Online Form Accessibility Checklist

Minimum manual check: Check the form as a sequence of understandable controls and recoverable states. Every control needs a persistent name, instructions must be available before they are needed, errors must identify a fix, and the whole task must work with a keyboard and enlarged content.

Published by Web-Forms.net on . Last editorial review: . Standards and operational claims are linked on this page.

Manual test evidence

Use this as an evidence log while operating the form. Record what happened, the affected control, and the repair that was verified.

  • Control name and group inspection
  • Keyboard path and focus order
  • Zoom and narrow viewport result
  • Empty, invalid, and corrected states
  • Success announcement
  • Tester, browser, assistive technology, and date

Before testing error recovery, verify the accessible form labels and instructions that give every control a stable name.

Inspect names, descriptions, and groups in the markup

A visible caption beside a field is not enough if it is not programmatically associated with the control. Use a label with a matching for and id pair, or an equivalent native relationship. Placeholder text is an example or hint, not a replacement for a persistent label.

It disappears during entry and can be difficult to review after an answer is present.

Group a set of radio buttons or checkboxes with fieldset and legend when the shared question gives each option meaning. Connect format help and constraints through visible text and an appropriate description relationship. Required status should be conveyed in text and markup, not only through color or an unexplained symbol.

Inspect the accessibility tree or use a screen reader to confirm the announced name and context.

Prefer native HTML controls before adding ARIA or custom interaction. Inspect the computed accessible name and description, not only the source label, because duplicate labels, hidden text, and description attributes can create a different announcement than expected. For each conditional control, confirm that its label remains unique when the new region appears.

Record the actual speech or accessibility-tree output for a small representative set so later changes can be compared against evidence.

  • Every input has one useful accessible name
  • Choice sets announce the shared question
  • Required and optional status is understandable
  • Hints do not repeat or contradict labels

Complete the task with keyboard and enlarged content

Start at the browser address bar and use Tab, Shift plus Tab, arrow keys, Space, and Enter. Focus must remain visible, follow a meaningful order, and reach every control and action. A custom choice control that only responds to a mouse is not complete.

Opening a calendar, select menu, help panel, or error link should not trap focus or move it without explanation.

Increase browser zoom to 200 percent and test a narrow viewport. Labels, help, errors, and buttons should remain available without content overlapping or requiring horizontal scrolling for the main task. Do not rely on a desktop arrangement to convey relationships.

Read the page without color and confirm that focus, errors, required fields, and selected states still have another indicator.

Repeat the keyboard test at each responsive breakpoint instead of assuming the desktop path scales down. At high zoom, check that sticky headers, cookie notices, chat controls, and validation banners do not cover the focused field. Make sure focus is never placed inside collapsed or visually hidden content.

Test with browser and operating-system text enlargement where available, because zoom, larger default text, and a narrow viewport can expose different clipping and reflow failures.

Trigger errors and verify the completion announcement

Submit the empty form and create each format error. The response should identify the affected field, describe what is wrong in plain language, and preserve valid entries. For several errors, place a summary near the start with links to the controls, then ensure each field has its own connected message.

Avoid blaming language and avoid examples that exclude valid international formats.

After a successful submission, move or announce status in a predictable way. The page should confirm what was received and the next step. If the same screen is reused, a screen reader user must not be left at a Submit button with no indication that the state changed.

Retest after visual redesigns, new conditional paths, validation changes, and third-party widget updates.

Verify the full spoken sequence for one failed submission: summary heading, number of errors, linked message, field label, entered value, constraint, and repair. After correction, confirm that stale error text and invalid state are removed. For success, test both a full page response and any dynamic status region.

A visible green panel is not sufficient evidence; the person must receive an understandable state change and retain a reliable route to the next action.

What an automated scan cannot confirm

A scanner may detect an input without a label, but it cannot reliably decide whether Account reference is understandable in this workflow, whether an error gives a useful repair, or whether focus moves predictably after a conditional section appears. Keep automated checks because they catch repeatable markup failures quickly, then add human task tests.

Save evidence as short notes: browser, zoom level, input method, page version, task, result, and defect. This makes a later retest possible. An unlabeled screenshot or a passing score without the tested path is weak evidence because the form may have changed or the difficult state may never have been reached.

What a checklist can and cannot prove

Can an automated checker prove a form is accessible?

No. Automated checks find some markup failures, but human testing is still needed for wording, keyboard flow, error recovery, zoom, and completion feedback.

What should be retested after a form changes?

Retest labels, choice groups, keyboard order, errors, focus, narrow layouts, and the success state. Conditional fields and third-party widgets need the same checks.

Accessibility standards behind the checks

Apply the checks to a complete form journey

Choose a realistic form and test the task from its first instruction through validation, submission, and confirmation.

Apply the checklist while operating the Web-Forms.net form builder at phone width and with a keyboard.

Apply the narrow-width checks in the online form builder preview before publishing an adapted template.