Accessibility of Internet Websites through Time
Stephanie Hackett, Bambang Parmanto, Xiaoming Zeng · 2004 · Proceedings of the 6th International ACM SIGACCESS Conference on Computers and Accessibility (Assets 04) · doi:10.1145/1028630.1028638
Summary
This paper presents a longitudinal analysis of web accessibility from 1997 to 2002 using archived snapshots from the Internet Archive's Wayback Machine. The researchers evaluated 240 archived instances of random websites (40 per year from the Alexa top 500) alongside 22 government websites, using the Web Accessibility Barrier (WAB) score — a metric developed at the University of Pittsburgh that provides a continuous quantitative measure based on 25 WCAG checkpoints weighted by priority level. The WAB score normalizes the number of actual violations against potential violations across all pages, producing a score where zero means no barriers and higher values indicate worse accessibility. The study also developed a complexity metric that assigns weighted values to HTML elements: tags receive a value of 1, script elements receive 10, and object/embedded elements receive 100, reflecting the increasing barriers these more complex technologies pose. The research was motivated by the observation that as the web evolved from simple text-based pages to rich multimedia experiences, the increasing design complexity might inadvertently increase accessibility barriers, particularly for users of assistive technologies like screen readers.
Key findings
The results showed a clear divergence between random websites and government websites over the six-year period. Random websites became progressively more inaccessible, with mean WAB scores increasing significantly from year to year (F(5,234)=7.246, p<0.0001). By 2001 and 2002, these websites had crossed the threshold score of 5.5 that separates accessible from inaccessible sites. Complexity of random websites also increased significantly each year (F(5,234)=16.52, p<0.0001), with the number of scripts rising dramatically from 697 in 1997 to 9,906 in 2002. A strong positive correlation was found between complexity and accessibility barriers (r=0.463, p<0.01), confirming that increasing web complexity contributes to decreasing accessibility. By contrast, government websites maintained relatively stable WAB scores near the accessible threshold throughout the entire period, even as their complexity also increased (r=0.14, p<0.041 — a weak correlation). Examining specific checkpoints revealed that the number of images used in websites increased enormously through the years, while the percentage of images with missing alternative text decreased from 63% in 2000 to 41.7% in 2002, suggesting growing awareness of at least some accessibility guidelines. Event handlers that require a mouse and non-descriptive link text were consistently among the most prevalent barriers.
Relevance
This study established an important empirical baseline for understanding how web accessibility evolves over time, demonstrating that technological advancement does not automatically improve accessibility — in fact, it can make things worse. For accessibility practitioners, the key finding is that government websites maintained accessibility despite increasing complexity, proving that accessibility and rich web experiences are not mutually exclusive when there is regulatory motivation (Section 508). The methodology of using the Wayback Machine for retrospective accessibility analysis was novel and has since been widely adopted in accessibility research. The WAB score's approach of weighting violations by WCAG priority and normalizing against potential violations addressed real limitations of simpler pass/fail evaluation methods. The study's main limitations are its reliance on automated evaluation (which cannot detect all accessibility issues) and the relatively small sample sizes. However, the clear statistical trends and the contrast between government and random websites provide compelling evidence for the importance of accessibility regulation and standards enforcement.
Tags: web accessibility · accessibility evaluation · longitudinal study · WCAG · Section 508 · web complexity · government websites · metrics · automated testing
Standards referenced: WCAG 1.0 · Section 508