Tables and Figures: When to Use Each

Tables and figures carry a disproportionate share of the attention in a results section. Reviewers often read the tables and figures first, scanning them before any prose, and journal editors sometimes judge a manuscript's quality by how its visuals are built. Weak tables or figures signal weak thinking, and strong ones can carry a paper even when the prose around them is uneven. Getting the choice right between a table, a figure, and plain prose is one of the highest-leverage decisions in writing up quantitative or mixed-methods research.

This guide covers when to use a table, when to use a figure, when to use neither, and how to build each so reviewers trust them. For the broader question of how to structure the results section itself, see the guide to writing a results section. For statistical notation inside tables and prose, see the guide to reporting statistical results in APA format.

The Core Decision

Every result in a paper can be presented as prose, as a table, or as a figure. The choice depends on what the reader needs to do with the information.

• Prose works when the finding is a single value or a short comparison the reader will absorb once and move on.
• Tables work when the reader will compare precise values across rows or columns, or when exact numbers matter.
• Figures work when the reader needs to see a pattern, a trend, a distribution, or a relationship that would be hard to see in a table of numbers.

That hierarchy applies in roughly that order of frequency in most quantitative papers. The majority of results are reported in prose. A smaller subset earn a table. A smaller subset still earn a figure. Over-illustrating is a far more common problem than under-illustrating.

When to Use a Table

Tables earn their place when the reader will do one of three things: compare precise values across groups or time points, look up specific numbers, or evaluate a set of related statistics the paper references repeatedly.

Clear cases for a table:

• Descriptive statistics for multiple groups or variables. Means, standard deviations, and sample sizes for three groups across six measures belong in a table. Spelling them out in prose produces a paragraph the reader will skip.
• A correlation matrix. Any time you report more than three or four correlations, a matrix is clearer than prose.
• Regression results with multiple predictors. Beyond two or three predictors, a table showing coefficients, standard errors, confidence intervals, and significance is standard practice.
• Measurement properties. Reliability coefficients, factor loadings, and item statistics across multiple scales fit naturally in a table.
• Participant characteristics. A demographics table lets reviewers evaluate the sample without wading through a paragraph of frequencies.

The test for whether a table is justified: would a reader want to compare any two numbers in it? If yes, use a table. If the numbers are a list rather than a comparison, prose is usually enough.

When to Use a Figure

Figures earn their place when the finding is a pattern the eye can read faster than the reader can parse numbers. The question isn't whether the data could be shown as a figure. Most data could. The question is whether the figure communicates something a table cannot.

Clear cases for a figure:

• Trends over time. Line graphs showing how an outcome changes across multiple time points reveal patterns a table buries.
• Interactions. A two-way ANOVA interaction is almost always clearer as a figure (with lines or bars for each condition) than as a table of means.
• Distributions. Histograms, box plots, and density curves show shape, spread, and outliers in ways numbers cannot.
• Relationships between continuous variables. Scatter plots show whether a relationship is linear, curvilinear, clustered, or driven by outliers.
• Study flow. CONSORT diagrams for clinical trials and participant flow diagrams for longitudinal studies communicate the structure of the study more efficiently than prose.
• Qualitative themes and their relationships. Thematic maps or network diagrams show how codes and themes connect in qualitative research.

For specific guidance on choosing between chart types (bar charts, scatter plots, box plots, histograms), see the guide to choosing the right visualization for your data.

When Neither a Table nor a Figure Is Needed

Most individual results do not need a visual. Prose handles them more efficiently and keeps the paper focused.

Prose is better when:

• The finding is a single statistical test with a short result ("Participants in the treatment group scored higher than controls, t(58) = 2.41, p = .019, d = 0.63").
• A comparison involves only two values that the reader will not revisit.
• The finding is secondary and does not warrant visual real estate.
• The visual would simply restate what one sentence can say.

A figure or table that duplicates one sentence of prose is noise, not communication. Reviewers often flag unnecessary visuals as padding, which hurts credibility more than the visual helps.

The Three-Part Decision Framework

Three questions in that order resolve most decisions. If a result fits the first question, make it a table. If it fits the second, make it a figure. Otherwise, prose.

Building a Table Reviewers Trust

A good table is not just a grid of numbers. It has a structure that reviewers scan quickly and specific conventions that mark it as professional.

The elements every table needs:

• A number. Tables are numbered in the order they are first mentioned in the text (Table 1, Table 2, Table 3). Numbering restarts at the beginning of each chapter in a dissertation but not in a journal article.
• A title. The title appears below the table number in APA style and is italicized. It should describe what the table shows in a complete noun phrase, not a sentence.
• Clear column headings. Column headings describe what is in the column. Group variables logically from left to right. Use descriptive headings rather than abbreviations when space allows.
• Row labels. Row labels on the left describe what each row represents. Use consistent capitalization and align text to the left while aligning numbers on the decimal point.
• A note. General notes appear below the table and define abbreviations, explain unusual symbols, state the significance convention (such as asterisks for p-values), and credit the data source if applicable.
• Appropriate rules. APA style uses horizontal rules above the header row, below the header row, and below the last row of data. No vertical rules and no interior horizontal rules.

