01
Using the wrong reporting guideline for the study design
Which guideline applies depends on the design, not the topic. A randomized trial uses
CONSORT, whether the field is cardiology, psychiatry, or surgery. An observational cohort uses
STROBE. A systematic review uses
PRISMA 2020. A diagnostic accuracy study uses STARD. An animal study uses ARRIVE 2.0. Submitting a meta-analysis with a CONSORT checklist or a cohort study with PRISMA signals to the editor that the authors do not understand their own design — exactly the kind of structural concern that triggers a rejection without further review.
The fixMatch the guideline to the design before you start writing — not after. The
EQUATOR Network maintains a complete decision tree, and our
reporting guidelines hub maps every common study type to its checklist. If your design uses an extension — cluster trials, network meta-analyses, scoping reviews — load the extension alongside the parent guideline.
02
Citing the guideline but not actually addressing each item
Pasting "Reported in accordance with CONSORT 2025" into the methods section and attaching a completed checklist at submission is not the same as compliance. Reviewers and editors increasingly cross-check the checklist against the manuscript text — and when an item is marked "page 4" but the page does not contain the required content, the manuscript looks worse than one that never claimed compliance at all. The most common failures here are randomization detail, blinding mechanism, allocation concealment described as a separate item from sequence generation, and per-item participant accounting in the flow diagram.
The fixTreat the checklist as a writing tool, not a submission formality. Go through each item, locate the specific text in your manuscript that satisfies it, and mark items complete only when the corresponding content is genuinely there. If you cannot find the sentence, the item is not addressed — write it, then mark the checklist.
03
Missing required methods items — sample size, randomization, blinding
This is where most trial and observational manuscripts actually fail. The chronic offenders: sample-size justification (the calculation with anticipated effect size, variance, alpha, and power — not just the final N), randomization sequence generation and allocation concealment treated as two distinct items, blinding (who was blinded, how, and the implications for outcome assessment if blinding was not possible), and confounding control for observational designs. Vague phrasing — "patients were randomized," "the trial was double-blind," "adjusted for confounders" — does not survive a careful reviewer, and editors with statistical training spot these omissions in minutes.
The fixWrite the methods section assuming a hostile statistician will read every sentence. Specify how the random allocation sequence was generated, how it was concealed until assignment, who was blinded after assignment and how, how missing data were handled (with the specific method named), and how confounders were measured and controlled. "Computer-generated random sequence with central web-based allocation" beats "randomized by computer" every time.
04
Incomplete flow diagrams and participant accounting
CONSORT and PRISMA both require a flow diagram, and the numbers across the diagram, the methods, the baseline characteristics table, and the results section must reconcile. They frequently do not. Authors lose track of screened vs. eligible vs. randomized vs. analyzed; they drop dropouts without explanation; they present per-protocol counts as if they were intention-to-treat. A manuscript where 450 participants are randomized but only 340 are analyzed — with no reason given for the missing 110 — reads as either careless or evasive, and it is one of the most common reasons reviewers recommend major revision or outright rejection.
The fixBuild the flow diagram before drafting Results, not after. Account for every participant from screening through final analysis, with reasons for each exclusion and dropout. Confirm the totals in the diagram match the totals in the baseline table and the analysis. If you are reporting intention-to-treat, every randomized participant should appear in the primary analysis — and if not, the missing-data method should be explicit.
05
Submitting without a systematic pre-submission compliance check
The first four mistakes are catchable. The fifth is the meta-mistake: not running a systematic compliance check at all. Most authors submit having only skimmed the abstract-level checklist — and they find out which items they actually missed when the rejection arrives weeks later, with no opportunity to fix anything. A formal pre-submission pass — manuscript text on one side, the relevant
reporting checklist on the other, item by item — eliminates the majority of these gaps.
The fixRun a structured check before you submit. You can do this manually with the checklist PDF and a careful read of your own manuscript, or you can have
PeerReviewAI do it automatically — every review auto-detects your study type, applies the correct guideline, and evaluates each checklist item against your manuscript text, with specific notes on what is missing and how to address it.