EDUKATE SINGAPORE · GOLD STANDARD PRACTICAL CASE STUDIES · CASE 123
Mathematics is not accessible if its structure changes when read aloud.
A workbook passes a text-field accessibility check but an assistive reading route changes 1/(x + 1) into a different expression. This original fictional Gold Standard case follows the actual mathematical meaning from authoring to the student, with a worked substitution and practical correction.
All students, tools, test results and institutions are fictional. W3C and Section 508 resources provide accessible-document guidance, not certification of any real workbook or an assertion that one presentation method works for every learner.
Chapter index
- 1. The workbook opens, and the equation means something else
- 2. A fraction has structure, not merely its characters
- 3. Why a green accessibility score did not test this expression
- 4. W3C and Section 508 guidance explain the missing structure
- 5. The student needs a usable question before a grade
- 6. Correct the authored equation, not just the screen-reader setting
- 7. A student-friendly explanation and an accessible assessment are not identical
- 8. Test actual mathematical journeys, not just file properties
- 9. The quality report should say what was actually verified
- 10. Practice: the same symbols, two different meanings
1. The workbook opens, and the equation means something else
A fictional lower-secondary mathematics team publishes a digital workbook for thirty-two learners. Each page has selectable text, headings and descriptive alternatives for diagrams. The document production team’s automated checker reports no missing alternative-text fields in the elements it examines, and the school labels the resource ready for accessible independent study.
When learner Priya uses her usual assistive reading setup, an important fraction is announced as a flat sequence of symbols without a usable distinction between numerator and denominator. On the visible page the expression is one divided by the quantity x plus one. In the spoken rendition, the bracket relationship disappears and she hears an expression that can reasonably be interpreted as one divided by x, then plus one.
The equation has become a different mathematical problem during the handoff from the authored page to an alternative reading mode. The Gold Standard question is not whether the workbook contains all its words. It is whether the mathematical structure and intended learning task survive the presentation method the student is expected to use.
2. A fraction has structure, not merely its characters
The fictional worksheet asks learners to evaluate 1/(x + 1) when x = 3. The intended answer is 1/(3 + 1) = 1/4, or 0.25. If the reading system instead conveys (1/x) + 1, the evaluated result becomes (1/3) + 1 = 4/3, approximately 1.33. Those are different expressions with different meanings.
A text extraction that lists the same digits, symbols and letters does not prove the fraction was conveyed correctly. The location of a numerator above a denominator, the extent of a radical, an exponent, or the nesting of parentheses may be central to the mathematics. Removing those relationships can change the answer even while preserving each character.
The document’s visible mathematical typography is one representation. Its programmatically available structure and any spoken or navigable alternative are related but distinct. The test must verify that the alternative conveys a correct meaning to the intended user under the actual delivery conditions.
GS-R02 requires acceptance criteria that respect mathematical meaning, while GS-R08 asks what the checker actually established. Passing a missing-text field scan is not proof that a complex expression remains interpretable.
3. Why a green accessibility score did not test this expression
The production team used an automated review that inspected the presence of alternative labels and ordinary document tags. It did not traverse the mathematical notation using the combination of reader, format and device available to Priya. The tool’s scope was real but narrower than the school’s statement that the entire workbook was fully accessible.
Some formulae were exported as grouped drawing objects. Others were rendered as images with flat strings describing the symbols. A few were copied from a web editor whose visual layout survived while semantic markup did not. The centre has no consistent record of which route produced each equation.
The checker can accurately report that every drawing group has a nonempty label. It cannot then be treated as evidence that the label preserves exponent grouping or operator order. A complete accessibility evaluation needs appropriate automatic, manual and user-journey checks, with the scope stated honestly.
That is an evidence-design weakness rather than a claim that every automated accessibility tool is useless. Automation helps find problems within its capability; management must identify what it cannot verify.
4. W3C and Section 508 guidance explain the missing structure
The World Wide Web Consortium’s Understanding WCAG 2.2 Success Criterion 1.3.2: Meaningful Sequence explains that where presentation order affects meaning, at least one correct reading sequence should be programmatically determinable.
The W3C’s MathML Core specification describes mathematical markup that preserves elements such as fractions, including numerator and denominator structure. The US government’s Section 508 resource on accessible equations and formulas explains the role of mathematical structure, meaningful alternatives and MathML for accessible content.
These sources do not establish that every browser and assistive technology supports every form of mathematical notation identically, nor do they certify the fictional document. The W3C MathML Core source is a Candidate Recommendation snapshot rather than a claim of universal conformance. Actual authoring and testing must fit the document format and learner technologies used.
The Gold Standard’s operating question is whether the approved mathematics task retains one intelligible mathematical interpretation in the delivered student experience, and whether appropriate reviewers have evidence of that.
5. The student needs a usable question before a grade
Priya’s incorrect response might initially look like a misconception about fractions. If the workbook has presented the wrong expression, scoring the response as a knowledge deficit would misread the source of the problem. The tutor should distinguish the learner’s mathematical reasoning from a flawed delivery of the question.
The immediate teaching decision is to provide the correct expression through a suitably accessible, authorised route and ensure the student has a fair opportunity to attempt it under the centre’s real learning and assessment policies. That may involve clear verbal reading, a properly structured digital format, tactile mathematics or other relevant methods selected with competent staff and the learner.
