Charitable accessibility, 2023

Introduction

The website of a charitable organization is a major instrument for telling the public about the mission, ideas, activities, and achievements of the charity, for reaching potential beneficiaries of its services, and for soliciting donations.

Among the attributes of an effective website, one is accessibility. An accessible website adopts industry standards and practices that make it easy for visitors to navigate, find information, and take action. Accessibility aims to facilitate browsing for everybody. This includes even people who have trouble manipulating a mouse, using a keyboard, seeing, distinguishing colors, hearing, understanding technical terms, remembering, calculating, etc.

One might expect all owners of websites to adhere to these standards and practices. After all, who wants to impair interactions with the world? But everyday experience tells us that web dysfunctionality is the norm. The average home page of popular websites contains 50 accessibility errors, according to one analysis.

Public charities get tax exemption, and give their donors tax deductions, because they do good. Their missions may require them to make their websites inclusive and thus accessible. And there are pragmatic incentives: Inaccessible websites may interfere with donations and may expose their owners to claims of disability discrimination. Conversely, some donors may (as I do) treat an accessible website as evidence that the charity is both well-meaning and efficacious.

Findings

Suppose, then, we perform accessibility tests on websites of charities. What do we find?

I did that in 2021 and 2022. In both cases, the tests produced a wide range of accessibility scores.

I have done it again in 2023, using a more comprehensive battery of tests. Once again, there is a spectrum of scores. As before, a score is a measure of how poor the accessibility is, so 0 would be the best imaginable score. The 73 charities are, again, ones that I chose to test because of my interest in their work, so the sample is not at all representative. As before, the tests were performed only on the home, or landing, pages of the websites of the charities. And, again, the scoring procedure is one of many that might be adopted; different procedures would obviously produce different results.

The table below summarizes the results. Blind Institute of Technology obtained the best score, and Open Collective Foundation the worst. The worst score is 17 times as great as the best. It may seem natural that 3 disability-oriented charities got scores among the best 7, but Benetech, Mobility International, Humanity and Inclusion, and Partnership for Inclusive Disaster Strategies are also disability-oriented, and they landed in the middle or lower. You can see the pages that were tested by following the links in the first column. You can see digests of the testing reports by following the links in the second column. And within each digest is a link to the complete report that the digest summarizes.

Diagnosis

Why did no charity get a perfect 0 score?

A score of 0 is possible, but unlikely. The procedure makes use of about 900 tests created by 9 different organizations, representing diverse notions of accessibility. Some of the tests explicitly produce warnings of suspected but not definitive errors. And tests defined by humans can be faulty, just as humans themselves can be. So no web page should be condemned as inaccessible for getting a score greater than 0. For comparison, the same battery of tests yielded a score of 172 on the page that contains the table of contents of this blog, 659 on this page, and 491 on the widely used example page example.com.

What can a charity do to improve its website accessibility?

The reports give advice about this.

Consider Project Gutenberg, which got the 3rd-best score. Some of the tests complained that the top heading, Welcome to Project Gutenberg, is not inside any landmark—an invisible piece of the code that identifies a part of the structure of the page. The standard practice is to include a main landmark in every page, enclosing all the content except site-wide headers and footers. The Gutenberg page has no main landmark. Landmarks, although invisible, are used by software that helps users understand and navigate the web. For example, screen readers use landmarks to interpret page organization for users (including blind users and those who have trouble reading text) by speech.

A frequent complaint about the Solidarity Collective page was poor text contrast. That means text superimposed on a background that is not very distinct from the text itself, or that varies in color or darkness. As shown below, the page contains a white text that is partly easy and partly hard to read, depending on whether the portion of the background photograph behind each letter happens to be black, gray, or off-white. Even if the text is legible, reading it becomes more difficult when its background is very nonuniform. Elsewhere on the same page some text has been placed over a solid contrasting background, with the photograph farther behind. That eliminates the problem, and the same treatment could cure this defect.

The home page of Sanitation and Water for All was faulted for many issues, including content that moves forever without the visitor asking it to. Some visitors with motion sensitivity can become disoriented or even suffer seizures from repetitive motion. Others, who need more time than usual or get distracted, have content yanked from them while they are trying to finish reading it. The page rotates a section of content every 3.5 seconds, giving visitors only that much time to read and understand texts such as H.E. Dr Han Seung Soo, former Prime Minister of Korea and SWA Global Leader, on the payoffs of investing in water, sanitation and hygiene. Accessibility benefits when visitors are given control: when they are permitted to decide which sections of content to read and when, instead of having content fed to them on a fixed schedule. On an accessible page, any motion is also stoppable by the visitor with a keyboard action, but the page does not give the visitor a way to do that.

Conclusion

I gave 3 examples above of the hundreds of Web accessibility issues documented by the tests. Why so many?

Web developers have often learned their craft when web accessibility was an esoteric topic and perceived as a frill. By now it has become a mainstream aspect of code quality and risky to ignore. Today it is realistic for any charity to demand that its website creators adhere to all generally accepted accessibility practices. Moreover, tools for automated accessibility testing have proliferated and become more powerful and easier to use. Everybody can use the same tools I used, which are documented in the reports. And, for the most powerful testing, web developers should combine multiple tools.

Human testers, including some with various disabilities, should ideally try to make realistic use of a website and document accessibility issues that they find, since automated tests do not exist for some issues. But human testing is much more expensive and takes longer to perform. Consequently, it is rational to run the available automated tests first and correct any defects they discover, before human testing begins. That will leave fewer defects for human discovery and make the human testing less expensive.

The famed 80/20 rule holds for accessibility defects. The report on your page likely shows that a few issues account for most of the score. That typically means that there are many instances of each of those issues on the page (and elsewhere on the site). Sometimes a single setting (such as a style definition) can be revised, and then all the instances will be instantly repaired.

Thus, the long list of issues in a report may seem discouraging, but it contains guidance on beginning where you can achieve the greatest impact. Ideally, each repair will also serve to establish an accessible design and implementation pattern, so future website revisions will maintain high code quality, including accessibility.