Figures in a Research Paper – Types, Captions, and Examples

Figures in a research paper are visual displays—such as graphs, charts, photographs, maps, diagrams, and flowcharts—that communicate data or concepts more clearly than text alone. Effective figures are necessary, accurate, readable, self-contained, cited in the text, and formatted according to the required style guide or target journal.

Introduction

A strong research figure does more than make a paper look attractive. It helps readers identify patterns, compare groups, understand a process, inspect evidence, or follow a conceptual argument. A poorly designed figure can have the opposite effect: it can obscure results, exaggerate differences, introduce errors, or make an otherwise sound paper difficult to evaluate.

This guide explains what counts as a figure, how figures differ from tables, which figure type to choose, how to write captions, and how to integrate visuals into an academic argument. It also covers APA 7 formatting, journal submission specifications, accessibility, copyright, image integrity, supplementary figures, digital tools, and responsible use of artificial intelligence.

Key Takeaways

• A figure should communicate information more effectively than text or a table.
• Tables and figures are normally classified and numbered separately.
• Every figure should be mentioned by number and discussed in the main text.
• Formatting rules vary by style guide, discipline, institution, and journal.
• A figure must remain accurate, readable, and understandable at its final published size.
• Reused, adapted, digitally modified, or AI-assisted visuals may require attribution, permission, disclosure, or all three.

What Is a Figure in a Research Paper?

A figure is a numbered visual display used to present data, evidence, processes, objects, locations, relationships, or theoretical ideas. It complements the written discussion but should not be added merely as decoration.

Common figures include:

• Line graphs
• Bar charts
• Scatterplots
• Histograms
• Box plots
• Maps
• Photographs
• Microscopy images
• Medical images
• Flowcharts
• Conceptual frameworks
• Process diagrams
• Network diagrams
• Timelines
• Infographics
• Technical drawings
• Multi-panel composites

In APA Style, visual displays other than tables are generally treated as figures. Individual journals may use narrower or more specialized categories, such as artwork, plates, schemes, graphical abstracts, or extended-data figures.

Are Tables Considered Figures?

No. In most academic style systems, tables and figures are separate display elements.

A table presents information in rows and columns. A figure presents information visually through shapes, positions, lines, images, symbols, or spatial relationships. They are normally numbered in separate sequences:

• Table 1, Table 2, Table 3
• Figure 1, Figure 2, Figure 3

A chart created from a dataset is usually a figure. The spreadsheet-like grid containing the exact values is a table.

Why Are Figures Used in Research Papers?

Figures are useful when visual presentation makes evidence or an idea easier to understand.

An effective figure can:

• Reveal trends that are difficult to detect in a paragraph of numbers.
• Show the relationship between two or more variables.
• Compare categories, groups, locations, or time periods.
• Display a distribution and its outliers.
• Explain a research procedure or analytical workflow.
• Show the appearance, location, or structure of an object.
• Summarize a theoretical or conceptual model.
• Reduce the amount of repetitive description in the main text.
• Help readers evaluate whether the evidence supports the author’s interpretation.

Figures should not be used to repeat information that is already easy to understand from the text. They should add information value rather than visual decoration.

Should You Use Text, a Table, or a Figure?

Choose the format according to what the reader needs to learn.

Reader’s need	Best format	Example
Understand one or two important values	Text	“The response rate was 76%.”
Look up exact values for several groups	Table	Means, standard deviations, and sample sizes for six groups
Recognize a trend over time	Line graph	Monthly temperature across one year
Compare a few categories visually	Bar chart or dot plot	Mean scores for four teaching methods
Examine the relationship between continuous variables	Scatterplot	Study time and examination score
Understand the shape of a distribution	Histogram, density plot, or box plot	Distribution of household income
Follow a procedure or decision sequence	Flowchart	Participant screening and selection
Understand a theoretical relationship	Conceptual diagram	Relationships among motivation, engagement, and achievement
Inspect a physical or visual phenomenon	Photograph or scientific image	Tissue sample, study site, artifact, or experimental apparatus

A display is unnecessary when a sentence communicates the information more efficiently. Conversely, a long paragraph that forces readers to mentally compare many values may be better converted into a table or figure.

