Accessible online exams: WCAG basics for training providers
Accessible online exams are exams that candidates with disabilities can perceive, operate and complete without barriers that have nothing to do with the skill being tested. For a training provider, a practical technical yardstick is WCAG 2.2, the W3C’s standard for web content, applied to the exam pages candidates actually see. Accommodations such as extra time cover what a page alone cannot.
This guide is for training providers who deliver exams online and need a working grasp of WCAG and reasonable accommodations, not a legal brief. It maps the success criteria that matter most for a timed exam to specific exam features, gives you a test you can run in 15 minutes, and sets out a decision table for accommodation requests. Legal and standards facts are quoted from primary sources and dated as of October 2026. This is not legal advice, so have a qualified adviser read the legal section before you rely on it.
Key takeaways
- WCAG conformance belongs to the delivered web pages and the whole exam process, not to a question bank, so whoever builds or configures the learner-facing pages owns it.
- WCAG 2.2 requires a time limit to be turned off, adjusted or extended unless an exception applies, so build per-candidate time settings before exam day instead of relying on an exception.
- A 15-minute keyboard, zoom, screen reader, contrast and timer check on your real exam finds common barriers, but it is a screen and not a conformance audit.
- Accommodation requests need a written procedure: a prompt answer, documentation limited to what is reasonable, and a check that the change does not alter the skill being measured.
What does WCAG require, and who owns it for accessible online exams?
WCAG 2.2 is the W3C Recommendation that sets success criteria at levels A, AA and AAA for web content. It became a Recommendation on 5 October 2023, and the current text is dated 12 December 2024. Conformance is judged on web pages and complete processes, not on content in the abstract. An exam is therefore accessible only as delivered: the pages, the timer, the review screen and the submit step.
W3C’s WCAG 2.2 Recommendation and its conformance guidance are specific about scope. Conformance “cannot be achieved if part of a web page is excluded,” and for a multi-step process “all web pages in the process conform at the specified level or better.” For Level AA, the page “satisfies all the Level A and Level AA success criteria.” An exam is a process: sign-in, instructions, questions, review, submit and results. One inaccessible step undermines the whole claim.
This is why a question bank cannot be “WCAG conformant.” The bank holds items, but the pages that render them decide whether a screen reader user can answer. Item writers control alt text and wording. The delivered interface controls focus, labels, timing and contrast.
If you deliver exams on a headless, API-first platform such as SimpliAssess, check who builds and maintains the learner-facing pages. SimpliAssess’s FAQ describes a headless design as one that separates the data backend from the user interface. Its published features, FAQ and training-company pages do not describe WCAG conformance or accommodation settings, so ask any vendor, including this one, for dated evidence about the interface you will actually use.
Which WCAG 2.2 criteria matter most for a timed online exam?
Seventeen WCAG 2.2 success criteria map directly onto the parts of a timed online exam: text alternatives, structure, contrast, zoom, keyboard use, time limits, focus, dragging, errors, submission, sign-in and screen reader support. This is our selection, not a W3C list of exam criteria. The table below pairs each with the exam feature it affects and the check to run.
The table shows each criterion, its level, the W3C wording (quoted or closely summarized) and what to check in an exam. The last column is our application of the criterion, not W3C text.
| Success criterion (level) | What W3C’s text says | What to check in an exam |
|---|---|---|
| 1.1.1 Non-text Content (A) | “All non-text content that is presented to the user has a text alternative that serves the equivalent purpose.” If an item is a test that would be invalid in text, the alternative need only identify it descriptively. | Diagrams, charts and photos in question stems have alt text that describes without giving away the answer. |
| 1.3.1 Info and Relationships (A) | “Information, structure, and relationships conveyed through presentation can be programmatically determined.” | Each multiple-choice group is a fieldset with the question as its legend, and every option has a label. |
| 1.4.3 Contrast (Minimum) (AA) | A contrast ratio of at least 4.5:1 for text, and 3:1 for large-scale text. | Question text, option labels, the timer and button text, in every theme you offer. |
| 1.4.4 Resize Text and 1.4.10 Reflow (AA) | Text can be resized “up to 200 percent” without loss of content or functionality. Vertical-scrolling content works at a width of 320 CSS pixels without two-dimensional scrolling, which W3C equates to 400% zoom from a 1280 pixel viewport. | The timer, Next button, question navigator and Submit stay visible and usable when zoomed. |
| 2.1.1 Keyboard and 2.1.2 No Keyboard Trap (A) | “All functionality of the content is operable through a keyboard interface,” and focus can be moved away from a component using only the keyboard. | Answering, flagging, moving between questions and submitting, plus embedded media or code widgets that can trap focus. |
| 2.2.1 Timing Adjustable (A) | Each time limit can be turned off, adjusted over a range at least ten times the default, or extended after a warning, unless an exception applies. | The exam clock. The next section covers it in detail. |
