Accessible Texts for Autism: An Eye-Tracking Study
Victoria Yaneva, Irina Temnikova, Ruslan Mitkov · 2015 · ASSETS '15: Proceedings of the 17th International ACM SIGACCESS Conference on Computers & Accessibility · doi:10.1145/2700648.2809852
Summary
This pioneering study is the first to use eye-tracking technology with autistic participants to evaluate text document accessibility. People with Autism Spectrum Disorder experience difficulties with reading comprehension—including processing complex sentences, understanding figurative language and abstract words, and grasping context—combined with atypical attention patterns that make designing accessible documents challenging. The researchers recruited 20 adults with confirmed autism diagnoses (no co-occurring learning disabilities) and 20 neurotypical controls to read 9 easy-read documents while their eye movements were recorded. The study investigated four questions: (1) whether autistic and non-autistic readers differ in how they allocate attention between images and text, (2) whether photographs or symbols impose different cognitive loads, (3) how adults with autism perceive the difficulty of Plain English documents, and (4) what text presentation formats autistic readers prefer. Materials were drawn from a corpus of 100 easy-read documents from UK charity organizations, government departments, and healthcare services, with Flesch Reading Ease scores ranging from 60-95. The researchers used gaze metrics including average time viewed, number of fixations, and revisits to areas of interest (images vs. text paragraphs).
Key findings
The study found significant differences in attention allocation between groups. Autistic participants spent 20.32% of their reading time looking at images compared to 13.42% for controls—a statistically significant difference (p=0.003) even when controlling for longer overall reading times in the ASD group. This confirms that images capture substantially more attention from autistic readers, making image selection critically important. Contrary to assumptions in prior literature that autistic individuals struggle with abstract symbols due to impaired generalization abilities, the study found no significant difference in processing time between photographs and symbols for adult participants without developmental delay. Both image types imposed similar cognitive loads, suggesting either can be used in documents for this population—though this finding may not generalize to children or adults with co-occurring learning disabilities. Regarding difficulty perception, autistic participants rated the easy-read texts across a wider range (from "very easy" to "difficult") while controls rated nearly all texts as "very easy." This suggests Plain English documents are appropriately challenging for autistic adults—comprehensible without being trivially easy. A majority of autistic participants (58.81%) preferred texts with images, compared to only 30% of controls (60% of whom were indifferent to image inclusion).
Relevance
This research provides concrete, empirically-grounded guidelines for creating accessible content for autistic adults—a population often overlooked in cognitive accessibility work that focuses primarily on intellectual disabilities. The eight recommendations include: illustrate main ideas with relevant images; avoid irrelevant images, logos, or advertisements that may distract; place images close to related text; use Plain English (Flesch score >65); and allow readers to proceed at their own pace. For web developers and content creators, the key insight is that images matter more to autistic readers than intuition might suggest—they are not mere decoration but active comprehension aids that consume significant attentional resources. This means irrelevant or decorative images may harm rather than help. The finding that both photographs and symbols work equally well for adults simplifies implementation, though the authors caution against overly abstract symbols requiring domain-specific knowledge. The study has acknowledged limitations: small sample size (common in autism research due to population heterogeneity), and results should not be generalized to children or adults with learning disabilities. The research also notes that only 6% of UK adults with autism are employed full-time, making reading accessibility improvements relevant to educational and employment outcomes.
Tags: autism · ASD · eye tracking · reading comprehension · easy-read · cognitive accessibility · text simplification · Plain English · visual attention