Examples of Figures in Research Paper

Examples of Figures in Research Papers or Thesis are as follows:

Line graphs Example

Bar graphs Example

Pie charts Example

Scatterplots Example

Tables Example

Photographs or images Example

Diagrams or schematics Example

Types of Figures in Research Papers

Line Graphs

A line graph is best for showing change across a continuous dimension, especially time. The horizontal axis usually represents time or another ordered variable, while the vertical axis represents the measured outcome.

Use a line graph when:

• Measurements occur at multiple ordered points.
• The direction or rate of change matters.
• Several groups follow comparable trajectories.

Do not connect categories with a line when there is no meaningful continuity between them.

Bar Charts and Dot Plots

A bar chart compares values across distinct categories. A dot plot can often communicate the same comparison with less visual weight.

Use these displays when:

• The horizontal variable consists of categories.
• The reader needs to compare magnitudes.
• There are relatively few categories.

Avoid decorative three-dimensional bars. They distort perceived size and make values harder to compare.

Scatterplots

A scatterplot displays the relationship between two quantitative variables. Each point represents an observation.

Scatterplots are useful for showing:

• Positive or negative associations
• Nonlinear patterns
• Clusters
• Outliers
• Variation around a fitted line

A trend line should not replace the underlying observations unless space or confidentiality requirements make this necessary.

Histograms and Density Plots

A histogram groups observations into intervals and shows the shape of a distribution. A density plot presents a smoothed estimate of that distribution.

These displays help readers examine:

• Skewness
• Multiple peaks
• Gaps
• Extreme values
• Approximate normality

Histogram appearance depends on bin width, so the chosen bins should not be manipulated to create a misleading pattern.

Box Plots and Violin Plots

A box plot summarizes the median, quartiles, spread, and possible outliers. A violin plot adds information about distribution density.

These figures are useful for comparing distributions across groups. When sample sizes are modest, adding individual data points often gives readers a more complete view than a summary shape alone.

Error-Bar and Interval Plots

Error bars can represent standard deviations, standard errors, confidence intervals, or other quantities. The caption must define them precisely.

“Error bars show variability” is not sufficient. The reader should be told whether the bars indicate:

• Mean ± standard deviation
• Mean ± standard error
• 95% confidence intervals
• Minimum and maximum
• Interquartile range
• Another stated measure

Heatmaps

A heatmap uses color intensity to represent values in a matrix. It is useful for correlation matrices, gene-expression data, geographic intensity, response patterns, and other high-dimensional relationships.

A heatmap requires:

• A clearly labelled scale
• A perceptually ordered color scheme
• Sufficient contrast
• An alternative to color-only interpretation where possible

Flowcharts

Flowcharts represent sequences, decisions, exclusions, or procedures. They are commonly used for:

• Participant recruitment
• Systematic-review screening
• Experimental protocols
• Data-cleaning workflows
• Algorithms
• Clinical decisions

Every shape, arrow, abbreviation, and branching rule should have a clear meaning.

Conceptual Frameworks and Models

A conceptual figure shows proposed relationships among constructs, variables, mechanisms, or stages.

Use one when a visual model is clearer than a long verbal explanation. Distinguish observed evidence from proposed or hypothetical relationships through labels, line styles, or explicit caption wording.

Maps

Maps may show study locations, spatial distributions, boundaries, routes, or geographic comparisons.

A research map may require:

• Scale
• Orientation
• Coordinate system or projection
• Legend
• Data source
• Boundary source
• Appropriate treatment of politically disputed areas

Photographs and Scientific Images

Photographs may document a study site, participant activity, specimen, artifact, experimental setup, clinical condition, or observed phenomenon.

Scientific images often require scale bars, acquisition details, staining information, orientation, panel labels, and disclosure of processing. Images of identifiable people may also require consent to publish.

Multi-Panel Figures

A multi-panel figure combines related visuals under one figure number. Panels are usually labelled A, B, C, and so forth.

Combine panels only when they contribute to one coherent message. A multi-panel figure should not become a storage space for unrelated results.

Essential Parts of a Research Figure

The exact arrangement depends on the style guide or journal, but a complete figure may contain the following elements.

Figure Number

The figure number identifies the display and permits accurate cross-referencing. Figures are usually numbered according to the order in which they are first mentioned.

Figure Title

The title briefly identifies the subject of the figure.

A useful title is specific:

Change in Mean Reading Score Across the 12-Week Intervention

A weak title is vague:

Study Results

Some journals do not display a separate title and instead begin the caption with the figure number and a descriptive sentence.

