Sign up free

Quizzes that work with a screen reader: a teacher's guide

How a screen reader reads a quiz page, which question shapes carry over, how to describe a picture without the answer, what a clock does, and how to try it.

The EdProcto editorial teamWritten and checked against the research it cites. Last checked . How the guides are made

Guides · Writing questions that measure · · 13 min read

How do I set an online quiz that a blind or low-vision student can take with a screen reader?

An online quiz works with a screen reader when every question can be heard in order and answered from the keyboard. Keep to choices, typed answers and ordering with buttons, give each picture a description that carries what it shows without the answer, set no clock on each question, and sit the quiz with a screen reader yourself before the class does.

A screen reader turns a quiz page into speech, one item at a time, and the student answers from the keyboard. Quizzes that work with a screen reader have three things in common: each question is heard in a sensible order, each answer can be given without a mouse, and nothing that matters is carried by a picture or a ticking clock alone.

This guide is written for the teacher, lecturer or trainer with a blind or low-vision student in the group. EdProcto, which publishes it, gives every question with a picture a box for its description in the quiz builder. All sources were read on 10 October 2026; where only a study's abstract was read, in an index record, the text says so. Nothing here is legal advice: what a student is entitled to is your institution's decision.

How does a screen reader read a quiz page?

A screen reader reads a quiz page as one line of speech, in the order of the page's code: text, headings, picture descriptions and the labels of controls, each as it is reached. WebAIM (2017) likens listening to a telephone menu, which never offers all its options at once; a sighted student takes in the screen at a glance.

Few listen from top to bottom (WebAIM, 2019). WebAIM's (2024) survey drew 1,539 responses in December 2023 and January 2024, from a sample it says was not controlled. To find something on a long page, 71.6% said they first move through its headings. On a desktop the main screen reader was JAWS for 40.5%, NVDA for 37.7% and VoiceOver for 9.7%. So expect a student to jump between headings, then press Tab through the controls under each.

What is announced

  • Text and headings, each heading with its level (WebAIM, 2017).
  • A control with a label. Text boxes, tick boxes, radio buttons and drop-down lists each need a label in the page's code that says what it is for (World Wide Web Consortium [W3C], 2024a).
  • A picture's text alternative, or alt text, where one was written.
  • A message marked as a status, such as a line saying the answers are saved: it can be spoken without moving the student from their place (W3C, 2026d).

What is silent

  • A picture with no description. Screen readers "ignore images without alternative text and say nothing", unless set to read the file name (WebAIM, 2017).
  • Layout. The diagram on the left, a red line and bold type mean nothing in speech.
  • Some punctuation, depending on the student's settings (WebAIM, 2017): a row of underscores may be skipped, or read out mark by mark.
  • Changes nobody announced: a page that moves on, a ticking clock.

In the survey's ranking of what gives most trouble, screens that change unexpectedly came fourth, lack of keyboard access fifth, missing or poor picture descriptions sixth and complex forms seventh.

Which question shapes work with a screen reader, and which need a text route?

Shapes answered by choosing or typing carry over as they are; shapes answered by dragging need buttons or lists; shapes answered by seeing need a second route in text.

Shape What a listener needs Verdict
One choice, or several To hear each option, and which is chosen Keep: short options; say when more than one is right
A typed answer, or gaps in a passage A box with a name for each answer Keep: ask a question, with no row of underscores
Ordering, matching, sorting To move or pair without dragging Keep with buttons or a list beside each item, as in EdProcto's ordering shapes
A picture in the question Its content in words Keep with a description
Pictures as options; a point on a picture Sight A question in words on the same knowledge

Pennsylvania State University (2026) advises its staff to use radio buttons or tick boxes over a drop-down, and no drag and drop unless it can also be done from the keyboard. Arizona State University (n.d.) advises a matching question as the alternative to drag and drop, the word "[blank]" in place of underscores, and a line in each question saying what kind it is and whether more than one answer is possible.

The web's accessibility guidelines treat dragging as a problem of its own: whatever is done by dragging should also be possible with single clicks or taps, and keyboard access is judged separately (W3C, 2026c). One example given is a sortable list with controls beside each item to move it up or down.