| 2.4.3 Focus Order (A) and 2.4.7 Focus Visible (AA) | Focusable components receive focus “in an order that preserves meaning and operability,” and the keyboard focus indicator is visible. | Tab order runs stem, options, then navigation, and focus is visible after any timer warning appears. |
| 2.5.7 Dragging Movements (AA) | “All functionality that uses a dragging movement for operation can be achieved by a single pointer without dragging,” unless dragging is essential. | Drag-and-drop ordering and matching items have a click or tap alternative, such as move up and move down controls. |
| 3.3.1 Error Identification and 3.3.2 Labels or Instructions (A) | Detected errors are identified and “described to the user in text.” “Labels or instructions are provided when content requires user input.” | Unanswered-question warnings and fill-in-the-blank instructions, such as the expected format or units. |
| 3.3.4 Error Prevention (Legal, Financial, Data) (AA) | Applies to pages that “submit user test responses.” At least one must be true: submissions are reversible, input is checked, or a mechanism exists for “reviewing, confirming, and correcting information before finalizing the submission.” | A review screen and a confirm step before final submission. This criterion names test responses in its scope. |
| 3.3.8 Accessible Authentication (Minimum) (AA) | “A cognitive function test (such as remembering a password or solving a puzzle) is not required for any step” in an authentication process, unless the step offers another method, an assistive mechanism, object recognition or personal content. | Exam sign-in: math CAPTCHAs, puzzles and memorized codes need another way in, such as an emailed link, or support for pasting a password. |
| 4.1.2 Name, Role, Value (A) and 4.1.3 Status Messages (AA) | For interface components, “the name and role can be programmatically determined.” Status messages can be presented by assistive technology “without receiving focus.” | Custom answer controls and flag toggles have names, and notices such as “answer saved” or a time warning are announced. |
W3C’s forms tutorial on grouping controls says radio button groups should always be grouped using a fieldset. That is the multiple-choice question, the most common item in any exam. Its caution is worth heeding: depending on configuration, some screen readers read out the legend with every form element, once or, rarely, not at all. Keep the legend short and make each option label clear on its own.
Pro tip: Write Alt Text With the Item. Treat alt text as part of item writing, not a late upload. Describe what a candidate needs to answer, never the answer itself. If the picture is the skill being tested, such as reading a chart, build a second item version for text-based delivery and record which objective it covers.
How should an exam handle time limits and extra time?
Build the exam clock so the limit can be adjusted for each candidate before the exam starts, warn when time is nearly up, and decide whether speed is part of what you measure. WCAG 2.2.1 allows essential limits, but treating your own exam limit as essential should be a documented decision, not a default.
WCAG 2.2.1 says that for each time limit set by the content, at least one option must be true. The user can turn it off before meeting it. The user can adjust it over a wide range “that is at least ten times the length of the default setting.” Or the user is warned and given “at least 20 seconds to extend the time limit,” at least ten times. Exceptions cover real-time events, limits that are essential, and limits longer than 20 hours.
W3C’s Understanding document for 2.2.1 ranks the options: disabling time limits “is better than customizing the length of time limits,” which “is better than requesting more time before a time limit occurs.” It also notes that where timing is not an intrinsic requirement, “a third party can control the time limits for the user,” with the example “granting double time on a test.” That supports an administrator-set multiplier, but only where the limit is not itself what the exam measures. The same page illustrates the essential exception with a ticket-purchasing example and does not say that exams qualify.
So ask a design question first: does speed matter to the skill? W3C’s AAA criterion 2.2.3 gives the example of a test “designed so that time to complete the test does not affect the scoring.” If your objectives are about knowledge rather than speed, a generous limit or none is the simplest accessible design. A tight limit used as an integrity control has its own problems, which the post on reducing cheating without full proctoring rates as weak evidence with an accessibility cost.
When you keep a limit, build it this way:
- Set a per-candidate time multiplier before the exam starts, and show the candidate the final allowance in the instructions.
- Warn at fixed points, such as halfway and five minutes left, instead of announcing every second to a screen reader.
- Save answers continuously, so a dropped connection or a pause does not cost completed work.
- Decide in advance whether a rest break stops the clock, and write the rule down.
How do you test accessible online exams in 15 minutes?
Run five checks on the real delivered exam: keyboard only, 200 percent zoom, a screen reader spot check, a contrast check, and a timer and submit check. W3C describes its own easy checks as “designed to be quick and easy, rather than exhaustive,” so passing this screen does not prove conformance.
The zoom and visible-focus steps follow W3C’s easy checks. Use a practice exam with a timer, and log the date, browser and exam version.
