The Pre-Submission Checklist: 12 Things Reviewers Flag in the First 10 Minutes
A field-agnostic checklist of the 12 problems editors and reviewers spot on the first pass, from scope fit and missing statements to numbers that disagree between the abstract and the tables. Fix them before you submit.
Two people skim your manuscript before anyone reads it closely. The handling editor looks at the cover letter, abstract, and figures to decide whether the paper goes out for review at all. Then each reviewer makes a fast pass of their own (the abstract, the last paragraph of the introduction, the main table, the conclusions) before settling in for a careful read.
If your paper looks careless in those first ten minutes, the reviewer will read the next three hours with suspicion. If it looks clean, you get the benefit of the doubt.
First-pass problems are predictable, and you can fix most of them in an afternoon. The checklist below works for a cardiology trial, a qualitative sociology study, a systematic review, or a lab paper, and it's grouped by who catches each problem.
Part 1: What the Editor Checks at the Desk
The editor's question is narrow: does this belong in my journal, and is it complete enough to send out? Fail it and you get a desk rejection, sometimes within days, with a form letter that tells you nothing.
1. The paper fits the journal's scope, and the cover letter says why
Read the aims and scope page, then skim the journal's last two or three issues. If nothing recent looks like your paper, pick another journal. Editors reject for scope mismatch all the time, and a strong paper gets bounced as fast as a weak one.
Your cover letter should say in two or three sentences why this journal's readers need this result. Skip "we believe this work will be of interest to your readership." Write something like "Your journal published the two prior trials of this intervention, and ours is the first in an outpatient population."
2. The title and abstract claim what the design can support
Editors read the abstract for overreach. The classic version is causal language attached to an observational design: "X reduces Y" when the data show an association. Other versions are "novel" for a question others have studied, and "first" without a literature search to back it up.
Check each verb in your conclusion sentence. If the study can't show causation, the abstract shouldn't claim it. Editors won't penalize "was associated with" or "predicted," but they will remember an abstract that oversold.
3. You followed the author guidelines to the letter
Word limit, abstract structure, reference style, figure resolution and file type, line numbering, section order. An editor can spot a format problem at a glance. A manuscript still in another journal's reference style tells them you're recycling a rejection, which may be true, but you don't need to announce it.
4. You used the right reporting guideline and attached the checklist
Most designs now have a reporting standard, and many journals require the completed checklist at submission. The EQUATOR Network indexes them. The common ones:
| Design | Guideline |
|---|---|
| Randomized trial | CONSORT |
| Observational study (cohort, case-control, cross-sectional) | STROBE |
| Systematic review or meta-analysis | PRISMA |
| Qualitative research | SRQR or COREQ |
| Animal research | ARRIVE |
| Diagnostic accuracy | STARD |
| Prediction model | TRIPOD |
| Psychology and social science | APA JARS |
Fill the checklist in with real page and line numbers, not "see Methods." Reviewers use it as a map. If an item doesn't apply, say why.
5. Every required statement is present
Editorial offices send papers back for a missing statement before any reviewer sees them. Check for each of these your journal asks for:
- Ethics approval and consent, with the approving board named and the protocol number
- Registration, for trials (ClinicalTrials.gov, ISRCTN), systematic reviews (PROSPERO), or pre-registered studies (OSF), with the registration number and date
- Funding, including the funder's role in design, analysis, and the decision to publish
- Conflicts of interest for every author, not only the corresponding one
- Data and code availability, with a repository link or a specific access procedure. A growing number of journals refuse "available on reasonable request"
- Author contributions, in CRediT format if the journal uses it
- Use of generative AI, if the journal requires disclosure (many now do, following ICMJE guidance), stating what tool you used and for what
Part 2: What the Reviewer Notices on the First Read
A reviewer wants to know whether they can trust the work. On the first pass they look for signs of care or carelessness, and they form an opinion fast.
6. You state the research question in plain words
The last paragraph of your introduction should say, in one or two sentences, what you set out to find. Reviewers jump there first. If they have to infer the question from the Methods, they will assume you weren't sure either.
A good test: could a reader in an adjacent field write your hypothesis on a sticky note after reading that paragraph?
7. The Methods would let someone repeat the study
Reviewers skim Methods for the gaps that cause trouble later:
- Who or what was included and excluded, and why
- How the key variables were defined and measured
- How the sample size was chosen. Give a power calculation, a saturation rationale for qualitative work, or an honest statement that you used all available data
- Which analyses were planned in advance and which were exploratory
- How missing data were handled
- For observational work, which confounders you adjusted for and why those
We covered the most common methodology problems in a separate post. The short version: if a choice would surprise a reader, justify it in a sentence.
