Accessibility is usually framed as a compliance issue for websites. In publishing it is simpler than that: some of your readers cannot see your figures, some cannot distinguish the colours in them, and some navigate your paper by heading structure rather than by eye. Colour-vision deficiency is far more common in men than in women. Researchers with low vision use screen readers daily. Designing for them costs almost nothing and improves the paper for everyone.
Figures that survive colour-blindness — and greyscale
A frequent accessibility failure in research is a figure where colour is the only thing distinguishing the data.
- Never encode meaning in colour alone. Add distinct shapes for points, different line styles (solid, dashed, dotted), or hatch patterns for bars.
- Avoid red-green and green-brown pairings, which readers with red-green colour-vision deficiency cannot separate.
- Label directly where you can, rather than sending readers to a colour key.
- Check in greyscale. Print your figure in black and white: if you can still read it, it will survive most colour-vision deficiencies too.
- Use a palette built for colour-vision deficiency. The Okabe–Ito palette and the viridis colour maps were designed so that their colours stay distinguishable.
Our guide on preparing figures and tables covers the resolution and format side.
For contrast, the Web Content Accessibility Guidelines (WCAG), published by the World Wide Web Consortium (W3C), give a useful benchmark. At level AA, normal text needs a contrast ratio of at least 4.5:1 against its background, and large text at least 3:1. Graphical parts needed to understand a chart, such as lines and markers, need at least 3:1 against adjacent colours.
Write alt text for every figure
Alt text is a short description read aloud in place of an image. It is not the caption repeated — the caption says what the figure is; the alt text says what it shows. Aim for one or two sentences conveying the finding:
Caption: Figure 2. Mean HbA1c by treatment group over 12 months.
Alt text: Line chart showing HbA1c falling from 8.1% to 7.0% in the intervention group and remaining near 8.0% in the control group over twelve months.
Only the author knows what matters in the image, which is why journals increasingly ask you to supply it.
The figure and numbers in that example are invented to illustrate the structure; they are not real data.
Rules for alt text that works
- Skip "image of". Screen readers already announce that an image follows.
- Lead with the finding. Name the chart type, then the pattern the reader should take away.
- Do not rely on colour. "The red line rises" means nothing to a reader who cannot see red; name the group instead.
- Keep complex figures short. For a dense multi-panel figure, summarise the main finding in the alt text and put the detail in the text or in a table.
- Make the numbers available. If a chart carries values a reader may need, give them in a table, or in a supplementary file as described in preparing supplementary materials.
Structure your text properly
Screen-reader users navigate by headings, exactly as sighted readers skim them. Use real heading levels in order rather than bold text that only looks like a heading, keep them descriptive, and write meaningful link text — "see the reporting guidelines" rather than "click here". Spell out abbreviations at first use; assistive technology reads unfamiliar acronyms letter by letter.
In Microsoft Word, apply the built-in heading styles rather than enlarging the font by hand. Styled headings carry through to the navigation pane and to a tagged PDF. Word also includes an accessibility checker that flags missing alt text and tables without header rows. Run it before you submit.
Tables and equations
Give tables genuine header rows so their structure is machine-readable, avoid merged cells where you can, and never use a table purely for visual layout. Where a journal supports it, supply equations in a proper markup format rather than as images — an equation embedded as a picture is silent to a screen reader.
A common table error is using blank rows or merged cells to create visual groups. A screen reader then reads empty cells or loses track of which header applies. The fix is to add a grouping column, or to split the table in two.
Supplementary files, video and audio
Accessibility applies to every file you upload, not only the main text. Supply supplementary documents as text-based files rather than scanned pages, which a screen reader cannot read. Give video captions, and give audio a transcript.
Accessibility checklist before submission
| Check | How to test it | Fix if it fails |
|---|---|---|
| Figures readable without colour | View in greyscale or a colour-blindness simulator | Add shapes, line styles or direct labels |
| Text and lines have enough contrast | Measure with a contrast checker against WCAG | Darken the colour or change the background |
| Every figure has alt text | Run Word's accessibility checker | Write one or two sentences giving the finding |
| Headings are real headings | Open the navigation pane | Apply heading styles in order |
| Tables have header rows and no layout-only cells | Read the table cell by cell | Mark the header row; remove blank and merged cells |
| No pictures of text or equations | Try to select the text | Retype it, or use the equation editor |
| Video and audio can be followed without sound | Play with the sound off | Add captions or a transcript |
Why it is worth the effort
Accessible papers reach more readers, and the same features that help assistive technology — descriptive headings, image descriptions, meaningful links, real text rather than pictures of text — are what indexing services and search engines read. Making your work usable by everyone and making it findable turn out to be largely the same task.
Updated 17 September 2026: We expanded this guide with WCAG contrast thresholds, alt-text rules, Word tips, a supplementary-file section and a checklist table.