Step 1: Keyboard only (3 minutes). Put the mouse aside. Using Tab, Shift+Tab, Enter, Space and arrow keys, answer one multiple-choice item, flag one, move on, open the review screen and reach Submit. You should always see where focus is (2.4.7), the order should make sense (2.4.3), and you should never get stuck (2.1.2).
Step 2: Zoom to 200 percent (3 minutes). Narrow the browser, then press Ctrl or Cmd with the plus key until you reach 200 percent. W3C’s check asks that all text stays readable, nothing hides behind other content, and horizontal scrolling is not needed. Then try 400 percent in a full-size window. The timer and navigation buttons must stay reachable.
Step 3: Screen reader spot check (4 minutes). Use NVDA, a free, open source screen reader from NV Access that runs on Windows, or the screen reader built into your operating system. Listen to one question. The stem and options should be announced together, the selected state should be spoken, the flag control should have a name, and “answer saved” or a time warning should be announced without moving focus.
Step 4: Contrast (2 minutes). Paste your text and background colors into a contrast checker such as WebAIM’s, which reports the 4.5:1 and 3:1 thresholds. Test question text, option labels, timer text and button text. Repeat for any dark or high-contrast theme you offer.
Step 5: Timer and submit (3 minutes). Start the exam with an extra-time setting and confirm the clock shows the right allowance. Let a warning fire and follow your extend or pause rule. Then submit with one question unanswered. The error should be described in text (3.3.1), and a review or confirm step should come before final submission (3.3.4).
Pro tip: Test Every Delivery Mode. Repeat the five checks in each mode you sell: practice, timed, review and, if you use one, proctored or lockdown mode. A page can pass in practice mode and fail once a security layer loads. The proctoring guide for training providers covers the accessibility duties that webcam rules can create.
Which accommodations should you offer, and who decides?
Offer accommodations through a written request process that answers promptly, asks only for documentation the need justifies, and tests whether the change alters what the exam measures. DOJ guidance lists extended time, braille or large-print exam booklets, screen reading technology, scribes and distraction-free rooms among its examples of testing accommodations.
The table below turns common requests into a decision aid. Extra time, screen reading technology, scribes, alternative formats and distraction-free rooms echo the DOJ testing accommodations guidance, which shows a last-updated date of February 28, 2020 and does not discuss online exams. Rest breaks, read-aloud and the webcam row are our additions, and all delivery and decision columns are our suggestions, not DOJ text.
| Request | How a training provider can deliver it | Decide by asking |
|---|---|---|
| Extra time | A per-candidate time multiplier set before the exam starts. | Does the candidate have documentation or a past accommodation for similar tests? Is speed part of the objective? |
| Rest breaks | A pause the candidate can use, with a written rule on whether the clock stops. | Can seen items be reviewed after a break? Does the break expose question content? |
| Screen reader or magnifier use | The candidate’s own assistive technology on accessible pages, tested in advance. | Have you run the 15-minute test in the exact mode this candidate will use? |
| Read-aloud or text-to-speech | Assistive technology on the candidate’s device, or a human reader. | Does the objective measure reading itself? If it does, reading aloud may change what the exam measures. |
| Speech-to-text or scribe | Dictation software, or a scribe who types what the candidate says. | Does the item score spelling or typing? If it does, dictation may change what is measured. |
| Large print, braille or other format | An alternative-format version of the exam, prepared in advance. | How much lead time do you need, and is the alternative item equivalent? |
| Distraction-free setting | A quiet room or a rule for the candidate’s own space. | Can your integrity controls still work in that setting? |
| Alternative to a webcam or lockdown rule | In-person delivery or another arrangement with comparable conditions. | Can you offer it as often and as promptly as the standard route? |
The same DOJ page says a testing entity must respond to requests in a timely manner, that documentation requests should be reasonable and limited to the need, and that proof of past accommodations in similar settings is generally sufficient. It also addresses flagging, which it describes as announcing to anyone receiving scores that the test-taker has a disability, so read it before you design certificates or transcripts.
Which accessibility laws might apply to a training provider?
Several laws may apply depending on where you and your clients are, and none of them makes an exam question bank “compliant.” The table below summarizes what each primary source says as of October 2026. It points your counsel to the right text. It does not decide whether any rule applies to you.
