One of the hardest decisions in scientific writing is knowing when a manuscript is actually finished.
There is almost always another experiment that could be performed, another analysis that could be added, or another paragraph that could be refined. Even when the data are solid and the manuscript has been revised repeatedly, pressing Submit can feel surprisingly difficult.
I encountered this often when I ran an academic research laboratory. Trainees sometimes assumed that a paper was finished only when every possible question had been answered. Early in my own career, I was not immune to the same problem. I could keep revising a manuscript that was essentially complete, looking for one more change that would make it finally feel finished.
Over time, I learned that a manuscript rarely announces that it is done. At some point, completion becomes a judgment call. The important question is not whether the manuscript is perfect. It is whether the manuscript is complete, coherent, and defensible.
A complete manuscript does not answer every possible question
Scientific research is inherently open-ended. A good study frequently generates as many new questions as it answers. That means completeness cannot be defined by exhaustion. If every interesting question raised by a study had to be resolved before publication, many papers would never be submitted. Instead, ask what question the manuscript is actually trying to answer.
A manuscript is approaching completion when its central question has been addressed, the evidence supports a clear contribution, and the story holds together on its own terms. Additional experiments may still be interesting and scientifically valuable, but that does not necessarily mean they belong in the same paper. This distinction can be surprisingly difficult to recognize from inside a project.
I remember lab meetings in which a trainee would present a series of experiments and describe the project as unfinished. Halfway through the discussion, however, it would become apparent that the story was already there. The central question had been answered. The remaining experiments addressed related questions, but they were really the beginning of the next paper rather than the end of the current one.
Follow-up papers are not a concession. Separating related but distinct questions can allow each to be addressed with appropriate depth and can make the original manuscript more focused. The question is therefore not simply, “What else could we do?” In science, there is almost always something else you could do. A better question is, “What else do we need to do for the claims in this manuscript to be supported?” Those are very different standards.
Your claims should match your data
Before submission, look carefully at the relationship between the data and the claims you make about them. A defensible manuscript does not need to present the final word on a subject. It does need to be accurate about what the study demonstrates. That means conclusions should not extend beyond the evidence. Non-significant results should not quietly become “trends” simply because they point in the desired direction. Associations should not become causal relationships. Findings from a particular model, population, or experimental setting should not automatically be generalized beyond it.
This is one of the most useful ways to assess whether a manuscript is ready: imagine discussing it with a skeptical but reasonable reviewer.
- Can you explain why each major conclusion follows from the data?
- Can you acknowledge the limitations without undermining the central finding?
- Can you distinguish what the study demonstrates from what it suggests?
If so, the manuscript may be much closer to submission than it feels.
The entire manuscript should tell the same story
Scientific manuscripts often evolve considerably between the first experiment and the final draft. The problem is that not every section necessarily evolves with them. The introduction may still emphasize the question the investigators originally expected to answer. The abstract may contain language left over from an earlier version of the story. The discussion may give disproportionate attention to a result that seemed central six months earlier but is no longer the main finding.
Before submission, step back from sentence-level editing and look at the manuscript as a whole.
- Does the introduction lead logically to the research question that the study actually addresses?
- Do the Results present the evidence in an order that makes the scientific story easy to follow?
- Does the Discussion interpret those results rather than simply repeating them?
- Does the abstract accurately represent the final manuscript?
- Do the figures, text, and conclusions all support the same central message?
A manuscript can contain excellent science and perfectly grammatical sentences while still feeling unfinished because these larger elements are not aligned. At that point, another round of proofreading will not solve the problem. The manuscript needs structural attention.
Look at the manuscript the way a reviewer will
Authors read manuscripts differently from reviewers. By the time a paper is ready for submission, the authors may have spent months or years thinking about the project. They know why each experiment was performed. They know what every abbreviation means. They remember the controls that ruled out alternative explanations. Connections that are obvious to the research team may not actually appear on the page. A reviewer has none of that background. Before submission, try to create some distance and read the paper as if you were encountering the work for the first time.
