The pre-submission checklist that fits on an index card

Editorial desks reject a substantial share of submissions before peer review, and the reasons cluster hard. Out of scope. Over the word limit. Reference list in the wrong style. Figures at screen resolution. A missing declaration. None of these are judgments about the work; they are judgments about whether the work has been prepared, and they cost weeks.

What follows is the checklist our client desk runs before a manuscript goes back to an author for submission. It is deliberately short enough to keep beside you.

Before anything else: scope

Run this first, because everything below is wasted if it fails.

• Read the journal's aims and scope, and then read the titles in the two most recent issues. If nothing there resembles your paper, the scope statement is being optimistic on the journal's behalf.
• Confirm the article type exists — original research, brief report, review, methods — and that yours matches the one you selected.
• Check the word limit for that article type, and whether it includes the abstract, references, and figure legends. Journals differ, and the difference is often two thousand words.
• Check any structural requirements you cannot retrofit later: a structured abstract with named headings, a required reporting checklist, a data availability statement.

The manuscript file

• Title within the character limit, and free of abbreviations the journal disallows in titles.
• Abstract within its own word count, structured if required, and containing no citations or undefined abbreviations.
• Keywords: the number the journal asks for, drawn from a controlled vocabulary where one is specified.
• Headings match the required structure, at the required levels, with consistent capitalisation.
• Line numbers and double spacing if requested — a small thing that reviewers notice immediately when it is missing.
• Every abbreviation defined at first use in the abstract and again at first use in the body.
• Units consistent, with the spacing convention the style requires between number and unit.
• Anonymised if the journal is double-blind: no author names in the file, no identifying acknowledgements, no institution in the methods, and no author metadata left in the document properties. That last one is the trap.

References

The single most common source of mechanical rejection, and the most tedious to fix late.

• Every in-text citation appears in the list; every entry in the list is cited in the text. Both directions.
• The style matches the journal's, including the author cutoff before et al., journal-title abbreviation, and page-range convention.
• Numeric citations are in order of first appearance and renumber correctly after every edit.
• DOIs resolve. Check them — a DOI that points at the wrong paper reads as carelessness at best.
• Preprints are marked as preprints, and replaced by the version of record where one now exists.
• No duplicate entries for the same source under different metadata.
• Reference manager field codes removed or the citations unlinked, on a copy, so the bibliography does not re-render on someone else's machine.

Tables and figures

• Every table and figure is called out in the text, in order, and the numbering has no gaps.
• Figure resolution meets the journal's minimum — commonly 300 dpi for photographs and higher for line art. Screenshots pasted from a slide deck will not pass.
• File formats are the ones requested, not the ones you have.
• Legends stand alone: a reader should understand the figure without the body text, including every abbreviation used inside it.
• Axis labels carry units. Error bars are defined in the legend — standard deviation, standard error, or confidence interval, stated explicitly.
• Colour choices survive greyscale printing and are distinguishable to readers with colour vision deficiency.
• Any reused figure has documented permission, and the permission line is in the legend.

Declarations and the paperwork that holds submissions

This is the section authors leave until the upload form demands it, which is exactly when it causes a delay.

• Author list final, in agreed order, with correct affiliations and a designated corresponding author.
• ORCID iDs where required — increasingly for all authors, not just the corresponding one.
• Contributorship statement, if the journal uses a taxonomy such as CRediT.
• Funding statement with grant numbers as the funder wants them written.
• Competing interests declared for every author, including the ones who believe they have none.
• Ethics approval with the committee name and reference number; consent statements where applicable.
• Data availability statement that says something specific, with a repository and accession number if the data are deposited.
• AI use declared if any tool was used in producing the manuscript.
• Language editing acknowledged, with a certificate if the journal asks for one.
• Reporting checklist completed and attached — CONSORT, PRISMA, STROBE, ARRIVE, whichever applies.
• Trial or review registration number included where one exists.

The cover letter

• Addressed to the current editor, by name, spelled correctly.
• Names the journal — the correct journal. Cover letters naming a previous submission target are more common than anyone would like.
• States in two or three sentences what the paper shows and why this journal's readership specifically.
• Confirms the work is original, not under consideration elsewhere, and approved by all authors.
• Notes any prior submission history or preprint posting, and any suggested or excluded reviewers with reasons.

The last pass

Three things worth doing after everything above is done, in this order:

• Read the paper on paper, or at least in a different application. You will see things you have stopped seeing on screen.
• Read the reference list on its own, top to bottom, without the manuscript. Errors cluster at the beginning and end.
• Open the submitted files as a stranger. Download what you uploaded, open every file, and check they are the versions you meant. Submitting the wrong draft is not rare.

Why the order matters

Scope first, because it can end the exercise. Then the manuscript, then references, then figures, then declarations — because each stage can change the one after it. Reformatting references before you know the target journal is work done twice, and it is the single most common wasted effort we see.

Most of this list is mechanical, and mechanical work is what our macros and pipeline exist to absorb: reference reconciliation, callout verification, style enforcement, abbreviation tables. What no program will do for you is decide whether the paper belongs in that journal, or whether the discussion is claiming more than the results support. Run the checklist so an editor's attention goes to those questions instead.