The Visual Display

The visual is the graph, map, photograph, diagram, or other image itself. Its design should make the principal comparison or message easy to perceive.

Labels and Units

Axes, categories, locations, components, and measurements must be clearly labelled. Units should be stated wherever they are needed for interpretation.

Write:

• Temperature (°C)
• Concentration (mg/L)
• Time (weeks)
• Response time (ms)

Do not assume the unit is obvious from context.

Legend or Key

A key explains what colors, symbols, line types, patterns, or shapes represent. It should be positioned where readers can connect it easily to the visual.

In some scientific disciplines, the term figure legend refers to the entire explanatory caption rather than the key inside the figure. Authors should follow the terminology used by their target journal.

Figure Note

A note may define abbreviations, explain statistical markers, identify a source, provide copyright attribution, or clarify information that does not fit naturally inside the visual.

Scale Bar

Photographs, microscopy images, and spatial images should include a scale bar when physical size cannot otherwise be judged.

Panel Labels

Multi-panel figures should use consistent labels such as A, B, and C. The caption should explain each panel in the same order.

How to Create and Add a Figure to a Research Paper

Step 1: Define the Figure’s Main Message

Complete this sentence:

“After viewing this figure, the reader should understand that…”

A figure without a clear message often becomes cluttered because it tries to display every available result.

Step 2: Identify the Reader’s Task

Decide whether readers need to:

• Look up exact values
• Compare categories
• Follow change over time
• Inspect a distribution
• Evaluate an association
• Understand a process
• Examine an object or location
• Interpret a theoretical model

The task determines the appropriate display.

Step 3: Choose the Correct Figure Type

Match the figure to the data structure and research question. Do not choose a graph merely because the software makes it easy to generate.

Step 4: Create the Figure from Verified Data

Use a cleaned and documented dataset. Check that:

• The plotted variables are correct.
• Filters and exclusions are intentional.
• Units match the manuscript.
• Category labels have not been switched.
• Missing values have been handled consistently.
• Statistical summaries match the reported results.

Maintain the underlying data and editable source file.

Step 5: Design for the Final Display Size

A figure that looks clear on a large monitor may become unreadable when reduced to one journal column.

Before export:

• Reduce the figure to its expected publication size.
• Check the smallest text.
• Inspect line thickness and symbols.
• Confirm that panel labels remain visible.
• Remove unnecessary borders, backgrounds, and gridlines.

Step 6: Make the Figure Accessible

Use color as a supporting cue, not the only cue. Combine it with:

• Different symbols
• Solid and dashed lines
• Direct labels
• Patterns
• Clear panel separation

Test the figure in grayscale and, where possible, with a color-vision simulator.

Step 7: Write a Self-Contained Caption

The caption should let readers understand the essential display without searching through the paper for basic definitions.

Include only the methods and statistics needed to interpret the figure. The caption should not become a duplicate Methods or Discussion section.

Step 8: Refer to and Interpret the Figure in the Text

Mention the figure by number before or near its appearance:

Reading scores increased during the intervention and remained stable during follow-up (Figure 2).

Then explain the relevant pattern:

The largest increase occurred between Weeks 2 and 6, after which the rate of improvement slowed.

Do not merely repeat every visible value.

Step 9: Validate the Figure Against Submission Rules

Before submission, check:

• Number of figures permitted
• Figure dimensions
• Accepted file formats
• Required resolution
• Font and line-size limits
• Color mode
• Separate-file requirements
• Caption-file requirements
• Supplementary figure conventions
• Accessibility or image-description fields
• Permission and disclosure requirements

The target journal’s current instructions override generic online advice.

How to Write a Figure Caption

A figure caption identifies the figure and supplies the information needed to interpret it correctly.

A strong caption commonly answers five questions:

1. What does the figure show?
2. What do the panels, groups, lines, symbols, or colors represent?
3. What sample, conditions, or method are essential for interpretation?
4. What do the error bars or statistical symbols mean?
5. Is a source, permission statement, or disclosure required?

General Figure-Caption Template

Figure X. [Concise title]. [What the figure or each panel shows.] [Essential sample, measurement, condition, or analytical information.] [Definition of error bars, symbols, abbreviations, or scale.] [Source, permission, or disclosure when required.]