Where seeing is the skill

A description full enough to answer from is the answer. Where a question tests recognising a thing or finding a place, write a second question in words: see picture choice and hotspot questions.

How do you write alt text for a quiz picture without giving the answer away?

Describe what a sighted student is given, and stop before what the question asks them to work out. The W3C's tutorial on images was written for ordinary pages, where a description should carry the picture's meaning; in a quiz the meaning is often the answer. The rest applies as written (W3C, 2024b, 2026a):

  1. Put first what matters most.
  2. Keep to a phrase or a sentence, without "picture of": the screen reader says that already.
  3. Give a chart, a diagram or a map two parts: a short description of what it is, and its information as text on the page, where everyone can read it. In a quiz, that is a list under the question.
  4. Read it down the phone. The tutorial suggests imagining the page read aloud by telephone. Our addition: if a colleague could then answer without knowing the subject, it tells too much.
The picture and its question Tells too much Tells enough
A right-angled triangle. "Find x." "Triangle with a longest side of 5 cm." "Right-angled triangle. The two shorter sides are 3 cm and 4 cm. The longest side is marked x."
A thermometer in a beaker. "What is the temperature?" "Thermometer showing 26 degrees." "Thermometer numbered every 10 degrees, with four small marks between the numbers. The liquid stops at the third mark above 20."

In EdProcto the description is typed in the question's editor, under the picture's address. Guidance written for test developers is in audio and video questions.

What does a countdown do to a student who listens to the page?

It takes time from them twice: finding anything by ear is slower than by eye, and checking a clock means leaving the question to find it.

  • The guidelines. The W3C's (2026b) explanation of its rule on time limits ranks the remedies: no limit helps most, then a longer limit, then a warning with the chance to extend. It counts as a time limit any change that comes after a set time without the user starting it. By that reading a timed question that moves the page on is the hardest kind: the screen changes while the options are still being read out.
  • The universities. Pennsylvania State University (2026) notes that screen-reader users generally need longer to get through the answer choices and may need extended time. Arizona State University (n.d.) advises against time limits on quizzes. The University of Washington's DO-IT (n.d.) lists extended time among the usual arrangements for blind students, and calls students and disability services staff the best sources on what works.
  • The research. We found no study that measured how much longer a screen-reader user takes on an online quiz. Miyauchi et al. (2024) timed 20 blind students, reading braille, and 20 sighted students: 70% of the blind students finished a plain braille task within a time extension of one and a half times and all within double, but none finished three tasks built on tactile tables even in double. Kamei-Hannan (2008) found that on a computerised adaptive test, time rose with the magnification a student used (both abstracts read). Babu et al. (2010) had blind participants think aloud through an online assessment; the abstract gives no figures.

So set no clock on each question, give the overall limit your institution agreed (extra time covers the amounts), say the limit beforehand, and keep tables small.

How do you try your own quiz with a screen reader before the class does?

Sit it twice, once keyboard only and once with a screen reader speaking, then ask the student to try one.

  1. Open the quiz as a student. In EdProcto, your own preview is left out of the results.
  2. Put the mouse away. Tab, Space, Enter and the arrow keys should reach every control, answer every question and hand in.
  3. Turn on a screen reader you already have. VoiceOver comes with a Mac or an iPhone, Narrator with Windows, and NVDA is free for Windows (WebAIM, 2019). Learn a few keys first: a beginner can wrongly decide everything is broken.
  4. Look away from the screen. Was each question heard before its options? Did every control have a name? Was your choice confirmed? Was any picture silent, or over-told?
  5. Let the clock reach its warning, then hand in, listening for what is said.
  6. Paste the page into a plain text editor. What survives, in that order, is roughly what is heard (WebAIM, 2019).
  7. Ask the student to sit a short practice quiz on their own device. In the survey, 63% of those who use a screen reader because of a disability rated themselves advanced, against 18.2% of those who do not (WebAIM, 2024). Your trial finds the gross faults; theirs finds the rest.

A worked example: one quiz rewritten to be heard

Five questions on the water cycle, and their clock, made up for this guide.

