Preparing Supplementary Materials

A well-run paper is disciplined about what it shows. It presents exactly what the reader needs to follow the argument, and no more. Everything else that supports the work but would clog the narrative — extended methods, extra figures, large tables, raw outputs — goes into supplementary materials. Used well, the supplement keeps your paper readable while making your work more complete. Used badly, it becomes a place to hide things reviewers should have seen.

What supplements are for

Supplementary materials are the right home for content that is useful but not essential to the main story:

• Extended methods or protocols that would interrupt the flow.
• Additional figures and large data tables.
• Videos, audio, or interactive content that print cannot carry.
• Full survey instruments, checklists, or code.
• Sensitivity analyses and extra subgroup results that support, but do not change, the main findings.
• Full search strategies for a systematic review.

What must stay in the paper

Here is the line that trips authors up: anything a reader needs to understand or trust your conclusions belongs in the main text. Your key results, the analyses that support your claims, and the figures that carry your argument cannot be demoted to the supplement to hit a word limit. Reviewers may not read supplements closely, and readers certainly skim them — so burying a key result there weakens your paper and invites the suspicion that you were hiding it.

Test: if a reader read only the main paper, could they judge whether your conclusions are justified? If not, something has been misfiled in the supplement.

Main text or supplement: a decision table

The table below sorts common items. Where an item moves to the supplement, the main text should still say in a sentence what it showed.

For systematic reviews, PRISMA 2020 asks authors to present the full search strategies for all databases, registers and websites, including any filters and limits. A supplement is the usual place for them, because they are long.

Worked example (invented to illustrate the structure): a retrospective cohort paper on wound infection after surgery keeps its primary comparison, its adjusted model and its patient flow in the main text. The supplement holds the data collection form, the complete variable definitions, a table of results with missing data imputed, and the code. The Results section adds one sentence saying that the imputed analysis gave the same conclusion.

Format and cite them properly

Give supplementary items a clear numbering scheme (Supplementary Figure S1, Supplementary Table S1) and refer to each one in the main text, in order. Provide them in open, accessible formats — PDF for documents, standard formats for data — not proprietary files a reader may not be able to open. Each supplementary figure and table still needs a caption that stands alone.

Many journals publish supplementary files as supplied, without copy-editing or typesetting. Check them as carefully as the main text, because any error in them will be published as it stands.

• Keep numbering in step. When a revision removes or reorders a supplementary item, update every S-number and every in-text reference to it.
• Use open data formats. Tabular data travels better as comma-separated values (CSV) than as a spreadsheet with formulas, colours and hidden sheets.
• Explain the variables. A data file needs a short data dictionary giving each variable's name, meaning, units and codes.
• Document code. Include a README that states the software, its version and the order in which to run the scripts.

Check supplements for identifiers and author details

Supplements leak information more easily than the main text, because fewer people read them. Before you upload, check each file for patient identifiers and, where review is double-blind, for author identity.

• Remove names, hospital numbers, full dates of birth and exact admission dates from data files, including hidden columns, hidden sheets and comments.
• Remove author names from file properties and tracked changes.
• Replace institution names in protocols and consent forms with a neutral placeholder such as [institution] for the review version.

The full routine for blinding every file is in how to anonymise a manuscript for double-blind review.

Make large data findable, not loose

A big dataset attached as a spreadsheet is hard to find and easy to lose. For anything substantial, deposit it in a repository following the FAIR principles, cite it with its DOI, and let the supplement point to it. That way your data is preserved, citable and reusable — which a file bolted onto a PDF never is.

To find a repository suited to your discipline, search the registry of research data repositories at re3data. Where no discipline repository fits, a general repository that issues DOIs will do. Then state where the data sit, and any access conditions, in your data availability statement.

Supplementary materials checklist

• A reader of the main paper alone can judge whether the conclusions are justified.
• Every supplementary item is cited in the main text, in order, with matching S-numbers.
• Files are in open formats, and data files come with a data dictionary.
• Code comes with a README stating software and versions.
• No file contains patient identifiers or, for double-blind review, author details.
• Large datasets are in a repository and cited with a DOI.

In a systematic review the full search strategies go in the supplement, but the record counts they produce stay in the main text; how to draw a PRISMA flow diagram shows how to fill each box of the PRISMA 2020 template.

Updated 17 September 2026: We expanded this guide with a main-text-or-supplement decision table, an invented worked example, a file-preparation list and a checklist.