The centre should not assume one alternate format works for every student who uses assistive technology. Preferences and needs differ. The authorised education team must confirm that the essential mathematical meaning is communicated, rather than merely switch to a new file type and declare victory.
GS-R09 addresses the exception. The responsible action protects the student’s learning task while the wider document is reviewed for similar formula-structure issues.
6. Correct the authored equation, not just the screen-reader setting
The document team traces the expression from its mathematical source through conversion, export and delivered file. It identifies whether the intended grouped denominator is represented in accessible structure or merely placed visually below a line. A source file containing a correct formula is not evidence that every downstream export retained that information.
The team may choose an appropriate semantic math format, a properly designed accessible alternative or a combination of representations suitable for the lesson and delivery platform. Technical decisions belong to qualified accessibility and publishing personnel. The important control is that every stage has a test for the meaning the next user must receive.
A change to a template, converter, export option or learning platform can affect many equations at once. A local correction to one visible page should therefore trigger a proportionate check of other expressions that used the same production path, not an assumption that the defect was unique.
GS-R07 protects the handoff from the mathematics author to the document producer to the learner. GS-R12 protects that handoff when tools or formats change.
7. A student-friendly explanation and an accessible assessment are not identical
The school also needs to consider the purpose of the workbook. A student practising fractions may need to navigate numerator and denominator independently. A learner studying algebra may need to examine nested structure step by step. A flat description of an entire page as ‘math problem’ cannot support that work.
An accessibility alternative should not accidentally give away an answer or change the difficulty relative to the intended task. For example, reading a fraction as ‘one over the quantity x plus one’ conveys grouping without solving the substitution. Describing it as ‘one divided by four’ before the student inserts x = 3 would alter the assessed step.
Consistency matters too. If a document uses ‘over’ for division in one item and omits grouping in another, learners may encounter unnecessary ambiguity. The maths tutor should review language and structure alongside technical staff so that the accessible version remains pedagogically valid.
The Gold Standard serves education when accessibility and subject accuracy are tested together, rather than being treated as unrelated checkboxes.
8. Test actual mathematical journeys, not just file properties
The next controlled test includes a simple fraction, an expression with a grouped denominator, an exponent and a word problem with labelled units. Reviewers follow the expressions in the actual student-facing environment using appropriate assistive methods and check whether the intended grouping can be understood and navigated.
The test deliberately includes 1/(x + 1), which must be distinguishable from (1/x) + 1. The reviewer evaluates how the expression is conveyed, not whether the alternative text merely contains the characters 1, x and +. The maths expert confirms the intended meaning, while accessibility expertise and relevant user experience check the presentation route.
A second test checks whether the learner can submit or communicate an answer through the actual workbook activity. Correct reading without an operable response channel may still leave the student unable to complete the planned task.
A passing rehearsal supports the tested combinations and expressions. It does not guarantee that every future equation, device or assistive technology is covered; meaningful changes require further review.
9. The quality report should say what was actually verified
The centre can report that its automated checker found no missing alternative-text fields within the elements inspected. It can also say that manual testing found a material fraction-reading defect in the learner’s delivered version. Those are compatible observations with different scopes.
After a corrected release, management can record the equations and user journeys tested, what was fixed, who approved the result and what remains outside the evidence. A global claim that all mathematics content is ‘100% accessible’ would require far stronger support than one successful test of a short worksheet.
Parents and students deserve practical information: which version to use, whether previously issued tasks are being corrected, and how to report an expression that does not make sense in the intended reading mode. The centre should not expose student information in public bug reports.
GS-R10 is meaningful only when the repair changes the actual learning experience. A polished audit badge cannot substitute for a problem that the learner can read correctly.
10. Practice: the same symbols, two different meanings
Exercise. A worksheet visibly asks for 1/(x + 1) at x = 3. A text-based reading route conveys (1/x) + 1 instead. The automated audit reports that all math images have nonempty labels. State the two resulting values, a bounded finding and one end-to-end acceptance test.
A reasoned answer is: ‘The intended expression gives 1/4, or 0.25. The alternative interpretation gives 4/3, approximately 1.33. The existing automated check supports label presence but does not demonstrate preservation of mathematical grouping. The document should not be described as accessible for that task until a suitable representation is reviewed and tested.’
The next test must use the actual learner-facing route to verify numerator, denominator and grouped expression boundaries without revealing the answer. The Gold Standard lesson is that accessibility succeeds only when the content retains its meaning for the person who needs to use it.
References and connected Gold Standard casework
- W3C WAI · WCAG Meaningful Sequence
- W3C · MathML Core, Fractions and Structure
- US Section508.gov · Accessible Equations and Formulas
- eduKate · Secondary 1 Mathematics Small-Group Tutorials
- Gold Standard · Information and Record Integrity
- Gold Standard · Core Requirements and Evidence
- Case 55 · Keyboard-Inaccessible Application
- Case 121 · Tactile Wayfinding Meaning
- Gold Standard Casebook
- Master Gold Standard Hub
The GS-R01–GS-R12 management framework is eduKate’s educational reference. Actual accessibility conformance must be evaluated for the relevant format, assistive technologies and users.