Statistical Figure Example

Figure 2. Weekly reading scores during the intervention and follow-up periods. Points represent group means for 84 students; shaded areas indicate 95% confidence intervals. The vertical dashed line marks the end of the 12-week intervention.

This is a hypothetical formatting example, not a report of an actual study.

Multi-Panel Caption Example

Figure 3. Participant recruitment and outcome distribution. (A) Flow of participants from eligibility assessment to final analysis. (B) Distribution of post-intervention scores by study group. Boxes indicate the interquartile range, center lines indicate medians, and points represent individual observations.

Photograph Caption Example

Figure 4. Northern sampling location during the dry season. The marked area indicates the 20 m × 20 m vegetation plot. Photograph by the author, September 2025.

Conceptual-Model Caption Example

Figure 5. Proposed relationship among academic support, self-efficacy, engagement, and persistence. Solid arrows represent relationships tested in the present study; the dashed arrow represents a proposed indirect relationship for future investigation.

Caption Versus Interpretation

The caption explains how to read the figure. The main text explains why the result matters.

Caption:

Bars show group means; error bars indicate 95% confidence intervals.

Main-text interpretation:

The confidence intervals for Groups A and B overlapped substantially, suggesting that the observed difference should not be interpreted from bar height alone.

How to Format Figures in APA 7

APA 7 treats graphs, charts, maps, drawings, photographs, and other non-table visual displays as figures.

A standard APA figure contains:

1. A bold figure number above the figure
2. An italicized title in title case on the next line
3. The figure image
4. A legend or key inside the figure when needed
5. A note beneath the figure when clarification, attribution, or statistical explanation is required

Example structure:

Figure 1

Mean Test Scores Across Four Assessment Periods

[Insert figure]

Note. Error bars represent 95% confidence intervals.

APA Numbering

Number figures independently from tables and in the order of first mention:

• Table 1
• Figure 1
• Figure 2
• Table 2

Do not number figures according to the page on which they appear.

APA In-Text Callouts

Refer to each figure by its number:

As shown in Figure 1, completion rates increased after the intervention.

Avoid location-dependent wording:

The figure below shows…

Page layout can change during editing, typesetting, or digital display, making “above” and “below” unreliable.

APA Figure Placement

APA 7 generally permits two approaches:

• Embed each figure in the text after it is first mentioned.
• Place figures on separate pages after the reference list.

An instructor, department, thesis handbook, or journal may require one option. Follow the most specific applicable instruction.

APA Notes and Sources

An original figure created entirely from the author’s own data normally does not require a source citation.

A figure reproduced or adapted from another work requires acknowledgement in the figure note and a corresponding reference-list entry. Copyright or licence wording may also be required.

Journal and Discipline-Specific Formatting

There is no universal figure format for every research paper.

A journal may require:

• Captions in a separate section
• One figure per file
• Figures removed from the manuscript document
• Figure-placement markers in the text
• A particular naming convention
• Specific maximum dimensions
• A limited number of panels
• Defined fonts and line widths
• A particular color space
• Separate graphical abstracts or highlights
• Original, uncropped image files

A university assignment may instead require figures embedded in a Word document. A dissertation may require a list of figures even though the chosen citation style does not require one.

Use this order of authority:

1. Target journal or publisher instructions
2. University, department, or instructor requirements
3. Required style manual
4. General academic-writing advice

Figure File Formats and Technical Quality

Raster Versus Vector Figures

A raster image is built from pixels. Photographs, scans, microscopy images, and many screenshots are raster images.

Common raster formats include:

• TIFF
• PNG
• JPEG

A vector image is built from mathematical shapes and paths. It can be enlarged without becoming pixelated.

Common vector-capable formats include:

• SVG
• EPS
• PDF
• AI

Graphs, diagrams, line drawings, and conceptual models are often best preserved as vectors when the journal accepts them.

Choosing a File Format

Figure type	Commonly suitable format	Main consideration
Photograph	TIFF or high-quality JPEG	Preserve detail; avoid repeated lossy compression
Line graph	SVG, EPS, or vector PDF	Keep text and lines scalable
Diagram	SVG, EPS, PDF, or editable source	Preserve shapes, labels, and arrows
Combination of image and labels	TIFF or journal-approved vector/raster format	Maintain both image detail and readable text
Web figure	SVG or PNG	Check accessibility and browser support
Initial Word assignment	Embedded high-quality image	Retain the editable original separately

The journal’s accepted formats take priority.

