Italian Web Accessibility 2026: An Observational Analysis of 10,022 Domains and 34,641 Web Pages
Year: 2026
Type: preprint
Abstract: Italian Web Accessibility 2026: An Observational Analysis of 10,022 Domains and 34,641 Web Pages This paper presents the Italian Web Accessibility Study (IWAS), a large-scale, reproducible observational analysis of automatically detectable accessibility barriers across popular Italian websites. The study examines 10,022 .it domains selected from the Italian portion of the Tranco Top 1 Million ranking (list 46XZX, dated July 8, 2026). The same domain sampling frame was previously used in the Italian Web Security Study (IWSS), providing a common foundation for future comparisons between web accessibility and security. Accessibility measurements were collected between October 9 and October 13, 2026, using Chromium and Playwright with axe-core 4.14.0. A frozen, deterministic sampling protocol selected up to five pages per domain, covering homepage, functional, content, deep-path, and seeded-page categories. Automated checks targeted the WCAG 2.0, 2.1, and 2.2 Level A and AA rule sets. Results The acquisition pipeline produced 36,038 sampled-page records, of which 34,641 were successfully evaluated. Of the 10,022 sampled domains, 7,573 (75.6%) yielded evaluable accessibility results, while 2,449 (24.4%) could not be evaluated. Among the evaluable domains, 7,328 (96.8%) presented at least one automatically detected accessibility violation. The most widespread accessibility barriers included: Insufficient colour contrast, affecting 6,129 domains (80.9%). Links without accessible names, affecting 4,894 domains (64.6%). Insufficient interactive target sizes, affecting 3,080 domains (40.7%). Missing or inadequate image alternatives, affecting 3,016 domains (39.8%). Overall, the study recorded 106,815 violation-rule occurrences involving 890,680 DOM nodes. The proportion of evaluable domains with at least one detected violation exceeded 96% across all four Tranco popularity quartiles. A sensitivity analysis excluding 470 cross-domain redirects produced a similar prevalence estimate (96.7%). Interpretation and limitations The findings indicate that automatically detectable accessibility barriers are widespread across the evaluated sample, independently of website popularity. However, the results do not establish complete WCAG conformance or non-conformance. Automated accessibility testing cannot replace manual evaluation, keyboard interaction testing, assistive technology assessment, or comprehensive website audits. The 24.4% of domains without evaluable results represents a substantial source of potential selection bias. The Tranco sampling frame is also not statistically representative of the entire Italian web. The paper documents the acquisition methodology, sampling procedure, detected accessibility barriers, impact classifications, measurement failures, sensitivity analyses, and limitations. Data and reproducibility The accompanying IWAS dataset is publicly available on Zenodo: https://doi.org/10.5281/zenodo.23346985 The research pipeline, methodology, tests, and reproducibility documentation are available at: https://github.com/giovannimanetti11/italian-web-accessibility-study Project website and further reading Research project by PerseoDesign Research: https://perseodesign.com An Italian-language overview of the methodology, findings and limitations is available on the PerseoDesign blog: https://perseodesign.com/blog/studio-accessibilita-web-italiano-iwas-2026 Related research This study shares its original domain sampling frame with the Italian Web Security Study (IWSS): Paper: https://doi.org/10.5281/zenodo.21322437 Dataset: https://doi.org/10.5281/zenodo.21323346
Author: Manetti, Giovanni
Language: English
Subfield: Human Factors and Ergonomics
Field: Social Sciences
Domain: Social Sciences
SDG: Reduced inequalities
Open Access status: green
OpenAlex ID: W7221654499