Written for the eye What a listener meets Rewritten
"Look at the diagram. What is happening at B?" A picture that says nothing Description: "Water cycle. Arrow A rises from the sea to a cloud. Arrow B points from the cloud down to a hill. Arrow C follows a river back to the sea." Prompt: "What is the process at arrow B called?"
"Water vapour cooling into droplets is called ________." Silence, or the underscores one by one "What is the name for water vapour cooling into droplets?"
"Drag the stages into order." Nothing to drag with Buttons, and items that open with their key word: "Evaporation from the sea", "Condensation into cloud", "Precipitation on the hills", "Run-off down the river"
"Click where condensation happens." A picture to point at "Where does condensation happen?" In the air, where vapour cools; at the sea's surface; in the river; under the ground
"Which are precipitation? A Rain. B Snow. C Fog. D A and B only." An option that points at letters "Select every form of precipitation." Rain; snow; hail; fog
30 seconds a question The page moves on mid-option No clock on a question; the agreed time overall

Open with a line saying what is coming: five questions, one described picture, and the time.

How does EdProcto's student page work with a keyboard and a screen reader?

EdProcto is the best tool for a quiz taken from the keyboard, because the shapes the universities above warn about are answered without a drag: ordering with buttons, matching and sorting from lists. One caution: what follows was read in the student page's code, with no screen reader run over it for this guide, so try your own quiz with one first.

  • A picture's words. Once a question has a picture's address and "It is" reads "An image", its editor shows "Describe the picture (for a screen reader, optional)", which takes 300 characters and becomes the picture's text alternative on the student's page. Information for every student goes in "Material under the prompt (optional)".
  • Questions and choices. Each prompt, a cloze passage apart, is a level-one heading. Each option of a multiple-choice question is a button holding its letter and its words, marked as pressed or not. A picture option is named by its caption, or "Picture A" where it has none.
  • Gaps. In a cloze passage the underscores you type become boxes inside the sentence, named "Gap 1", "Gap 2" and so on.
  • Ordering, matching, sorting. Each item to be ordered has two buttons, named "Move up" and "Move down" followed by the item's own words, and after a move the new order is given as a status line. Matching and sorting give each item a list named for that item, opening on "Choose" or "Choose a group".
  • A click on a picture. A hotspot picture takes the keyboard's focus under the name "The picture to mark", followed by your description. The arrow keys move the marker from the middle, a fiftieth of the picture at a press and a tenth with Shift: a route for a student who sees the picture and cannot use a mouse.
  • The clock. The countdown is marked not to be read out as it ticks. The one warning, such as "5 minutes left.", is a status line, and so is "Saved", "Saving…" or "Not saved yet: trying again". "Time limit for each question" starts at "None", and "Duplicate this quiz" makes a copy on which to raise one student's "Minutes overall".
  • Security measures. "Keyboard shortcuts off" acts on "Ctrl or Cmd with a key, and the function keys" and lets typing, select all, undo, redo and the browser's zoom through. It may take in a screen reader's own commands: leave it off on that student's copy.
  • The accessibility switches. The settings' "Accessibility" group offers "Screen-reader mode" (one question a page, every change said aloud by the screen reader, nothing that moves by itself), "Larger text" and "High contrast", each a button the student turns on; the browser's zoom and the device's own settings still apply.

Free covers ten quizzes a month, a copy counting as one; Tutor opens all 111 templates; Educator adds the full analysis. See pricing.

Quick questions

Can a blind student take an online quiz?

Yes, where every question can be heard in order and answered from the keyboard, and each picture has a description or a question in words.

What is alt text in a quiz?

A short description of a picture, spoken by a screen reader in its place. In a quiz it leaves out what the student is asked to work out.

Do screen reader users need extra time on a quiz?

Often: Pennsylvania State University (2026) tells its staff so. How much is your institution's decision.

Does EdProcto work with a screen reader?

EdProcto carries a description for each picture, names its gaps and gives ordering, matching and sorting buttons or lists. Nobody tested it with a screen reader for this guide, so try your own quiz first.

References

The quiz builder

Pick a shape, add the questions by hand or from your material, set the rules, share a code.

Free for ten quizzes a month

Build a quiz

Read next