Resolution

Resolution requirements depend on figure type and publisher. Photographs are often accepted at lower pixel density than monochrome line art because fine lines and text need greater sharpness.

Common publisher recommendations may range from approximately:

• 300 ppi for color photographs
• 600 ppi for combination artwork
• Up to 1,200 ppi for monochrome line art

These are examples, not universal requirements. Exporting a low-resolution image at a larger nominal ppi does not recreate missing scientific detail.

PPI and DPI

Pixels per inch, or ppi, describes the density of pixels in a digital image. Dots per inch, or dpi, properly describes printer output, although publishers often use the terms interchangeably in author instructions.

Follow the terminology and measurement method used by the target publication.

Fonts and Labels

Use a consistent, readable font throughout the paper’s figures. Check that:

• Capitalization is consistent.
• Variables use the same symbols as the text.
• Decimal places are appropriate.
• Greek letters and mathematical symbols export correctly.
• Labels remain readable after reduction.
• Text is editable or fonts are embedded when required.

Avoid Screenshots When an Export Is Available

A screenshot of a graph often produces blurred text and an unsuitable background. Export the figure directly from the program in a vector or high-resolution raster format.

Statistical and Visual Integrity

Figures are part of the research evidence. They must represent data honestly.

Use Appropriate Axes

A shortened vertical axis can make a small difference look dramatic. A non-zero baseline is not automatically wrong, especially for line graphs, but it must not conceal scale or exaggerate meaning.

When an axis break or truncated range is necessary:

• Mark it clearly.
• Explain the decision if it affects interpretation.
• Consider showing the full range in supplementary material.

Show Uncertainty Clearly

Define every interval or error bar. Where appropriate, report:

• Sample size
• Number of independent replicates
• Statistical test
• Confidence level
• Exact or threshold p-values
• Adjustment for multiple comparisons

Do not use statistical symbols that are explained only in another figure.

Distinguish Observations from Summaries

A bar showing a mean can conceal distribution shape, unequal variance, and outliers. Where feasible, add raw points, interval plots, box plots, or distribution displays.

Avoid Selective Presentation

Do not:

• Remove inconvenient observations without a stated analytical reason.
• Show only the most favorable experimental image.
• hide background bands or signals.
• change one image differently from comparable images.
• splice images without visible boundaries and explanation.
• duplicate a panel while presenting it as a different condition.
• use a color scale that creates a false threshold or pattern.

Retain the original data and unprocessed image files. Document legitimate cropping, contrast adjustments, normalization, and processing.

How to Discuss Figures in the Main Text

Every figure should have a purpose in the paper’s argument.

A useful figure discussion usually performs three actions:

1. Directs the reader to the figure.
2. Identifies the relevant pattern.
3. Explains its importance to the research question.

Example:

Figure 2 shows that both groups improved during the first four weeks. However, only the intervention group continued to improve between Weeks 4 and 8, producing a wider group difference at the final assessment.

This is more useful than:

Figure 2 shows the results.

Do not reproduce every value in prose. Highlight the comparisons, trends, anomalies, or relationships that matter.

When to Mention the Figure

Mention the figure before it appears or in the same paragraph as its first appearance, unless the required style says otherwise.

Figures may appear in different sections:

• Introduction: A conceptual model, location map, or background framework
• Methods: Experimental setup, sampling map, protocol, or participant-flow diagram
• Results: Graphs, images, and outcome visualizations
• Discussion: Usually interpretation of existing figures rather than presentation of new results
• Appendix or supplement: Supporting analyses and extensive documentation

Multi-Panel, Supplementary, and Appendix Figures

Multi-Panel Figures

Use one number with lettered panels:

• Figure 2A
• Figure 2B
• Figure 2C

The caption should explain panels in order. Labels should be consistent in size, position, and style.

Supplementary Figures

Supplementary figures support the paper but are not essential to its central argument. They may contain:

• Sensitivity analyses
• Diagnostic plots
• Additional experimental controls
• Full versions of cropped displays
• Secondary outcomes
• Extended methodological documentation

Common numbering includes Figure S1, Figure S2, and Figure S3. Follow the journal’s required convention.

A supplementary figure still needs a caption and an in-text callout.