8. The numbers agree everywhere they appear
Reviewers catch this error all the time, and it does more damage than its size deserves. The abstract says 1,204 participants and Table 1 says 1,201. Percentages in a column add up to 103%. A hazard ratio in the text doesn't match the one in the figure. A p-value of 0.04 sits next to a confidence interval that crosses 1.
A mismatch like this rarely changes the conclusion, but it makes a reviewer wonder what else you got wrong. Do one pass where you read only the numbers: check each one in the abstract against the text, then against the tables and figures.
9. Results report effect sizes with their uncertainty
"Significantly improved (p < 0.05)" leaves a reviewer guessing at the size of the effect. Report the effect size, its confidence interval, and the exact p-value if you use one. If you ran many tests, say how you handled multiple comparisons. If you didn't correct, say so and label those results as exploratory.
10. Figures and tables make sense on their own
Many reviewers look at the figures before reading the text. Each figure and table should be readable in isolation:
- Axes labeled, with units
- Every abbreviation defined in the legend
- Sample size shown for each group or panel
- Error bars defined (SD, SE, or 95% CI are not interchangeable)
- Color choices that survive grayscale printing and color-blind readers
If a reviewer has to hunt through the Methods to work out what a figure shows, the figure isn't finished.
11. The limitations are specific and the conclusions stay inside the data
A limitations paragraph that says "this study has some limitations, including sample size" reads as a formality. Name the actual threats to your conclusion, say which direction they would bias the result, and say what you did about them.
Then reread the final paragraph of the Discussion. Co-authors edit conclusions over many drafts, and the claims creep: a result from one population becomes a recommendation for all of them, or "may suggest" turns into "demonstrates." Reviewers compare the last sentence of the paper with the main result in your primary table, so line them up before you submit.
12. Every reference is real, current, and relevant
Three checks:
- Existence. Fabricated citations are on the rise, mostly from AI tools inventing plausible references, and reviewers have started looking for them. One invented reference can sink an otherwise sound paper. Verify each entry resolves to a real paper with the authors, year, and journal you listed. Our free citation checker does this in bulk, and we explain the problem in more detail here.
- Retractions. Citing a retracted paper as support for a claim is an easy catch for a reviewer. The Retraction Watch database is now openly available through Crossref, and many reference managers will flag retracted papers for you.
- Currency. If the newest reference on your central question is five years old, a reviewer in the field will notice the gap and may know the paper you missed. Run one last search before submitting.
The Final 15 Minutes Before You Click Submit
The last few items concern the files you upload.
- Blinding. For double-blind review, remove author names from the manuscript, the file properties, the acknowledgments, and self-citations written as "our previous work."
- Tracked changes and comments. Accept all changes and delete all comments. Then open the file on a different computer to confirm they're gone.
- No hidden text. Some authors have tried embedding invisible instructions for AI reviewers. Journals now scan for it, and it backfires.
- Supplementary files. Every supplement is cited in the main text, and every citation points to a supplement that exists.
- Co-author sign-off. Every author has seen the final version. Journals can treat submission without it as misconduct, and you don't want that email from an editor.
The Checklist
Copy this into your submission notes.
Desk check
- Paper fits the journal's scope, and the cover letter says why
- Title and abstract claims match the study design
- Author guidelines followed (length, format, references, figures)
- Reporting guideline checklist completed with page numbers
- Ethics, registration, funding, COI, data availability, contributions, and AI-use statements present
Reviewer first read
- Research question stated in plain words at the end of the introduction
- Methods detailed enough to repeat, with sample size justified
- Numbers consistent across abstract, text, tables, and figures
- Effect sizes and confidence intervals reported, multiple comparisons handled
- Figures and tables readable on their own
- Limitations specific, conclusions within the data
- References verified as real, not retracted, and current
Files
- Blinded properly, if required
- Tracked changes and comments removed
- No hidden text
- Supplements all cited and attached
- All co-authors approved the final version
Getting a Second Set of Eyes
Items 8, 11, and 12 are the hardest to check on your own paper. You've read your numbers so many times you no longer see them, and you know what you meant to conclude, so you read that instead of what's on the page.
A colleague outside your group is the best fix. If you can't find one before the deadline, ManuscriptMind runs a structured review that checks numerical consistency across sections, flags claims your design doesn't support, and verifies each reference. You get a severity-ranked list of issues to fix before a reviewer finds them.
And when the reviews do come back, here is how to respond without tanking the resubmission.
Keep reading
How to Respond to Reviewer Comments Without Tanking Your Resubmission
A point-by-point template for revise-and-resubmit letters, including how to agree, push back, and decline reviewer requests without losing the editor.
Read5 Common Methodology Issues That Get Manuscripts Rejected
Learn the most frequent methodology problems peer reviewers identify and how to avoid them in your research manuscripts.
Read