When a student gets a Mathematics question wrong, “weak foundation” is only one possible diagnosis. The learner may be missing a prerequisite, but they may also understand the prerequisite and choose the wrong method, misread a condition, build the wrong representation, execute a valid method badly, fail to retrieve known material, or lose control under time. Different errors need different repairs.
This guide is about maths learning gaps diagnosis, maths diagnostic test, maths misconceptions and the distinction between a prerequisite problem and a method problem. Before sending a student back six months or six years, the aim is to locate the earliest unstable link that is actually carrying the current error.
The page sits beside BTT’s Lower-Floor Law of Mathematics. That article explains why advanced Mathematics rests on earlier floors. This article adds restraint: not every failure on an advanced question proves that an earlier floor is broken. Diagnosis has to distinguish dependency failure from representation, method selection, execution, retrieval and examination control.
Precision protects both struggling and strong learners. A struggling student can lose months if every error triggers a broad “go back to basics” programme. A strong student can become bored if secure prerequisites are repeatedly retaught. Good diagnosis seeks the smallest sufficient repair, then reconnects it to the current Mathematics.
1. Trace the reasoning chain before naming the problem
A Mathematics question can be read as a chain: read the conditions → identify objects and relationships → represent them → retrieve relevant knowledge → select a method → execute it → communicate the reasoning → check the result. An incorrect final answer tells us that something failed somewhere in the chain, not where.
The first diagnostic move is to locate the earliest point at which a valid route stopped being valid. If the representation is wrong, excellent algebra downstream cannot rescue the solution. If representation and method are sound but a negative sign is lost during expansion, the main repair should not begin with the whole topic. It should begin with that transformation.
This is why a marked script is more useful than a list of wrong chapters. “Trigonometry wrong” identifies a syllabus location. “Selected the correct ratio but rearranged the fraction incorrectly” identifies a mechanism. Mechanism statements can change teaching.
2. Seven error families
- Prerequisite: earlier knowledge required by the current task is unstable.
- Current concept: the learner has an incomplete or incorrect model of the idea now being taught.
- Representation: the situation is translated incorrectly into a diagram, equation, graph, table or symbolic form.
- Method selection: relevant methods exist in memory, but the learner chooses an unsuitable one.
- Execution: the route is valid but algebra, arithmetic, notation or calculator use breaks.
- Retrieval: knowledge exists but is not accessed reliably when needed.
- Control: timing, checking, attention or communication causes marks to be lost despite usable knowledge.
A single question can contain more than one family. The purpose of the categories is not perfect classification; it is to stop “weak at Maths” from replacing analysis.
3. What a genuine prerequisite failure looks like
A prerequisite is implicated when the current task depends on earlier knowledge that fails even after the higher-level surface is removed. If algebraic fractions are failing, strip away the variables and test the required fraction relationship with accessible numbers. If the same structural problem remains, the lower floor has evidence against it. If ordinary fractions are secure and the failure appears only when factors and variables are involved, the repair belongs higher.
Good prerequisite probes are short. If exponential equations fail, test the specific index law being used. If trigonometric manipulation collapses, test the algebraic transformation separately. If coordinate geometry breaks, probe gradient or equation rearrangement before reteaching the entire coordinate chapter.
The question is not “What did the student learn years ago?” It is “Which earlier capability is carrying this task right now, and does that capability fail when tested cleanly?”
4. What a method-selection problem looks like
Method-selection problems have a different signature. The student can perform several techniques when the method is named, but does not know which one belongs in mixed or unfamiliar work. This becomes increasingly important in Secondary, Additional and JC Mathematics because mature questions stop announcing the route.
To diagnose selection, temporarily remove execution. Show several problems and ask only: what structure is present, what method would you choose, and why? If the choice is poor but the student later executes the correct method accurately once told, the bottleneck is recognition or selection. Another page of routine execution may improve speed while leaving the real problem untouched.
The IES guide Teaching Strategies for Improving Algebra Knowledge in Middle and High School Students includes attention to algebraic structure and intentional choice among strategies. The wider principle is useful throughout Mathematics: possessing procedures and selecting among them are separate capabilities.
5. Representation can fail before method choice
Many apparent method errors begin earlier. A percentage question is modelled with the wrong base. A diagram is assumed to be to scale. A rate problem uses a total where a per-unit quantity is required. A word problem becomes the wrong equation. Once the representation is wrong, a student can execute the chosen method perfectly and still fail.
A clean representation probe stops before calculation. What quantities are present? What is known and unknown? What relationship connects them? Can the student draw, tabulate, graph or symbolise the situation? If execution becomes accurate as soon as the representation is supplied, the teaching target is the bridge into Mathematics.
This is why visual and connected representations recur in IES Mathematics guidance. Representation is not decoration. It is one of the languages through which structure becomes available for reasoning.
