Accessibility guru, heal thy site

Introduction

There are many experts that can help you make a website (or mobile app, PDF file, etc.) accessible—that is, inclusively functional.

Of course, you should expect those experts’ own websites to be paragons of accessibility. But guess again! Just as diversity and inclusion podcasts are often inaccessible, so are accessibility websites.

What I found

In July 2022 I fed the home pages of 98 experts in digital accessibility to procedures that automatically test and score web pages for accessibility. A perfect score would be 0.

In the table below:

• Each name in the Page column is a link to the page that was tested.
• Each number in the Score column is a link to a digest with a detailed report.

What do the data tell us?

The table above shows that none of the 98 pages got a perfect score of zero. Why?

One reason is that some of the tests deliver warnings about likely accessibility problems, so even a perfectly accessible page could get a non-zero score.

Another reason is that tests can be imperfect, so some reported faults may in fact not be harmful. Conversely, automated tests do not catch all faults, so these pages may have other accessibility problems not reported by any of the tests.

Moreover, there are disagreements on exactly what would make a web page perfectly accessible, and therefore on how to test and measure accessibility.

A high score does, however, justify concern and investigation. Most of the tests belong to widely used test packages.

Even the best-scoring page (Ab11y) fails some tests, because the page violates common accessibility norms. For example, the page has 2 main landmarks instead of only 1; its logo is wrongly coded as a heading; and its navigation links across the top are a list but are coded as if they were all phrases in a sentence.

The pattern here reminds me of George Bernard Shaw’s maxim, He who can, does. He who cannot, teaches. Many consultants claim they can make your website accessible, but you could reasonably ask them to demonstrate this competence on their own websites first.

Details of this work

The accessibility testing procedure was tp12a, part of Testaro. The scoring procedure was sp12a, part of Testilo.

The 98 digital-accessibility experts were identified mainly by means of lists published by the International Association of Accessibility Professionals and Raghavendra Satish Peri.

This report is a revision of an earlier one last revised in January 2022. The scores have changed, because:

• Some of the pages have changed, so the same procedures would have produced different scores.
• The set of test sources has grown. Tests other than my own now come from software packages developed by Deque, IBM, Siteimprove, Squiz Labs, Tenon, and WebAIM.
• Consequently, the number of tests has grown. The tp12 procedure runs 808 tests.
• Tests have been revised to correct bugs and improve validity, thanks in part to suggestions received.
• The scoring procedure has been revised to more accurately reflect the overlaps between test packages.

The digest for each page has been reorganized so that tests from various packages are grouped according to what they test.

Do you have a suggestion for further improvements? If it is for testing, please create an issue for Testaro. If it is for scoring, please create an issue for Testilo.