The abstract and figures deserve particular attention. They provide reviewers with a rapid overview of the study and can strongly shape the initial understanding of the work. If the abstract, figures, and main text appear to tell slightly different stories, the reviewer has to do the work of reconciling them. Ideally, that work should already have been done by the authors.
Figures should be understandable, legends should explain the necessary abbreviations, symbols, and statistical information, and every figure and table should be discussed appropriately in the text. The abstract should be written or thoroughly revised after the rest of the manuscript is essentially complete so that it reflects what the paper actually says, rather than what the authors originally expected it to say.
Know the difference between improving the manuscript and postponing submission
Late-stage revision can become deceptively comfortable. Changing a sentence, moving a paragraph, or adjusting a figure feels productive. Sometimes it is. But after enough rounds of revision, these changes may stop materially improving the manuscript. This is where perfectionism can become an obstacle.
A manuscript rarely reaches a state in which no sentence could possibly be improved. If that is the standard for submission, the paper may remain on a computer indefinitely. Ask whether the changes you are making still affect the science or the reader’s understanding of it.
- Are you correcting an ambiguity that could cause a reviewer to misunderstand the experiment? Keep working.
- Are you restructuring a discussion that does not adequately address an important limitation? Keep working.
- Are you debating for the fourth time whether one perfectly acceptable phrase would be marginally better than another perfectly acceptable phrase?
You may be done. “Done” is often a decision rather than a feeling.
Perform the mundane final checks
Once the scientific story is complete, there is still a less exciting definition of “ready for submission.” The manuscript needs to be technically ready.
Before submitting, check the journal’s current author instructions rather than relying on memory or requirements from another journal. Confirm word limits, reference style, figure requirements, supplementary-material requirements, and any required statements or forms. Then perform a final manuscript-level check. Make sure that:
- every figure and table is cited in the text;
- figure legends define abbreviations, symbols, and statistical notation;
- acronyms are defined when first used;
- references are complete and consistent;
- tracked changes and comments have been removed;
- placeholder text is gone;
- the abstract agrees with the results and conclusions;
- statements about statistical significance are accurate;
- limitations are addressed appropriately; and
- all required manuscript components are present.
These details may not determine whether the science is publishable, but avoidable errors can make a manuscript look less careful than the underlying work deserves.
Sometimes you need someone who is not inside the project
One of the biggest obstacles to judging manuscript readiness is familiarity. After months or years immersed in a project, investigators can lose perspective in both directions. Small uncertainties can begin to seem much larger than they really are. At the same time, gaps in logic can become almost invisible because the authors automatically fill them in from their own knowledge.
This is why an outside reader can be valuable before submission. That reader might be a colleague who is familiar with the field but was not involved in the project. It might be a collaborator who has been less immersed in the writing. Or it might be an experienced scientific editor who can evaluate both the science and the way it is being communicated.
The purpose is not to lower the standard for publication or to reassure the authors that everything is fine. Sometimes an outside reader identifies a real problem that needs to be addressed. The value is perspective.
When I was running a laboratory, I sometimes needed that perspective myself. Talking through a manuscript with a colleague could make it apparent that the story was already complete. At other times, the conversation revealed exactly what was still missing. Both outcomes were useful.
So, is your manuscript ready?
A manuscript is probably ready for submission when you can answer yes to a few fundamental questions:
- Has the central scientific question been answered?
- Do the data support the major claims?
- Do the introduction, results, figures, discussion, and abstract tell a coherent and consistent story?
- Have you been transparent about the limitations of the work?
- Can the manuscript stand on its own without requiring the reader to know what the authors know?
- Would additional work materially strengthen the conclusions, or would it simply expand the scope of the paper?
- Are you still improving the manuscript, or are you simply reluctant to let it go?
Submission is not a declaration that the work is flawless. Peer review exists in part because manuscripts can still be questioned, challenged, and improved. The goal before submission is something more realistic and more useful: a manuscript that is coherent, transparent, scientifically defensible, and ready to withstand scrutiny. At that point, another experiment or another week of polishing may not make the paper stronger. It may simply delay the moment when the science finally reaches its readers.
By Dennis Hughes, MD, PhD
Managing Director, Clarity in Science Editing and Writing
Comments are closed