6. Execution errors need narrower repair
Execution errors happen after a valid representation and appropriate method have been chosen. They include arithmetic slips, lost negative signs, illegal algebraic transformations, notation problems, calculator-entry mistakes and incomplete working. They can be costly, but they do not automatically justify a broad foundation rebuild.
Freeze the upstream decisions. Confirm that the learner understands the problem and why the chosen method fits. Then isolate the exact transformation that fails. If expansion is unstable, practise expansion. If calculator brackets are the problem, repair input and estimation. If indices fail only with negative exponents, teach that boundary. Precision saves time.
7. Retrieval can imitate missing knowledge
A learner may possess knowledge but fail to retrieve it under pressure or after time. One useful sign is rapid recovery after a small cue. If a keyword, diagram or first step causes the whole structure to return, the knowledge may exist but lack a reliable retrieval route. Full reteaching could be disproportionate.
Ask how much cueing is required. Does the student need the formula or only the concept name? A worked example or simply thirty seconds to think? Can they retrieve again after a spaced delay? These observations distinguish absent knowledge from inaccessible knowledge.
Retrieval problems call for reactivation, spacing and mixed recall. The repair should make the pathway back to knowledge more reliable without pretending the learner never learned it.
8. Control problems sit above knowledge
Some errors are performance-control failures: poor pacing, skipped conditions, insufficient working, abandoned questions, unverified results or attention that degrades late in a paper. To diagnose control, remove pressure. If the same problem is solved accurately untimed, compare what changes under examination conditions.
The rule is simple: do not teach the syllabus again when the missing capability is paper control; do not train paper control when the Mathematics itself is absent. The assessment condition is part of the diagnosis.
9. Forty diagnostic situations: foundation, method or something else?
1. Algebraic fractions collapse, ordinary fractions are secure
Observation. The student handles numerical fractions accurately but cancels invalidly when variables and factors appear. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The arithmetic fraction floor is probably not the main bottleneck; factor structure is suspect. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Give paired expressions where cancellation is valid and invalid, first numeric and then algebraic. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. If discrimination improves without broad fraction reteaching, repair algebraic structure and return to the current chapter. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
2. Algebraic fractions collapse, ordinary fractions also fail
Observation. Common denominators and equivalent fractions are unreliable even with simple numbers. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The current algebraic error is likely being carried by an earlier fraction prerequisite. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Use accessible numerical fractions to test equivalence, multiplication and addition separately. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Repair the failing relation, then reconnect to variables quickly so remediation has a destination. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
3. Quadratic formula remembered, quadratic equation not recognised
Observation. The learner substitutes correctly when told to use the formula but chooses unrelated methods in mixed work. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Execution knowledge exists; structural recognition and method selection are weak. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Mix quadratic and non-quadratic equations and require classification before calculation. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Train recognition cues and boundaries instead of another block of formula substitution. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
4. Quadratic equation recognised, formula execution fails
Observation. The student identifies the quadratic structure but loses signs or coefficients during substitution. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Selection is sound; notation and execution are the main issue. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Use short coefficient-identification and bracket-control exercises. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Restore mixed questions after the narrow repair to ensure the student can still select independently. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
5. Trigonometry answer fails because the triangle is represented wrongly
Observation. The student chooses a plausible ratio from a mistaken identification of sides or angles. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The method error began as representation. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Require labels for opposite, adjacent and hypotenuse before any ratio is chosen, including rotated diagrams. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Repair spatial interpretation rather than adding more ratio memorisation. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
6. Trigonometry ratio correct, rearrangement fails
Observation. The learner selects sine, cosine or tangent appropriately but cannot isolate the unknown. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The trigonometric concept may be usable while algebraic rearrangement is unstable. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Replace trig symbols with simple variables and test equivalent rearrangements. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Repair the algebraic transformation and immediately reinsert it into trigonometry. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
7. Percentage increase uses the wrong base
Observation. Arithmetic is accurate but the percentage is taken from the wrong reference quantity. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The relationship and base are misunderstood rather than the percentage calculation itself. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Use simple before-and-after examples and require the sentence “percentage of what?” before calculation. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Train base identification across growth, discount and reverse-percentage contexts. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
8. Ratio procedure works, word problem fails
Observation. Ratios simplify and scale correctly when explicitly presented, but verbal situations are mistranslated. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Translation into ratio structure is the bottleneck. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Ask for a bar, table or unit relationship before arithmetic. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Practise representation across contexts rather than more isolated ratio computation. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
9. Linear equations fail only with negatives
Observation. Positive-coefficient equations are stable; negative terms trigger sign errors. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. A narrower signed-number or transformation issue may be carrying the algebra error. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Test negative-number operations and one-step equations separately. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Repair the exact sign behaviour before returning to multi-step equations. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
10. Linear equations fail with simple positive numbers too
Observation. The learner moves terms mechanically and cannot explain equality preservation. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The current concept of equation equivalence may be incomplete. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Use balance models and equivalent transformations with accessible numbers. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Rebuild equality meaning before increasing symbolic complexity. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
11. Gradient is calculated but not interpreted
Observation. The formula is used accurately, yet the student cannot explain rate or units. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Procedure is ahead of conceptual interpretation. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Compare graphs with the same gradient in different contexts and ask for verbal meaning. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Teach rate-of-change interpretation rather than repeating the formula. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
