You are here

Case Discussion: Withdrawal of paper at proof stage

Case Summary

Case 19-11

An original paper was submitted to a journal, peer reviewed, revised by the authors, and accepted by the manuscript editor. It was scheduled for publication 3 months later. After the paper was copyedited and typeset, the corresponding author was informed of the acceptance and was asked to proofread the article. After 2 weeks, however, the corresponding author requested that the paper be withdrawn. The chief editor asked for an explanation and brought the case to COPE Forum, noting that the journal does not charge article processing fees.

Question(s) for the COPE Forum

  • Significant resources have been provided to improve the paper. Should we accept the withdrawal?
  • Is there anything else we should do if the authors do not respond to our request for an explanation?

Forum Advice

The Forum asked why the authors were not informed of acceptance before copyediting, and whether they had assigned copyright or agreed to publication. The authors could be asked if the journal process was the reason for withdrawal, or if they had discovered a critical error in the paper. Ultimately, authors may ask to withdraw a paper and editors have to honour the request. The journal should review its processes and ensure communications with authors are clear, perhaps making use of COPE’s journal audit tool for members, at


Case Discussion

This archived COPE Forum case is categorised under two COPE Core Practices:

  • Allegations of Misconduct, which states: “Journals should have a clearly described process for handling allegations, however they are brought to the journal's or publisher’s attention. Journals must take seriously allegations of misconduct pre-publication and post-publication. Policies should include how to handle allegations from whistleblowers.”
  • Journal Management, which states: “A well-described and implemented infrastructure is essential, including the business model, policies, processes and software for efficient running of an editorially independent journal, as well as the efficient management and training of editorial boards and editorial and publishing staff.”

The misconduct that might be involved depends on the corresponding author’s reason for the late withdrawal. Some situations may justify withdrawal but, akin to rejecting a manuscript submission because of misconduct, the editor also has a duty to follow up (eg, with the institution and possibly other editors). The level of publishing experience of the author can be taken into account, as well as the possibility of honest error in the publication process or in the paper’s content such that the conclusions are invalidated. COPE’s flowcharts can be helpful here, in addition to guidance such as “Cooperation Between Research Institutions and Journals on Research Integrity Cases: Guidance from the Committee on Publication Ethics (COPE)”.

Reasons may include prior submission to another journal, which may publish faster or have a (higher) journal impact factor. In addition, the paper or a similar one might have already been published, representing redundant publication during submission. The Forum asked if the authors had actually agreed to publication. If the authors disagree with the original or revised submission, or if there are false claims of authorship, correction may be an alternative to withdrawal, but the corresponding author’s actions would need investigation. If any authors do not agree with the edited and typeset version, the author dispute/s would need to be resolved by the authors and re-editing could be negotiated with the journal. If the corresponding author loses faith in the integrity of the work and suspects co-author misconduct such as fabrication, falsification, plagiarism, unethical research, or bias from a conflict of interest, these issues also need to be pursued. The institution may be requested to investigate or review their training in scholarly publishing.

The authors might confess that they plan to take their now-improved paper to a journal that has a (higher) journal impact factor.  A previous case brought to COPE Forum (case 17-19, Unethical withdrawal after acceptance to maximize the 'impact factor') describes how an author requested withdrawal of an accepted paper so as to transfer it to a journal in the publisher portfolio that had a journal impact factor. The publisher wanted to ban all the authors from submitting to their journals again, as well as levy the full article-processing charge (APC) for late withdrawal “in order to play impact factor games”. After the author’s institution investigated the case, the publisher attributed the misconduct to author inexperience. Although the author insisted on withdrawal, the publisher accepted the Forum’s advice of adopting an educational approach and did not charge the APC.

An educational approach could be used in the present case if the situation is similar. However, the editor noted that the journal does not have an APC. Unfortunately, some authors may take advantage of this operational model to try to receive free peer-review guidance (and even editing) of submitted papers. In discussing case 18-11 (Increased number of casual submissions), the Forum suggested implementing a small submission fee to cover administrative costs and deter “casual” or low-quality submissions. The fee policy and amount must be stated clearly in the submission instructions.

If the corresponding author does not reply, the editor would have to postpone publication of that article because all authors need to first approve publication. Likewise, a paper cannot be withdrawn without a written request signed by all authors that includes their reason for withdrawal. The editor would have to keep trying to obtain a withdrawal letter from the corresponding author and if that fails, a co-author could be asked. It is possible that not all the authors know about the intended withdrawal and they may eventually reverse the decision. The editor would also have to remind the authors that they cannot submit the paper elsewhere until the journal sends them a formal withdrawal notice. Deadlines should be provided and if there is still no reply, the institutions could be asked to try to trace the author/s. The paper may need to be withdrawn by default after a final deadline, and the authors/institutions would be informed of the disappointing outcome and reminded of expected future publishing etiquette.

While waiting for an author reply, the editor might wish to do Internet searches to see if the paper or a similar one has been published elsewhere and to check for plagiarism. The situation may be more complicated if the journal routinely assigns a DOI (digital object identifier) to accepted but unedited manuscripts and uploads them as an online-first version to their website and/or a repository. The timestamp would indicate whether the journal published the paper first. The editor/s of other publications then need to be informed.  As outlined in the COPE Retraction Guidelines, instead of issuing a Notice of Redundant Publication in the first published version and Notice of Retraction in the later version/s, the editor could use an Expression of Concern to alert readers of possible redundancy and ongoing investigation. If the online-first version is the later (redundant) one, or contains plagiarism, an Expression of Concern could be issued while the editor investigates whether withdrawal is appropriate. Journals may have different policies for withdrawing/retracting online-first documents from their journal and/or repositories, compared with retracting the final version (version of record), which is not removed but appropriately watermarked. However, specific policies should exist and be disclosed on journal websites.

