The Current Status of Accessibility in Mobile Apps
Shunguo Yan, P. G. Ramachandran · 2019 · ACM Transactions on Accessible Computing · doi:10.1145/3300176
Summary
This large-scale empirical study evaluated the accessibility of 479 Android apps from Google Play across 23 business categories using IBM Mobile Accessibility Checker (MAC), an automated tool that maps accessibility rules to GUI widget categories. The study examined over 610,000 GUI elements across nearly 14,000 pages, categorizing issues as violations (V), potential violations requiring manual verification (PV), or warnings (W). The researchers proposed two new accessibility conformance metrics: Inaccessible Element Rate (IAER), which estimates the percentage of GUI elements that are inaccessible, and Accessibility Issue Rate (AIR), which calculates the percentage of actual accessibility issues relative to the maximum possible. The study also introduced coverage measures to estimate what percentage of accessibility issues automated tools can detect—finding that MAC identifies approximately 67% of issues, with 33% requiring manual evaluation.
Key findings
The results revealed pervasive accessibility failures: 94.8% of apps had violations, 97.5% had potential violations, and 66.4% had warnings. Just five widget types (TextView, ImageView, View, Button, ImageButton) create 92% of GUI elements and cause 89%, 78%, and 86% of violations, potential violations, and warnings respectively. Six specific issues dominate: lack of element focus (53.4% of violations), missing element description (19.6% of violations, 77.8% of potential violations), low text color contrast (24% of violations), insufficient spacing between elements (45.6% of warnings), and undersized text fonts and elements (31.5% and 17.4% of warnings). Together these account for 97%, 77.8%, and 94.5% of V, PV, and W respectively. Average IAER scores showed approximately 30% of GUI elements had accessibility issues, while AIR scores indicated 15% of issues remained unfixed. Critically, no significant relationship was found between app popularity (ratings or installs) and accessibility conformance—popular apps are just as likely to be inaccessible.
Relevance
This study provides actionable data for prioritizing mobile accessibility efforts. Training should focus on five widgets and six issues to address the vast majority of problems. The finding that app popularity doesn't correlate with accessibility underscores that market forces alone won't drive improvement—app stores should integrate automated accessibility checking into acceptance criteria. The proposed IAER and AIR metrics offer a practical way to track accessibility progress during continuous delivery, allowing teams to set quantitative improvement targets. For organizations, the 67% automated coverage figure helps estimate testing costs: roughly one-third of issues will require manual evaluation. The study's widget-to-rule mappings can inform IDE plugins and linting tools that flag issues during development rather than post-release. Future work should extend coverage to cognitive accessibility requirements not addressed by WCAG 2.0.
Tags: mobile accessibility · Android · automated testing · accessibility evaluation · GUI accessibility · empirical study
Standards referenced: WCAG 2.0 · Section 508