Accessibility statement

What this site aims for

The site aims to meet the Web Content Accessibility Guidelines (WCAG) 2.2 at level AA. WCAG is the international standard, published by the W3C, for making web content usable by people with disabilities. It has three levels, A, AA and AAA, and AA is the middle one. The W3C’s overview of WCAG explains the levels in plain words.

“Aims for” is the honest wording. It is a goal, not a certificate. This page does not say the site meets every success criterion in WCAG, because the checks below cover part of the standard, not all of it.

How it is checked

The checks are scripts that run against the finished pages, so they run the same way each time. The publish script runs them before anything is uploaded, in a real browser (Chromium, the engine behind Chrome and Edge). If one fails, the version is not published. One more test reads this page and holds it to the run: it fails if the automated checker, the tests with scripts off, the tests with reduced motion, the print checks or the Skip to content test stop running.

• An automated checker. axe-core, a widely used open-source accessibility checker, reads each page and reports problems such as text too faint to read, buttons with no name and broken labels. It runs on each page on a desktop-sized screen in light and in dark, and on a phone-sized screen in light. It also runs with the search box open and with the phone menu open. A serious or critical finding fails the run. Milder findings are listed in the test output.
• Keyboard tests. A script uses the keyboard on the phone menu (Tab to move in, Escape to close) and on the search box (Ctrl+K or / to open, the arrow keys to move, Enter to open a result, Escape to close and return focus). It does the same on the interactive pieces: the role picker on the home page, “Where does your school stand?”, “How an engagement runs”, the race, the workflow self-check, “You approve, then it runs”, the cards under “Try the method yourself”, the filter on the Common questions page and the glossary’s letter buttons.
• The first Tab stop. On each page, the first press of Tab lands on the “Skip to content” link, which shows an outline, and Enter on it moves you to the main content.
• A phone check. Each page is opened at 390 pixels wide, a common phone width, and fails if it scrolls sideways. The same check runs at 700 and 1024 pixels.
• Diagram labels. The text inside the diagrams is measured as drawn, at phone, tablet and laptop widths. A label drawn smaller than 10.5 pixels fails the test.
• Scripts switched off. Pages are opened with JavaScript turned off, to confirm that the content and links are still there and that no button is left on screen that cannot work. This covers the menu and the role paths on the home page, the Common questions page, the interactive pieces, the “Try the method yourself” cards, the glossary, the articles, the case studies and the error page.
• Reduced motion. Pages are opened with the device setting for reduced motion switched on. This is tested on the race, on “You approve, then it runs”, on the reading progress bar (checked on one article or case study) and on the glossary.
• The basics, page by page. One main heading, no skipped heading levels, a text description on each image, a name on each link and button, and no leftover template code showing.
• Light and dark. The automated checker runs in both. Text contrast is also measured directly in the glossary and in the search box on the error page, and the light and dark switch is tested to remember your choice.
• Print. Articles and case studies are checked as they print: the site’s header, the progress bar and the Copy link buttons are left out, the text is a serif, and each link shows its address. The glossary is checked too: its filter and buttons are left out and its sources show their addresses.

What you can rely on

• It works without JavaScript. The text and the links are in the page itself. Scripts add the light and dark switch, the search box, the filters and the interactive pieces. With scripts off, the menu is a plain list of links and each interactive piece shows a plain version.
• You can use it with a keyboard. The menu, the search box and the interactive pieces named above work without a mouse, and the first Tab stop is a “Skip to content” link.
• You can see where you are. An outline marks the control that has keyboard focus.
• Light and dark follow your device. The site matches the light or dark setting of your device. A switch in the header changes it, and the choice is kept in your own browser.
• Less motion if you ask for it. When your device asks for reduced motion, the race shows its result at once, the rows in “You approve, then it runs” do not fade, no reading progress bar is drawn and the glossary’s Copy link buttons do not fade.
• Articles and case studies have a print version. The site’s header, the progress bar and the Copy link buttons are left out, and links show where they go. The glossary prints without its filter and buttons.

What the checks cannot show

• Automated tools find only part of the problems. They can tell that a button has no name or that an image has no description. They cannot tell whether a description makes sense, whether a page is easy to follow, or whether a screen reader reads a piece in a helpful order.
• The interactive pieces are tested by script, not yet by a person using a screen reader.
• One browser engine is in the test run. The browser tests use Chromium. Firefox and Safari are not part of the test run.
• The phone view in dark is not part of the automated pass. The checker runs on a phone-sized screen in light only.
• Other websites are outside these checks. The links that leave this site go to pages that other people run.

Report a problem

If something on this site gets in your way, please tell me. Email hello@datazak.com or use the contact page. It helps to say which page it was, what you were trying to do, and what device, browser or assistive technology you were using.

Date of this statement

This statement is dated October 25, 2026. It describes the site and its tests as they stood on that day.

Last updated October 25, 2026.