How to review
A useful report tells the authors what must change, what is optional, and why.
Read the whole package
Start with the abstract and the figures. Then the methods. Then the claims in the discussion. A beautiful narrative that the methods cannot support is a fault, not a style.
Structure the report
A short summary of what you understood the paper to be. Then numbered essential points. Then optional points, labelled optional. End with the recommendation and a sentence that matches it — do not recommend rejection in the form and praise in the letter.
Essential versus optional
Essential: a control that is missing, a statistic that does not match the design, a claim the data do not carry, an ethics statement that is absent where the work required one.
Optional: another experiment that would make a different paper, a preference for a figure style, extra citations to your own work unless the claim actually depends on them.
Statistics and figures
If you cannot assess the analysis, say so — that is a legitimate ‘unable to assess’, not a pass. Figures should have labelled axes, units, sample sizes and a caption that can be read without the main text. Do not demand decorative scientific imagery.
Using the workspace
The reviewer workspace lists the anonymised metadata, the abstract, and the files the authors deposited. Files that have not been deposited are marked as such — the journal does not invent a PDF. Submit comments and a recommendation there. Until the editorial system is connected, a demonstration session shows the layout; reports are not filed.