Research Guides
Test a Research Filter With a Keyboard Before Calling It Usable
A research page may offer excellent evidence and still prevent some readers from reaching it. Before sharing an interactive comparison, test one complete task without a mouse. A button that looks clickable is not evidence that the workflow is usable.
Choose an outcome, not a control count
Use a harmless example: select a category, inspect its results, clear the selection and return to the full list. Work with fictional data or a disposable test account. Do not activate purchases, paid research or destructive controls during an accessibility check.
Record the page URL, date, browser, operating system, starting state and intended result. This turns a vague report such as “the filter is broken” into something a developer can reproduce.
Separate operation from visibility
W3C's keyboard guidance explains that functionality should be available through a keyboard interface, with a narrow exception for inherently path-dependent input. A comparable keyboard route can satisfy the requirement; counting focusable controls alone is not a complete test. W3C: Keyboard
Focus visibility is a separate question. W3C explains that users need a visible indication of the focused item so they know what their next keyboard action will affect. W3C: Focus Visible
For your worksheet, keep two columns: “Could I perform the action?” and “Could I tell where I was?” A successful filter change does not answer the second question.
Run the short task
- Begin at the page's normal entry point, without preselecting the filter with a mouse.
- Navigate to the relevant controls using the keyboard.
- Change one filter using the control's expected keyboard interaction.
- Inspect whether the displayed results match the selection.
- Clear the filter and confirm the original result set returns.
- Continue to the next useful action and record any lost position or unexpected jump.
This is an editorial test procedure, not a substitute for a complete accessibility assessment. Different controls have different interaction conventions; do not assume every widget should use the same keys.
Write the defect so someone can fix it
A useful fictional report might read:
On the comparison page, keyboard navigation reaches the category selector. Selecting “Analytics” updates the result count, but the focus indicator disappears. I cannot reliably identify the next active control. Expected: the workflow remains operable and the focused control remains identifiable. Mouse operation was tested separately and succeeded.
Attach the exact steps and starting state. Do not attach private customer records simply to illustrate a UI problem.
Recheck the whole task after the narrow repair
A focus-style correction should not change selected values, erase filters or submit requests. Repeat the full selection-and-reset task after the patch, not just the previously failing keystroke. Keep the result bounded: “This keyboard workflow passed in the recorded browser” is more accurate than “the entire site is accessible.”
For a research publisher, the benefit is practical: readers have a better chance of reaching and interpreting the evidence you want them to cite. This check does not guarantee backlinks, rankings or compliance.