It can be difficult to sort through feedback on a paper/chapter when you have feedback from multiple collaborators. Sometimes collaborators might have a conflicting point of view, or making all of the improvements suggested by each would make the paper far too large.

What are some strategies on working through this feedback efficiently? Should you fully address each collaborator's comments one by one, attempt to merge all the comments together before editing the manuscript, or take some sort of hybrid approach?

3 Answers 3


It's a tricky question and I think the best strategy will ultimately vary from person to person. Nevertheless I have two concrete suggestions that has helped me a lot throughout the writing phases of my manuscripts and thesis.

  1. Get feedback in "rounds" instead of "continuously". When you have a version of the text on which you want feedback, send that to all involved parties at once. That prevents people commenting on each others suggestions and making a mess out of everything in the markup.

    Once you have feedback from multiple people, you can iteratively go through the feedback and add/ignore the changes on the master document. This is rather contradictory to the accept/reject change function on MS Word, but in my experience that's not really sustainable when you have a lot of cooks making a single soup.

    After the "round" is done, remove most (if not all) markup from people, have a new version of the document with whatever else you want to add/remove. If more feedback is needed have a new round.

  2. If there are co-authors who are more superior to you in the academic ladder and you are concerned about how to handle any contradictory/problematic comments, define one of your superiors to be the "alpha" and ask that persons help to try and settle disputes. It should be relatively easy in some fields where the last author usually has the ultimate responsibility for the project.

    Alternatively, if you have a supervisor/mentor who's gonna be a co-author but not the last; you can try to reason with that person to settle disputes. I have ended up in that situation where one of our co-authors decided to contradict with her own comments from before, and had these endless iterations of comments and changes. When I couldn't take it anymore, I talked to my supervisor and which point he took over the reigns and defined a new, final version prior to submission on which he asked all co-authors to have their final remarks.

It might appear as a bit totalitarian solution, but at that point the alternative was me completely dropping the project based on this one co-authors unexplainable and unbearable behavior.

  • I'm curious, in which fields is the last author the most important? In the ones where the order is relevant, it is always the first; the last author being either last, or second.
    – Davidmh
    Sep 15, 2015 at 6:25
  • @Davidmh Hard to give an exhaustive list, i'd say medical, biological and chemical sciences for sure. Also take a look at this question
    – posdef
    Sep 15, 2015 at 9:08

While your question seems to be on-topic, since it is formulated in the context of academic activities, I doubt that there exist any domain-specific strategies beyond general ones to handle feedback from people (or, generalizing, processing information streams) via analysis and synthesis.

Having said that, I think that there are various possibilities in dealing with feedback from multiple people, but they all stem from general main problem of handling conflicting requirements (obvious edge cases of integrating non-conflicting requirements are rather trivial and are solved by merging information, based on all feedback, provided each piece increases the value of the target artifact).

The above-mentioned possibilities and corresponding strategies IMHO can be based on different perspectives, from philosophical (i.e., thesis, antithesis, synthesis) and other high-level approaches (i.e., Collaborative Information Synthesis) to the ones, based on business practices (i.e., business requirements analysis, including impact, scenario and stakeholders analysis) and software engineering practices (i.e., trade-off analysis) to more technical, such as multiple-criteria decision analysis (the latter approach is likely an overkill for handling traditional academic feedback, but included here for completeness; plus, it might be used for creating automated feedback systems).


I got privileged to work with some collaborators (some of them, I didn't get chance to meet personally). I got some good quick results instead of waiting for next round. Here is one strategic plan as suggested by one of my superior collaborators:

  1. Keep all the conversation in a single thread.
  2. After getting response from one of the team (it is obvious that all the team members are busy with their academic work, they are really helpful to make your paper one level up), start working on it. Now, there is an issue how to deal with title of the manuscript. I followed V.round-name.iteration, thus it would be like V.2.1.
  3. Update all members when you have finished your task. Highlight the updated section or text. If you have something to say, comment it on the thread or in the manuscript clearly. Make your own identity while commenting.
  4. If you have any ambiguity, then discuss with all the team member very preciously. You may be curious to know that why other team members are not responding! They are listening, they will respond when they feel that their time has come to make it another level up.
  5. A good collaboration does not always mean to write the whole work in hand by the all team members, you have to take charge of that, a simple comment from a experienced person can be helpful to get you paper accepted.
  6. If you have any contradictory statement made by others, then try to digest it, if not, then send them to decide what to do. But do not send as it is, make some own comments, then only the team will realize that you are working instead of just paper published.
  7. Do not wait for receiving response from all members. That will kill your time. Work with the response received from atleast one of the individual.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .