A large-scale web accessibility analysis considering technology adoption

Abstract

Abstract This paper reports the results of the automated accessibility evaluation of nearly three million web pages. The analysis of the evaluations allowed us to characterize the status of web accessibility. On average, we identified 30 errors per web page, and only a very small number of pages had no accessibility barriers identified. The more frequent problems found were inadequate text contrast and lack of accessible names. Additionally, we identified the technologies present in the websites evaluated, which allowed us to relate web technologies with the accessibility level, as measured by A3, an accessibility metric. Our findings show that most categories of web technologies impact the accessibility of web pages, but that even for those categories that show a negative impact, it is possible to select technologies that improve or do not impair the accessibility of the web content.