| Rule | What the primary source says | Question for your adviser |
|---|---|---|
| US: ADA Title III, 28 CFR 36.309 | Section 36.309(a) covers “any private entity that offers examinations or courses related to applications, licensing, certification, or credentialing for secondary or postsecondary education, professional, or trade purposes,” which must offer them “in a place and manner accessible to persons with disabilities or offer alternative accessible arrangements for such individuals.” The eCFR showed Title 28 as up to date as of 10/14/2026. | Do our exams or courses fall under this section? |
| US: DOJ web guidance | A page dated March 18, 2022 says the Department of Justice “does not have a regulation setting out detailed standards” for websites of businesses open to the public, and that “existing technical standards provide helpful guidance.” A separate rule published April 24, 2024 covers state and local governments. | Which technical standard will our clients and counsel expect us to meet? |
| US: Section 508 | The Section508.gov page (reviewed July 2026) says the law “applies to all federal agencies when they develop, procure, maintain, or use” electronic and information technology, and that the 2017 update harmonized requirements with WCAG 2.0. | If a federal agency buys our exams, what will it ask us to document? |
| UK: Equality Act 2010, sections 20 and 29 | Section 29(7) says a duty to make reasonable adjustments applies to a service-provider. Section 20 requires “such steps as it is reasonable to have to take” where a provision, criterion or practice puts a disabled person at a substantial disadvantage. Legislation.gov.uk showed section 29 up to date with changes in force on or before 17 October 2026. | What counts as a reasonable adjustment for our exam format? |
| EU: Directive (EU) 2019/882, the European Accessibility Act | Article 2(2) applies to listed services “provided to consumers after 28 June 2025,” including e-commerce services. Article 4(5) exempts microenterprises providing services. The Commission’s overview lists e-commerce but does not mention e-learning or exams. | Is our online exam sale an e-commerce service, and are we a microenterprise? |
| EU: EN 301 549 | ETSI published V4.1.1 in September 2026, and its foreword says clauses 9, 10 and 11 were updated “to align with the WCAG 2.2 recommendation.” The previous V3.2.1 (March 2021) was based on WCAG 2.1. Secondary sources (Deque, Davis Wright Tremaine, September 2026) report that the new version had not yet been cited in the EU Official Journal, which is what gives a harmonized standard its legal effect. | Which version of EN 301 549, and which WCAG version, does a given tender name? |
Two practical points follow. First, state the WCAG version and the date whenever you describe your own conformance, because tenders may still name EN 301 549 V3.2.1 and WCAG 2.1 while newer versions exist. Second, do not claim conformance from a quick check. Conformance covers every page in the exam process.
What should you do next?
This week, run the 15-minute test on one real exam in timed mode and save the results with the date, browser and exam version. One page of dated notes is more useful in a client conversation than a general claim of accessibility.
Then write a one-page accommodation procedure from the decision table, name who answers requests, and set your default time multipliers. Ask item writers to store alt text and any alternative-format item with each question in the item bank, so accessibility travels with the item.
Finally, ask whoever delivers your exams for a dated conformance report for the learner interface. The assessment platform RFP checklist shows how to ask for one and how to read the partial rows. Have a qualified adviser read the legal section for each place you deliver exams.
Frequently asked questions
Is WCAG required for online exams?
Not by name in every case. DOJ says it has no regulation setting detailed web standards and calls existing technical standards such as WCAG helpful guidance, while 28 CFR 36.309 requires covered private entities to offer accessible exams or alternative arrangements. Client contracts and tenders may name WCAG or EN 301 549. Check your jurisdiction and contracts with a qualified adviser.
What WCAG level should an online exam meet?
Aim for WCAG 2.2 Level AA. AA conformance means a page satisfies every Level A and Level AA success criterion, so it includes the basics such as keyboard access and text alternatives. State the version and date in any claim. EN 301 549 V4.1.1 (September 2026) aligns with WCAG 2.2, but the earlier V3.2.1 was based on WCAG 2.1, so a tender may name an older version.
Does a timed exam fail WCAG 2.2.1?
Not automatically. The criterion requires a time limit to be turned off, adjusted or extended unless it is a real-time event, essential, or longer than 20 hours. Whether an exam limit is essential is a judgment call. W3C notes a third party can control limits, such as granting double time on a test, so build adjustable per-candidate limits.
How much extra time should you give?
The sources checked give no universal figure. W3C’s example is double time, and DOJ guidance lists extended time, including double time on standardized exams, as an accommodation decided through a request and reasonable documentation. Pick a small set of multipliers in advance, for example 1.25, 1.5 and 2, and map each documented need to one of them instead of improvising per request.
Is a quick accessibility check enough to claim WCAG conformance?
No. W3C describes its easy checks as “designed to be quick and easy, rather than exhaustive.” Conformance is judged on whole pages and complete processes, so a claim needs an evaluation of every page in the exam flow, ideally with candidates who use assistive technology. Use the quick check to find obvious barriers early.
Can candidates use text-to-speech in an exam?
Often yes, but decide per objective. DOJ lists screen reading technology among testing accommodations, and 28 CFR 36.309 asks that results reflect the skill being measured, not the candidate’s impaired sensory, manual or speaking skills, except where those skills are what the exam measures. If the exam measures reading itself, read-aloud may change what is measured, so document that reasoning and offer a comparable alternative.