A table that follows these conventions looks the way reviewers expect, which lets them focus on the content. A table that fights conventions distracts even when the numbers are correct.

Building a Figure Reviewers Trust

Figures carry the same requirements as tables and add a few more because they involve visual choices.

The elements every figure needs:

• A number. Figures are numbered independently of tables (Figure 1, Figure 2, Figure 3) in the order they are mentioned.
• A caption. The caption appears below the figure in most styles (APA 7 places it below the figure number and title). It describes what the figure shows and defines any abbreviations.
• Axis labels. Both axes need labels with the variable name and, where appropriate, the unit of measurement. Unlabeled axes are the most common reviewer complaint about figures.
• Legible scale. Tick marks should produce round numbers. Scales should start at zero unless there is a reason to start elsewhere, and if the scale is truncated, say so.
• A legend when needed. If the figure plots multiple groups or variables, a legend identifies each one. Place the legend inside the plot area when space allows.
• Sufficient resolution. Figures need to render well in print and online. For journal submission, 300 dpi is the usual minimum for raster images; vector formats (PDF, SVG, EPS) are preferred when the journal accepts them.
• Color choices that work in grayscale. Some readers will print your paper in black and white. Patterns, line styles, or shapes should distinguish series even when color is lost.

A figure that violates one of these rules is often rejected at copy editing, which adds a round to the submission timeline.

Referencing Tables and Figures in Text

Every table and every figure must be mentioned by number in the text before it appears. A table or figure that no sentence points to will be flagged as an orphan during review.

The reference pattern is straightforward:

• Introduce it before showing it. Place the sentence that mentions the table or figure before the visual itself in the manuscript. The visual then appears at the next logical break in the text (not necessarily immediately).
• Describe the key finding, not the table itself. "As shown in Table 2, the intervention group outperformed controls on all three measures" tells the reader what to look at. "Table 2 shows the results" tells the reader nothing.
• Do not restate the entire table. If a value is in the table, do not repeat it in the paragraph. Highlight the finding, point to the table, move on.
• Capitalize Table and Figure when numbered. "See Table 1" uses a capital T. "The table above" uses a lowercase t because it is not naming a specific numbered element.

Table and Figure Placement

Placement conventions vary by journal and submission stage. Three patterns cover most cases.

Embedded in text. Many journals now ask for tables and figures embedded close to their first mention in the submitted manuscript. The copy editor repositions them during typesetting.

Separate files. Some journals require tables as separate documents and figures as separate image files, with placement indicated by markers in the manuscript text ("[Insert Table 1 about here]").

End of manuscript. APA style traditionally placed tables and figures at the end of the manuscript, after references. This remains common in dissertation drafts and in some journal submission formats.

Check the submission guidelines of your target journal before deciding on placement. Journals vary, and submitting in the wrong format can cost a round at the desk-check stage.

Common Mistakes

Using a figure when a table would be clearer. Bar charts of three or four means with error bars are often less informative than a small table of the same means with standard deviations. The figure wastes space and adds no clarity.

Using a table when prose would be clearer. A table with two rows and two columns, showing a single comparison, is a sentence stretched across half a page. Write the sentence instead.

Restating the whole table in the text. If every value in the table appears in the surrounding paragraph, the table is redundant. Delete either the paragraph detail or the table.

Unlabeled or vaguely labeled axes. Reviewers lose patience quickly with figures that force them to guess what the axes measure. Label clearly, with units.

Decorative color. Rainbow bar charts and gratuitous 3D effects distract from the data. Use color only when it carries information (differentiating groups, encoding a third variable).

Inconsistent decimal places. A table with means reported to two decimals and standard deviations to three, with no reason for the difference, reads as inattention. Match decimal places within each column.

Figures or tables that are not mentioned in the text. Every table and every figure needs a sentence pointing to it. Orphan visuals get flagged at review.

Too many decimals. Reporting a mean to four decimal places implies precision that the measurement does not support. Two decimals is usually enough.

How Many Is Too Many?

A reasonable guideline for most quantitative empirical papers: no more than three to five tables and three to five figures total, across the entire manuscript. Dissertation chapters may run longer. Some journals impose explicit limits in the author guidelines.

The question to ask about any individual visual: does the paper lose something important if this is cut? If the answer is yes, keep it. If the paper would be fine without it, remove it. Reviewers reward restraint.

Before You Submit: A Self-Audit

Tables and figures are among the elements copy editors examine most closely before publication, and small errors are the kind of thing an author stops seeing after the fourth or fifth draft. If you have a manuscript in preparation and want a professional editor to review your tables, figures, results section, or full paper before submission, Editor World offers journal article editing by editors with subject-matter backgrounds. Clients browse editor profiles and choose the editor whose expertise best matches their field. A free sample edit on the first 300 words is available for every project, and a certificate of editing is available as an optional add-on for publishers or committees that require one.

Frequently Asked Questions