Knowing when to use tables vs figures research papers is less about aesthetics and more about reader cognition. Tables carry exact values for readers who need to compare numbers. Figures carry pattern for readers who need to see relationships. Picking the wrong format forces reviewers to do work the author should have done — and that friction shows up in the review.
This guide gives you a clear decision framework, the journal conventions that matter, and the mistakes that most often trigger reviewer comments.
The Fundamental Difference
A table is for precision. Use it when readers need to read exact values, compare specific cells, or verify numbers against a claim made in the text. A figure is for pattern. Use it when readers need to see a trend, a distribution, a relationship, or a process.
The simplest heuristic: if readers will quote a specific number from this display, use a table. If readers will describe a shape or trend, use a figure.
Use a table
Descriptive statistics, regression coefficients, demographic breakdowns, correlation matrices — anywhere precise cell values matter.
Use a figure
Time series, group comparisons, distributions, theoretical models, process flows — anywhere shape communicates faster than numbers.
When Tables Work Better
Tables do specific jobs better than figures almost every time. Reach for a table when:
Best for exact comparison
- Descriptive statistics (means, SDs, ranges)
- Regression coefficients with p-values
- Demographic breakdowns
- Correlation and factor loading matrices
- Sample characteristics across groups
- Content analysis coding frequencies
Best for showing pattern
- Trends over time
- Distributions and outliers
- Group mean comparisons
- Theoretical or conceptual models
- Path diagrams and SEM models
- PRISMA flow diagrams
A common mistake is using a bar chart where a simple two-row table would read faster. If your "figure" has fewer than six data points and no meaningful pattern to see, it is probably a table in disguise.
The Reviewer-Triggering Mistakes
2. Describing every number from the table in the text narrative.
3. Using a figure where a table would be more informative (and vice versa).
4. Tables with too many decimal places or unnecessary columns.
5. Figures with unreadable axis labels, missing units, or no legend.
Duplicating data in both formats wastes space and signals editorial imprecision. Pick one — the one that best serves the reader's task for that specific data point.
How Many Tables and Figures Is Normal?
Most Scopus and WoS journals accept somewhere between 4 and 8 displays total for an empirical paper, with the split depending on the study design. Quantitative papers typically weight toward tables; qualitative and review papers typically weight toward figures.
Check your target journal's recent issues for the norm. If three recent papers use 6 displays each (4 tables, 2 figures), that is your working budget. Submitting 12 displays where the journal publishes 6 is a flag.
The Rules Reviewers Apply (Unwritten but Universal)
Tables
- Three-line format (top, middle header line, bottom) is the academic standard. Avoid heavy gridlines.
- Numbers align right; text aligns left.
- Decimals consistent within a column — usually two places unless precision matters.
- Statistical significance marked with asterisks and explained in a footnote.
- Sample sizes (N) always declared.
Figures
- Resolution of at least 300 DPI for publication. Most journals now require vector format for line art.
- Colour used purposefully — and always legible in greyscale (many readers still print).
- Axis labels include units and font size is readable at print scale.
- Legend on the figure, not only in the caption.
- Error bars defined in the caption (SE, SD, or CI).
For guidance on building publication-quality displays, see How to Create Publication-Quality Figures and Tables.
The Text-Table-Figure Triangle
Strong papers treat text, tables, and figures as three channels carrying different content — never repeating each other. Text narrates and interprets. Tables hold precise numbers the reader may need to verify. Figures show patterns the reader needs to see at a glance.
When you write a Results paragraph, refer to the table or figure by number ("As shown in Table 2…") and describe only what the display cannot — the interpretation, the comparison, the headline finding. Do not re-narrate the table cell by cell. For more on this discipline, see our pillar guide on research paper structure for journal submission.
Journal-Specific Conventions to Check
Tables and figures are among the most template-sensitive elements of a manuscript. Before finalising:
- Check whether the journal requires tables and figures embedded in the text or in separate files.
- Verify the caption format — some journals put titles above tables and below figures; others differ.
- Confirm colour policy — a handful of journals still charge for colour figures.
- Review supplementary material policy — large tables can often be moved online.
The Bottom Line
When to use tables vs figures research papers comes down to a single question: what does the reader need to do with this data? If they need to read specific numbers, give them a table. If they need to see a pattern, give them a figure. Never both. Match your display count to your target journal's norm, make every display interpretable on its own, and keep the text narrative focused on interpretation rather than data re-description. For the full structural picture, see our pillar guide on how to structure a research paper for Scopus/WoS journals.
Need a Second Eye on Your Displays?
Our editorial team can review your tables and figures against journal-specific norms before you submit — part of our preparation support workflow.
Free Consultation