Making one’s website accessible for all, including persons who face obstacles in browsing the web, is a charitable thing to do. It is also a responsible thing to do, since nonprofit organizations, like businesses, can be held liable for disability discrimination if they do not. Therefore, it would be reasonable to expect charities to make their websites accessible.
But that expectation would be wrong. In fact, accessibility problems abound in websites of charities.
I tested the home pages of 116 mainly USA-based charitable foundations. I selected the charities from those working to advance various causes, chiefly justice, inclusiveness, civil liberties, ecological recovery, waste reduction, population sustainability, nutrition, disaster resiliance, peace, public safety, secularism, evidence-based policy, and knowledge access. They include the charities to which GiveWell awards its top rating for efficacy.
I used an automated testing procedure (version 7 of
a11y in Autotest). The procedure leverages over 400 tests in three widely used packages (Axe, Equal Access, and WAVE) and extends them with 16 custom tests of my own.
For the main page of each organization’s website, the procedure generated a score. A score of 0 would indicate that a page has passed all the tests. Any score greater than 0 indicates test failures.
I performed the tests in October and November 2021.
In the table below:
- Each name in the
Pagecolumn is a link to the page that was tested.
- Each number in the
Scorecolumn is a link to a detailed report.
Many of these charities have missions that would make them sympathetic with accommodations for diversity and inclusion.
Web accessibility mainly aims to ensure that people with various disabilities can successfully use web resources. But the scores of disability-oriented charities were widely dispersed, suggesting they have not given special attention to the accessibility of their own websites.
A similar result was found for international human rights organizations, whose home pages received scores ranging from about 450 to more than 3500.
Measures of accessibility vary, there is no consensus on the best or the correct way to measure it, and these tests, like all tests, are fallible. That is why this procedure runs many tests from several sources. Being fully automated, this procedure can be replicated and double-checked. Being open-source, the details behind the scores that it generates can be inspected and challenged.
More importantly, these organizations can study the individual reports about their web pages, in order to discover and dismantle website barriers, thereby making themselves more inclusive.