Good Fonts for Dyslexia
Luz Rello, Ricardo Baeza-Yates · 2013 · Proceedings of the 15th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS) · doi:10.1145/2513383.2513447
Summary
This paper presents the first experiment using eye tracking to objectively measure the impact of font type on reading performance for people with dyslexia. Using a within-subject design, 48 participants with clinically confirmed dyslexia (ages 11-50) read 12 comparable texts each presented in a different font: Arial, Arial Italic, Computer Modern Unicode, Courier, Garamond, Helvetica, Myriad, OpenDyslexic, OpenDyslexic Italic, Times, Times Italic, and Verdana. Fonts were selected to cover key typographic characteristics (serif vs. sans serif, monospaced vs. proportional, roman vs. italic) and included commonly used fonts, fonts recommended by dyslexia organizations, and fonts designed specifically for dyslexia. Texts were carefully controlled: all extracted from the same book, matched at 60 words each, with similar word lengths and no numerical expressions or acronyms. Texts were presented left-justified, 14-point, black on white, with columns no wider than 70 characters — following British Dyslexia Association recommendations. Reading performance was measured via two eye-tracking metrics (reading time and fixation duration) with comprehension questions as a control variable, plus subjective preference ratings.
Key findings
Font type has a statistically significant effect on both reading time and fixation duration for people with dyslexia. Arial Italic performed worst on both measures — significantly longer reading times than Helvetica, Arial, and CMU, and significantly longer fixation durations than eight other fonts. Courier produced the shortest fixation durations, significantly shorter than six other fonts. At the font-characteristic level, sans serif, monospaced, and roman styles significantly improved reading performance over serif, proportional, and italic styles respectively, though these effects were stronger for fixation duration than reading time. Italic fonts consistently decreased readability. Notably, OpenDyslexic — a font designed specifically for dyslexia — did not lead to better or worse readability compared to standard fonts, and participants significantly preferred Verdana and Helvetica over OpenDyslexic. The recommended "good fonts" considering both performance and preference are Helvetica, Courier, Arial, Verdana, and CMU. Participant preferences showed almost no correlation with actual reading performance (Pearson r = -0.13), meaning readers cannot reliably identify which fonts help them read faster.
Relevance
This study provides the first objective, eye-tracking-based evidence for font recommendations that had previously been based only on expert opinion or anecdotal guidance from dyslexia organizations. The finding that sans serif, monospaced, and roman fonts improve readability gives practitioners concrete, evidence-based guidance for text presentation. Particularly important is the finding that Arial Italic should be avoided entirely, and that italic fonts in general decrease readability for dyslexic readers — a direct challenge to the common use of italics for emphasis in web and document design. The lack of improvement from OpenDyslexic challenges the assumption that specially designed dyslexia fonts are inherently better than well-chosen standard fonts. For web developers and content authors, the practical takeaway is clear: use Helvetica, Arial, Courier, or Verdana; avoid italic styles; and prefer sans serif over serif fonts. These findings are relevant to WCAG compliance, though WCAG itself does not currently specify font recommendations for dyslexia.
Tags: dyslexia · typography · readability · eye tracking · font design · cognitive accessibility · text presentation · legibility
Standards referenced: WCAG