Multilingual Website Assessment for Accessibility: A Survey on Current Practices
Silvia Rodríguez Vázquez, Anton Bolfing · 2013 · Proceedings of the 15th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS) · doi:10.1145/2513383.2513428
Summary
This short paper presents results from a survey of 67 web accessibility professionals exploring how multilingual websites are assessed for accessibility. The authors note that when a monolingual website is localized into multiple languages, the accessibility level achieved in the original version may not be maintained across all language versions. Localization involves far more than translation — it encompasses adapting colours, images, date formats, menu sizes, page structure, and other semiotic and cultural elements, all of which can affect accessibility. The survey targeted experienced professionals with at least two years in web accessibility, recruited via snowball sampling through mailing lists, the WebAIM discussion list, Twitter, and LinkedIn. The study examined two areas: the procedures followed during multilingual accessibility assessment tasks, and perceptions of whether localization professionals could contribute to achieving higher accessibility in multilingual sites. The paper focuses on the first area, investigating whether standardized patterns exist for assessing localized websites.
Key findings
The survey revealed that current multilingual website accessibility assessment practices lack standardization. Evaluators tend to focus on elements within their own area of expertise but do not follow a consistent, well-defined pattern for ensuring accessibility is maintained across language versions. Perhaps most notably, the time spent assessing textual and culture-related elements — which remain key information assets on any webpage — was considerably low. This suggests that critical aspects of localized content, such as whether translated text maintains proper semantic structure, appropriate reading level, or correct language attributes, may be overlooked during accessibility evaluations. The findings also indicate that localization-specific knowledge and skills are needed but generally absent from current accessibility assessment workflows. No clear procedure emerged from the data for verifying that accessibility achievements in an original website are preserved through the localization process.
Relevance
This research highlights a significant blind spot in web accessibility practice: the assumption that an accessible website in one language remains accessible when localized. For organizations operating multilingual sites, this is a critical concern — accessibility audits conducted only on the primary language version may miss issues introduced during translation and cultural adaptation. The findings suggest that accessibility teams and localization teams need to collaborate more closely, and that assessment methodologies should explicitly account for multilingual content. For practitioners, the takeaway is clear: test every language version, not just the source. While the study is a short poster paper with a modest sample size and self-reported data, it opened an important area of inquiry at the intersection of localization and accessibility that remains underexplored.
Tags: web accessibility · multilingual web · localization · accessibility assessment · internationalization · evaluation methods
Standards referenced: WCAG