consortium meeting, so credit is not always harvested in one country. 5. Commit to capacity building in writing, with named trainees and a timeline. ## Common Mistakes and Limitations - **Adding the department chair at revision stage.** Gift authorship, once flagged by an editor, can trigger a correction or a misconduct inquiry that stains every author on the paper. - **Deciding order at submission.** By then, leverage has replaced judgment. Set the rule at kickoff. - **Relying on verbal promises across institutions and time zones.** If it is not written down, it did not happen. - **Assuming your home country's norms are the defaults.** Ask about order and acknowledgment conventions in the first meeting, explicitly. - **Leaving data ownership undefined until after collection.** This converts a data dispute into an authorship dispute, exactly the gap the funder analysis documented [3]. - **Treating a contribution share calculation as objective truth.** It is a prompt for discussion, not a measurement instrument. - **Naming acknowledged contributors without their consent.** Some people decline acknowledgment for professional or political reasons, and journals require their approval. - **Defaulting the corresponding author slot to whoever uploads the manuscript.** Assign the role deliberately and record it. On limitations: journal and funder policies differ, and this guide describes common practice rather than any single rulebook, so read the specific policies that govern your paper. Individual disputes can turn on facts that no general guide can weigh. When the stakes include a degree, a grant, or a career, take the file to your institution's research integrity office. ## Frequently Asked Questions ### What qualifies someone as an author on an international paper? Meeting all four criteria the target journal uses: substantial contribution to the work, drafting or critical revision, final approval, and accountability. Providing samples or funding alone earns an acknowledgment, not a byline. ### Who decides author order in a multinational team? The team decides, ideally in a written agreement at kickoff. Most biomedical fields put the largest direct contributor first and the supervisor last, but any recorded convention works when everyone consents. ### How do co-first authors show equal credit? With a footnote symbol beside both names and a statement such as "these authors contributed equally." Confirm the journal prints the statement, because formatting varies. ### What should I do when a co-author stops responding before submission? Document every attempt to reach them with dates, inform the remaining authors, and notify the editor of the non-response. Journals can proceed when the record shows reasonable effort. ### Are authorship rules the same in every country? No. Journal criteria converge because journals enforce them, but local norms on order, senior positions, and acknowledgment differ, which is why written agreements matter. ### Can a lab head or funder be listed as an author? Only by meeting all four criteria. A principal investigator who designed the work and revised the manuscript qualifies, while a program officer who administered funding does not. ### How do I list a large consortium paper on my CV? Cite the group name as the author, then add your own name if the consortium's citation policy allows it. Keep a copy of that policy with your records. ### When should an authorship dispute leave the research team? After internal mediation fails, or immediately if the dispute involves alleged misconduct such as a fabricated contribution. Institutional research integrity offices handle both, and journals expect internal channels to be tried first. ## Sources 1. [Global research synergy: Advancing fundamental Parkinson's research through international collaboration.](https://pubmed.ncbi.nlm.nih.gov/42627362/) 2. [The top 100 most cited peer-reviewed publications on global neurosurgery: Mapping research trends and gaps over four decades.](https://pubmed.ncbi.nlm.nih.gov/42215816/) 3. [The Role of Research Funders in Promoting Ethical International Health Research Collaborations.](https://pubmed.ncbi.nlm.nih.gov/41988915/) 4. [Responsible artificial intelligence in healthcare: a systematic review on the use of ethical principles in the development and deployment of artificial intelligence.](https://pubmed.ncbi.nlm.nih.gov/42712331/) 5. [Identifying practices for meaningful health research partnerships: series of rapid environmental scans.](https://pubmed.ncbi.nlm.nih.gov/42934760/) 6. [Mapping global research on labor and postpartum pain management: a bibliometric analysis (2000-2025).](https://pubmed.ncbi.nlm.nih.gov/41840591/) 7. [Rethinking ethical reflexivity and oversight in health research through an ecosystem approach: A workshop report.](https://pubmed.ncbi.nlm.nih.gov/42454216/) 8. [Spine Surgery Missions in Low-Resource Settings: Practical Considerations and Lessons Learned.](https://pubmed.ncbi.nlm.nih.gov/42777256/)