Appendix Figures

In theses and reports, appendix figures may be numbered according to the appendix:

• Figure A1
• Figure A2
• Figure B1

Institutional rules may use alternatives such as Figure A.1.

How to Cite, Adapt, or Reproduce a Figure

Citation and copyright permission are related but separate.

A citation tells readers where the material originated. Permission or an applicable licence gives the legal basis for reuse.

Original Figure from Your Own Data

A figure created entirely from your own data does not normally need a source citation. Keep the underlying dataset and editable figure file.

Original Visualization of Someone Else’s Data

When you create a new graph from a published dataset, the visual design may be yours, but the data are not. Cite the original dataset or publication in the caption or note and in the reference list.

Example:

Note. Figure created by the author from data reported by Ahmed and Chen (2024).

Adapted Figure

An adapted figure changes the original layout, labels, visual structure, or content while remaining recognizably based on the source.

Use wording such as:

Adapted from…

Cite the source and check its licence or obtain permission where required.

Reproduced Figure

A reproduced figure is copied substantially as published.

Use wording such as:

Reprinted from…

Citation alone does not automatically provide permission. Check the copyright owner, licence terms, publisher permissions service, and applicable legal exceptions.

Conceptual Figure Based on Existing Literature

A diagram that synthesizes ideas from several sources can be an original figure, but the underlying theories and relationships still require citation.

Example:

Note. Model developed by the author based on the frameworks of Rahman (2022) and Lewis et al. (2024).

Photographs of People

A photograph may require:

• Research-participant consent
• Consent to publish
• Removal of identifying information
• Ethics-committee approval
• Permission from the photographer or copyright holder

Anonymization should be considered before publication, not only after the figure has been designed.

Making Figures Accessible

Accessibility improves a figure for many readers, including people with low vision, color-vision deficiencies, reading disabilities, small screens, or black-and-white printouts.

Do Not Rely on Color Alone

Instead of distinguishing groups only as red and green, combine color with:

• Circle and triangle markers
• Solid and dashed lines
• Direct labels
• Pattern fills
• Separate panels

Use Meaningful Contrast

Adjacent colors should have sufficient luminance contrast. Very pale lines, low-contrast text, and rainbow color scales can make important differences difficult to perceive.

Write Alternative Text

Alt text briefly identifies the figure and communicates its purpose or principal result.

Example:

Line graph showing that the intervention group’s mean reading score increased from Week 1 to Week 12, while the comparison group changed little.

Do not begin every description with “Image of” unless the image type is meaningful. A complex chart may require both brief alt text and a longer description containing the important values, trends, or relationships.

Provide the Information in Text

The paper should state the principal finding represented in the figure. A reader who cannot see the visual should still be able to understand the argument.

Check Final-Size Legibility

Accessibility fails when labels become unreadable after journal reduction. Test figures at their expected final width, not only at full-screen size.

Digital Tools for Creating Research Figures

The best tool depends on the figure type and the need for reproducibility.

Task	Common tools
Basic charts	Excel, Google Sheets, LibreOffice Calc
Statistical figures	R with ggplot2, Python with Matplotlib, GraphPad Prism, SPSS, Stata
Vector diagrams	Adobe Illustrator, Inkscape, Affinity Designer
Scientific-image analysis	ImageJ or Fiji
Conceptual or process diagrams	PowerPoint, Visio, diagrams.net, Lucidchart
Typeset figures and manuscripts	LaTeX, TikZ, PGFPlots, Overleaf
Geographic figures	QGIS, ArcGIS, R, Python
Biological and medical illustrations	BioRender and discipline-specific icon libraries
Reproducible notebooks	R Markdown, Quarto, Jupyter notebooks

Reproducible Figure Workflows

A reproducible workflow generates the figure from a documented dataset and script rather than editing values manually.

A strong workflow preserves:

• Raw data
• Cleaned data
• Analysis code
• Plotting code
• Software and package versions
• Editable source files
• Export settings
• Final submission files

This makes corrections easier and reduces inconsistencies between the figure, statistics, and written results.

Artificial Intelligence and Research Figures

AI tools can assist with figure preparation, but they cannot replace author verification or research integrity.

