How to Create Tables for Research Publications

Tables present precise numbers that readers may want to compare, such as baseline characteristics, outcomes and model results. Each table should be understandable without reading the text, with a clear title, labelled rows and columns, units and footnotes.

When to use a table

• You have many numbers that readers may need exactly.
• You are comparing groups on several characteristics.
• A figure would hide detail.

Use text for one or two numbers, and figures for patterns and trends.

Layout rules

1. Title says what, who and when (for example "Table 1. Baseline characteristics of participants by treatment group").
2. Columns: groups and total, with n in the header.
3. Rows: variables, with units and categories indented.
4. Summary statistics: mean (SD) or median (IQR) and n (%), stating which in the row label or footnote.
5. Decimal places: consistent and sensible.
6. Footnotes: abbreviations, statistical tests, symbols.
7. No repeated data between table and text, other than key results.

Table 1 in trials and cohorts

Baseline characteristics describe participants, and in randomised trials guidelines such as CONSORT advise against significance tests of baseline differences, because any difference arises by chance; describe them instead.

Illustrative example

Invented for teaching. A Table 1 lists age (median, IQR), sex (n, %), comorbidity count and frailty score for early-surgery and delayed-surgery groups, with n in the column heads and the footnote "IQR, interquartile range; percentages are column percentages."

Choosing between a table and a supplement

Journals limit the number of tables in the main text. Keep tables that readers need to judge the main result (baseline characteristics, primary outcome, key models) and place long lists, extra models and raw extraction data in a supplement. Refer to each supplementary table in the text so that readers know it exists, and give it a self-contained title and footnotes like a main table.

Common pitfalls

• Too many tables for the journal's limit.
• Unlabelled units and abbreviations.
• Mixing mean (SD) and median (IQR) without saying.

Key takeaways

• Clear title, labelled units, consistent summaries and footnotes.
• Follow the journal's table style.

This page is general educational information about research methods. It is not medical, legal or regulatory advice; follow your institution's ethics board and your target journal's instructions.