More than Meets the Eye: A Survey of Screen-Reader Browsing Strategies
Yevgen Borodin, Jeffrey P. Bigham, Glenn Dausch, I. V. Ramakrishnan · 2010 · Proceedings of the 2010 International Cross Disciplinary Conference on Web Accessibility (W4A) · doi:10.1145/1805986.1806005
Summary
This comprehensive survey catalogues the browsing strategies that experienced screen reader users develop to overcome the accessibility and usability barriers they encounter on the web. Drawing from multiple user studies, the authors document how blind users are far from passive consumers of content — they actively develop, discover, and share sophisticated coping strategies. The paper covers the full landscape of screen reader interaction: basic navigation techniques (arrow keys for sequential reading, shortcut keys like H for headings, Tab for focusable elements, element-type lists), speed optimization (increasing speech rate to 400-500 words per minute for congenitally blind users who expect key-press-to-speech delay under 50-100ms), and the use of Braille displays (40-80 cell resolution, particularly valued by deaf-blind users and for text editing). The core of the paper organizes strategies around specific challenges: navigating unfamiliar sites (heading navigation with H key, keyword search, reading backwards to find content before known landmarks like 'add to cart' buttons), task-based adaptation (Tab key for form filling, X for checkboxes, E for editable fields), form filling (guessing labels for unlabeled fields by position, e.g., the 3rd textbox is likely 'Street Address'), inferring roles of inaccessible controls (exploring nearby elements, following image-links to read target page titles, remembering ordinal positions), simulating mouse interaction for non-focusable custom widgets, and dealing with dynamic content changes including AJAX updates, automatic page refreshes, and form validation errors.
Key findings
The paper identifies a critical insight: if an effective browsing strategy exists to bypass an accessibility problem, that problem may appear less important than others that have no workaround — meaning accessibility surveys and automated tools may undercount the severity of problems that experienced users have learned to circumvent but that remain barriers for novices. Referencing the WebAIM survey of over 1,000 screen reader users, the paper reports that skip links are used: whenever available 22%, often 16%, sometimes 28%, seldom 19%, never 10% — with many users citing broken skip links and fear of missing content as reasons for avoidance. Heading navigation is more popular: whenever available 52%, often 24%, sometimes 14%. The survey of dynamic content strategies reveals that most users' primary strategy is simply avoiding dynamically changing pages altogether. When forced to deal with dynamic content (e.g., form validation errors that appear visually but are not announced), users must manually refresh the screen reader buffer and search for changes. Auto-refreshing pages (affecting nearly 50% of major news sites at 5-10 minute intervals) are particularly disruptive because screen readers restart from the beginning of the page. Users develop landmark-based strategies — searching for known text strings like 'Current Bid' on eBay to jump directly to dynamically changing values. For template-based sites, users learn to skip repeated template content using heading jumps (H key), paragraph navigation (P key), or the N key for non-linked content blocks, selecting strategies based on how the HTML is structured.
Relevance
This paper remains one of the most cited and practically useful references in screen reader accessibility research. Its value lies in shifting perspective from 'what barriers exist' to 'how do real users actually cope' — a distinction that has profound implications for both web developers and assistive technology designers. For developers, understanding that users navigate by headings (52% whenever available) makes a stronger case for proper heading hierarchy than any guideline citation. Knowing that users read pages backwards from known landmarks to find nearby content explains why consistent page structure matters more than pixel-perfect visual design. The form-filling strategies reveal how users compensate for missing labels — inferring field purpose from position and surrounding context — which both explains why some sites 'work' despite failing automated tests and highlights how fragile these workarounds are. For assistive technology developers, the paper provides a roadmap of where screen readers fall short: dynamic content handling, mouse simulation for custom widgets, and the lack of tools to help users identify page changes after refreshes. The observation that congenitally blind users can comprehend speech at 400-500 words per minute quantifies the speed advantage that experienced users have over novices and suggests that speech rate customization is not a luxury feature but a critical efficiency tool.
Tags: screen readers · blindness · browsing strategies · web navigation · usability · keyboard navigation · dynamic content · ARIA live regions · user research · assistive technology
Standards referenced: WCAG · WAI-ARIA · UAAG · ATAG · Section 508