Potentially appropriate uses include:

• Brainstorming a conceptual layout
• Generating a draft caption from author-supplied information
• Suggesting accessible labels
• Identifying inconsistent terminology
• Producing a non-data conceptual illustration when permitted
• Checking whether a caption defines abbreviations and panels
• Converting a rough sketch into an editable diagram, subject to verification

High-risk or unacceptable uses include:

• Generating experimental results that were never observed
• Adding, removing, or altering features in microscopy, clinical, or laboratory images
• Inventing labels, scale bars, sample sizes, p-values, or mechanisms
• Producing a realistic image that readers could mistake for research evidence
• Copying the distinctive content of a protected figure without permission
• Uploading confidential participant images or unpublished data to an unapproved service

Publisher policies vary. Before using an AI-assisted visual:

1. Read the target journal’s current AI and image policies.
2. Check the tool’s licence and data-retention terms.
3. Retain prompts, versions, and original source material.
4. Verify every label, relationship, value, and citation.
5. Disclose the tool and its role when required.
6. Ensure that the author, not the tool, remains accountable for accuracy.

AI-generated captions should be treated as drafts. They may misidentify trends, confuse axes, omit limitations, or invent methodological details.

Advantages of Figures

Well-designed figures can:

• Communicate complex evidence efficiently.
• Reveal relationships that are difficult to describe.
• Improve navigation through long papers.
• Support comparison and pattern recognition.
• Make methods and conceptual models easier to follow.
• Allow readers to inspect evidence directly.
• Increase the usefulness of a paper across disciplines and formats.

Limitations of Figures

Figures also have limitations:

• Simplification can conceal important detail.
• Design choices may bias interpretation.
• Some readers need textual alternatives.
• Technical requirements vary across publications.
• Large numbers of figures can interrupt the argument.
• Visual summaries cannot replace methodological explanation.
• Reused images may involve copyright or consent restrictions.
• Static figures may not communicate uncertainty or multidimensional data adequately.

Use figures together with clear prose, appropriate tables, transparent methods, and access to underlying data where possible.

Common Mistakes

Mistake	Why it is a problem	Better approach
Adding a decorative figure	It consumes space without supporting the argument	Include only figures with clear information value
Calling a table a figure	It conflicts with standard numbering conventions	Number tables and figures separately
Referring to “the figure below”	Its location may change	Refer to “Figure 2”
Repeating all values in the text	It creates redundancy	Interpret the most important pattern
Using an unclear caption	Readers cannot understand the figure independently	Define groups, panels, symbols, and statistics
Omitting units	Values become ambiguous	Put units in axis or component labels
Using tiny text	Labels disappear after reduction	Design at final publication size
Relying only on color	The figure becomes inaccessible	Add symbols, patterns, or line styles
Exporting a screenshot	Text and lines may become pixelated	Export from the original program
Presenting only means	Distribution and outliers are hidden	Show points, intervals, or distribution summaries
Using undefined error bars	Uncertainty cannot be interpreted	State SD, SE, CI, or another measure
Copying a cited figure without permission	Citation may not authorize reuse	Check the licence and obtain permission
Selectively editing an image	It can misrepresent evidence	Apply permitted changes uniformly and disclose them
Trusting an AI-generated diagram without checking it	It may contain fabricated or incorrect content	Verify every element and follow journal policy

Pre-Submission Figure Checklist

Before submitting the paper, confirm that:

• Every figure has a clear purpose.
• Figures and tables are numbered separately.
• Numbering follows the order of first mention.
• Every figure is called out in the main text.
• The text interprets rather than duplicates the figure.
• Titles and captions follow the required style.
• Axes, units, groups, panels, and scale bars are labelled.
• Error bars and statistical symbols are defined.
• Values match the analysis and written results.
• No misleading axis, color, or scaling choice remains.
• Text remains legible at final publication size.
• Color is not the only method of differentiation.
• Alt text or an image description has been prepared where needed.
• The editable source and underlying data have been retained.
• Original scientific images have been preserved.
• Legitimate image processing has been documented.
• File format, resolution, dimensions, and naming meet journal rules.
• Supplementary figures follow the required sequence.
• Reused or adapted content is cited.
• Necessary copyright permissions and participant consents are available.
• AI assistance has been checked and disclosed when required.

Conclusion

Figures in a research paper should make evidence or ideas easier—not harder—to understand. The best figure is selected for a specific reader task, created from verified information, designed for its final size, explained by a self-contained caption, and interpreted in the text. Because APA, university, and journal requirements differ, authors should apply sound visual principles first and then follow the most specific current instructions governing their paper.