Development technologies impact in web accessibility
Carlos Duarte, Inês Matos, João Vicente, Ana Salvado, Carlos M. Duarte, Luís Carriço · 2016 · Proceedings of the 13th International Web for All Conference (W4A) · doi:10.1145/2899475.2899498
Summary
This study investigates whether the choice of web development technologies—programming languages, web frameworks, JavaScript frameworks, and content management systems—has a measurable impact on the accessibility of resulting web pages. The researchers crawled 1,669 pages from distinct domains, used Wappalyzer to identify the technologies behind each page, and evaluated accessibility using QualWeb, an automated tool that assesses pages after browser processing (post-rendering) rather than from raw source code. This is significant because prior research showed substantial accessibility differences between server-delivered pages and those evaluated after browser processing—the latter being what users actually experience. The study applied three accessibility metrics based on WCAG 2.0 pass/fail/warning ratios: Conservative (warnings counted as failures), Optimistic (warnings counted as passes), and Strict (warnings excluded). Technologies were included only if present on at least 15 pages. Statistical analysis used non-parametric Wilcoxon tests for individual technology comparisons and Kruskal-Wallis tests with post-hoc comparisons for within-category analysis, with Bonferroni correction for multiple comparisons.
Key findings
Six technologies produced pages with statistically significantly better accessibility than the overall average (p<0.001): PHP, Twitter Bootstrap, jQuery, Modernizr, Drupal, and WordPress. Bootstrap showed the strongest positive impact, maintaining statistical significance even under the optimistic metric. Within-category comparisons revealed important differences. Among CMS platforms, Drupal (M=0.698) scored significantly higher than WordPress (M=0.636, p=0.005), which the authors attribute to Drupal's commitment to WCAG 2.0 and ATAG 2.0 conformance in its core from version 7 onwards. For web frameworks, ASP.NET (M=0.589) scored significantly lower than both Ruby on Rails (M=0.661, p=0.038) and Twitter Bootstrap (M=0.676, p<0.001)—Bootstrap's advantage likely stemming from built-in ARIA markup in components like modal dialogues. Among programming languages, Ruby (M=0.667) produced significantly more accessible pages than Java (M=0.562, p=0.024). No significant differences were found between JavaScript frameworks, though jQuery and Modernizr stood out individually. The authors speculate that programming language differences may reflect developer community culture rather than language features.
Relevance
This study provides rare empirical evidence that technology stack decisions—made early in project planning—have downstream consequences for web accessibility. For organizations and developers concerned about accessibility, the findings offer actionable guidance: choosing Drupal over other CMS options, Bootstrap over ASP.NET, or PHP/Ruby over Java correlates with better accessibility outcomes. The finding that Drupal's formal commitment to accessibility standards translates to measurably better real-world outcomes validates the approach of embedding accessibility into platform cores rather than relying on add-on plugins. However, the study has important limitations: it relies solely on automated evaluation (which cannot detect many accessibility issues), evaluates only the first page of each domain, and cannot determine causation—more accessibility-conscious developers may simply gravitate toward certain technologies. The sample sizes for some technologies were small (e.g., 16 for MooTools and Joomla). Despite these caveats, the study opens an important research direction and provides evidence that technology choices matter for accessibility.
Tags: web accessibility · automated testing · web development · content management systems · JavaScript frameworks · WCAG compliance · accessibility evaluation
Standards referenced: WCAG 2.0 · ATAG 2.0