Skip to main content
Paper Buddy

Manuscript workflow

What to check before sharing your manuscript draft

Before sending a draft to your supervisor or co-authors, compare the research question, numbers, citations and conclusions. Prepare your feedback request with the seven checks and email example below. Once the draft is assembled or submission approaches, use Section Critique and Health Check to review logic, reporting and statistical descriptions.

What you will prepareIdentify sections needing focused revision, issues spanning the manuscript, revision priorities and facts the authors must confirm.

Seven checks before sending a draft to your supervisor or co-authors

You do not need a perfect manuscript before asking for feedback. Check the following so readers can focus on the research, and include your unresolved questions. This is an editorial checklist for sharing a draft, not an official reporting guideline or submission approval standard.

  1. Can you state the research question in one sentence?

    Check that the population, comparison and outcome have the same scope across the title, objective, Methods, Results and conclusion.

  2. Do numbers agree across the abstract, text and tables?

    Compare sample sizes, denominators, units of analysis, measurement times, units and estimates. Resolve discrepancies against actual analysis records rather than choosing a number by how often it appears.

  3. Are terms and abbreviations consistent?

    Avoid different names for the same population or outcome. Expand abbreviations at first use, and follow journal instructions for independently understandable abstracts and tables.

  4. Do tables and figures contain the information needed to read them?

    Check titles, legends, units, denominators, abbreviations and analysis conditions, and verify that the text refers to the correct table or figure.

  5. Does each citation support its associated claim?

    Check the source population, design and results as well as the paper’s existence. After AI editing, recheck whether the scope of the claim has changed.

  6. Does the conclusion stay within what the study established?

    Look for associations presented as causal effects, unmeasured benefits and limitations absent from the final interpretation.

  7. Is the feedback request specific?

    Identify the file version and date, unfinished sections, two or three priority questions and a preferred reply date. Check whether the timing works for the recipient, and share only material you are authorized to disclose.

When numbers differ, first check that they describe the same outcome

Fictional teaching example: for the same group’s attendance within 30 days, the abstract says 80%, but Results and the table say 82/100 (82%).

On small screens, scroll the table horizontally to compare every column.

Comparing numbers across the abstract, text and table
LocationValue in the draftWhat to verify
Abstract80%Are the analysis population, denominator and period the same?
Results82/100 (82%)Does this match the final analysis output and manuscript version?
Table 282/100 (82%)Do the title, footnote and denominator describe the same outcome?

If the verified final analysis in this fictional case is 82/100, revise the abstract to 82%. If the values instead describe different groups or analysis populations, correct the definitions and labels first. Do not change a value solely because AI flagged it or because another value appears more often.

What should you fix first?

  1. First: numbers, denominators or method inconsistencies affecting interpretation, and claims stronger than the evidence.
  2. Next: missing explanations, table and figure links, terminology and citation consistency.
  3. Last: prose and formatting. Keep unresolved scientific questions visible in the feedback request.

A review-request email to adapt

Replace the brackets with actual information. Do not describe unfinished checks as completed.

Subject: Draft for review — [study name, version and date]

Dear [name],
I am sharing a draft addressing [research question].
Checks completed: [items personally verified and revisions made].
Still unresolved: [unfinished sections or facts to verify].
I would especially appreciate your feedback on (1) [question and manuscript location] and (2) [question and location].
If possible, could you review it by [preferred date]? Please let me know if another timeline would work better.
Attached: [manuscript version], [tables and figures].
Thank you.

Source for manuscript consistency, tables and citation checks: ICMJE manuscript-preparation recommendations, checked November 8, 2026. The seven-part grouping, example and email are educational material created by Paper Buddy. ICMJE ↗

See how to revise repetition and overinterpretation in a Discussion →

Before you start

  • The manuscript sections to review, such as Methods, Results and Discussion, and the study type. Distinguish sections not yet written.
  • Materials that establish the analysis methods, sample sizes, denominators, units of analysis and actual results.
  • The venue’s current instructions and the reporting guideline applicable to the study. Authors must inspect the evidence needed for final decisions.

Steps in Paper Buddy

  1. Review one section’s logic and sequence

    If you are stuck on a particular section, such as the rationale in Introduction or interpretation in Discussion, start with Section Critique. Examine unclear connections and missing explanations, then verify that each comment applies to the actual text.

  2. Review the whole manuscript with Health Check

    After assembling a draft or completing a major revision, use Health Check to examine reporting items, consistency between study design and statistical descriptions, and the connection between claims and evidence. It reviews what the manuscript states: distinguish missing descriptions from a conclusion that something was never done.

  3. Verify findings and prioritize revisions

    Start with issues affecting interpretation, such as inconsistent sample sizes, denominators or units of analysis, mismatches between methods and results, and conclusions stronger than the evidence. Ask Paper Chat to explain unclear findings in the manuscript’s context; consult co-authors or a statistician when needed.

  4. Inspect the revised file and submission items

    Compare the abstract, main text, tables and figures after revision. Separately check references, submission format, author names and affiliations, ethics and competing-interest declarations against actual records and venue requirements. Open and read the final exported Word file too.

A request to adapt

Example: after a major revision, you want to recheck the connection between Methods, Results and the conclusion.

Check whether the unit of analysis in Methods matches the sample sizes and denominators in Results, and whether the conclusion is stronger than the actual findings. Distinguish what the manuscript establishes from facts the authors still need to verify.

This is an example request. Adapt it to your manuscript and verified facts.

What to check yourself

  • Do the research question, design, analysis, results and conclusions address the same scope?
  • Are sample sizes, denominators, units and numbers consistent across text, tables and figures?
  • Are required reporting items described according to what was actually done?
  • Have you avoided causal claims or generalizations stronger than the observations?
  • Have you checked references, author information, declarations, submission format and the exported file?

Frequently asked questions

When should I use Section Critique versus Health Check?

Use Section Critique for a focused review of the logic and structure of one section you are writing. Use Health Check after assembling the draft to examine reporting, statistical descriptions and claims and evidence across the manuscript.

Does Health Check decide statistical validity or publication likelihood?

No. It reviews manuscript descriptions and does not replace statistical reanalysis of raw data or predict acceptance. Verify findings against actual records; researchers judge the validity of the analyses and final interpretation.