There are some legal implications. As raised by the Forum, the editor would need to check whether the authors had assigned copyright, if applicable, or a publishing licence or other publishing agreement. The wording and conditions of such documents need close scrutiny, particularly related to manuscript version/s at submission, acceptance, or publication, and what happens at withdrawal or retraction. In a previous case (19-06, Dual submission and editor’s failure to take action), the copyright for a paper published in a journal was held by another journal because the authors did not formally withdraw their earlier submission.

This case illustrates what might happen if editorial office processes and communications are unclear or inconsistent. The second relevant COPE Core Practice is that of Journal Management. COPE recommends that all processes and policies be transparently and publicly disclosed on a publication’s website and that all publishing personnel be well trained. Other places where relevant policies and guidelines can be mentioned or linked to, as a reminder, include the manuscript submission platform, official forms (eg, author declarations, copyright transfer if applicable), and correspondence at key stages (eg, submission acknowledgement, first peer-review decision, acceptance letter).

There should be clear processes and policies for all areas of scholarly publication, including manuscript submission, peer review, publication, and post-publication stages. Guidelines should include topics such as authorship, intellectual property, maximising reproducibility and minimising bias, withdrawals/corrections, complaints/appeals, business practices, and any fees and waivers. In fact, publishers and editors should check they adhere to the all 10 COPE Core Practices, which are aligned with the 16 international Principles of Transparency and Best Practice in Scholarly Publishing.

Although COPE no longer publishes any codes of practice/conduct per se, publishers and publications are expected to use and link to the COPE Core Practices, guidance, and resources when creating their own guidelines and codes of practice/conduct. COPE publisher and editor members can routinely check their publication policies against the COPE Journal Audit and go through the COPE e-Learning modules. Non-members and novice editors can use COPE’s guidance contained in educational blogs in past COPE Digests, the infographic General Approach to Publication Ethics for the Editorial Office, and A Short Guide to Ethical Editing for New Editors.

Authors, and publication staff, should be clearly informed of the journal’s processes and kept up-to-date with the progress of submission and production, so as to ensure ethical and timely publication. The staffing model and business model used and who communicates with the authors should be explained on the journal’s website, together with ethics guidelines, typical workflows, timelines, and average processing and production times. The presented case contains a few red flags to do with journal procedures that may need review and improvement. Indeed, the Forum queried whether the journal process was to blame for the withdrawal decision.

First, the chief editor told the Forum that the “manuscript editor” accepted the manuscript after peer review and revision. This suggests that editorial decision-making is delegated to others. Different journals do have different gate-keeping structures, and the model used should be clearly explained to all. Commonly, each manuscript is assigned to a handling editor (also named academic editor, section editor, or manuscript editor, although the latter term is also used for staff copyeditors). The handling editor could be the chief editor, a senior/associate/assistant editor, an editorial/advisory board member, or a staff editor such as a managing/executive editor. This person identifies and may contact peer reviewers, assesses peer review reports, and makes a recommendation for the final decision on a paper. A handling editor who is not the chief editor might be given authority to make the final decision, referring to the chief editor only if there are problems. The handling editor may also correspond with the authors, or this is done by a staff or managing/executive editor.

It is unclear in this case if and when the authors were informed of manuscript acceptance and whether they were told of the publication plan. If the peer review, revision, and rereview process already took a long time, then adding 3 months to the timeline before notifying an author of the outcome is unreasonable. Authors should also be provided with a contact method and dedicated contact person to enquire about their submission, and replies should be timely. Author instructions should explain that a formal rejection notice is needed before submission elsewhere. Novice authors may mistakenly believe their paper has been rejected and may submit elsewhere if they have not been informed of the outcome after a few months and cannot successfully reach the editorial office.

Second, it is unclear in this case why it took the corresponding author 2 weeks to contact the office after receiving the proofs. Authors are commonly given only 48 hours to approve their proofs, although the time can be extended if there has been substantive in-house copyediting or if they need more time. If no deadline or comprehensive instructions were given, the author may not have known it was possible to ask for more time or for a re-edit. The in-house editing may have been extensive and present a challenge for all authors to check and comment on, especially if they did not agree to all the edits. The method of marking up proofs may have been unclear or too difficult to do. Periodically checking in-house production processes, communications, and user experience is recommended.

Finally, the journal should be proactive in checking its preventive measures to minimise misunderstanding and future problems. Staff workflows and training materials can be reviewed; staff can be encouraged to subscribe to COPE Digest at the bottom of the COPE website. Emails to the corresponding author could be copied to all authors, and author contribution declarations could be collected. The presence or absence of any simultaneous submissions, prior presentations/publications of similar work, and any posted preprints should also be declared. Journals should have a clear policy on unique submission, what counts as prior publication, and rejection and withdrawal procedures. If anomalous cases or surges in misconduct arise, the editor can use an editorial to remind potential authors of journal policies and the conduct expected of both authors and editorial staff.

Trevor Lane on behalf of the COPE Education Subcommittee


In our March issue of Digest Deborah Poff, COPE Chair, shares guidance and cases relevant to anyone who is new to publication ethics issues or needs a refresher. Following our recent Forum, we'd love to hear your views relating to editing of peer review comments in our survey. The results of the survey, together with the Forum discussion and comments on our website, will inform a COPE discussion document on the topic bringing together the shared views. Plus the latest publication ethics news as gathered by COPE Council.

Read March Digest: Publication ethics