12. Gradient meaning is sound but arithmetic fails
Observation. The learner understands steepness and rate but makes subtraction or fraction errors. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Conceptual understanding is stronger than execution. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Use clean coordinate pairs to isolate subtraction and fraction calculation. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Repair arithmetic and return to contextual graph problems. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
13. Simultaneous equations work when the method is named
Observation. Elimination and substitution are both executable in topical exercises, but mixed questions produce poor choices. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The bottleneck is selection criteria. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Ask the student to compare equation structures and choose a method before solving. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Train strategic discrimination, not a third procedure. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
14. Substitution chosen correctly but change-of-subject fails
Observation. The learner sees that substitution is efficient but cannot rearrange an equation reliably. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The strategic choice is sound; a prerequisite transformation is unstable. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Test changing the subject outside the simultaneous-equation context. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Repair that transformation and reconnect without discarding the correct strategy. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
15. Geometry theorem is remembered but used on the wrong configuration
Observation. The learner quotes a correct theorem in a diagram where its conditions are absent. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Retrieval is stronger than condition recognition. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Use contrast diagrams where the theorem does and does not apply. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Teach condition checking and diagram reading rather than more theorem memorisation. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
16. Geometry theorem appears after one keyword
Observation. A tiny cue brings back an accurate theorem statement and valid application. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The knowledge may exist but retrieval is weak. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Use spaced retrieval with progressively reduced cues and mixed theorem sets. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Avoid full reteaching unless meaning or application also fails. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
17. Statistics calculation correct, interpretation weak
Observation. The learner computes a statistic correctly but draws an unsupported conclusion. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Procedure is secure while statistical reasoning is weak. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Ask interpretation questions that require no calculation and compare datasets with similar summaries. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Teach what the statistic can and cannot support. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
18. Statistics concept clear, calculator menu fails
Observation. The student explains the required measure but cannot obtain it reliably from the calculator. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Technology execution is the bottleneck. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Practise the calculator sequence alongside rough estimation and manual sense checks. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Do not reteach the statistical concept without evidence that it is also unstable. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
19. Differentiation rules known, function structure misread
Observation. Power, product and chain techniques are known individually, but composite expressions are misclassified. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Selection depends on structural parsing that is not yet reliable. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Give expressions to classify without differentiating. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Train seeing structure before adding more derivative computation. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
20. Derivative correct, later algebra fails
Observation. The calculus step is correct but simplification or solving critical points collapses. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Calculus understanding may be stronger than the algebra carrying later steps. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Isolate equivalent post-differentiation algebra. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Repair the downstream algebra while preserving the correct calculus reasoning. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
21. Integration fails only when preparation is required
Observation. Standard integrals are recalled, but the learner does not rearrange, split or simplify before integrating. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Strategic preparation is the bottleneck rather than missing integration facts. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Show integrands and ask only what should change before integration. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Train pre-method transformation and recognition. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
22. Complex-number work repeats ordinary algebra errors
Observation. Imaginary-unit rules are understood, but expansion and factorisation errors match those seen with real variables. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The advanced topic is being carried by an older symbolic prerequisite. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Test equivalent real-variable expressions. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Repair the algebraic floor and return promptly to complex numbers. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
23. Vector symbols work, geometric meaning is lost
Observation. Symbol manipulation is accurate but lines, directions and positions cannot be connected to it. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Symbolic procedure is ahead of representation. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Move repeatedly between diagrams, vector equations and verbal descriptions. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Build representational links instead of adding manipulation volume. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
24. Word problem abandoned before any mathematics appears
Observation. The student says “I don’t know” without identifying quantities or relationships. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The failure occurs before method selection and may involve language, representation or entry confidence. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Ask only for knowns, unknowns and one relationship. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Diagnose the entry process before assuming the underlying topic is absent. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
25. Word problem represented correctly but not solved
Observation. The right equation or diagram is built, then the learner stalls. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Representation is not the main bottleneck; downstream retrieval or execution is. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Remove context and test the mathematical object directly. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Repair the later link rather than reteaching question reading. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
26. Errors cluster late in long papers
Observation. Early work is accurate; signs, copied values and omissions rise after sustained effort. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Knowledge may be sound while attention or pacing degrades. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Compare a short timed set with the late-paper segment and inspect timing. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Train paper movement, checking windows and pacing rather than foundations. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
27. Errors occur from the first question even untimed
Observation. Signs, copied values and arithmetic slips appear immediately. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The issue is unlikely to be only end-of-paper fatigue. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Inspect fluency, visual tracking, working layout and checking one at a time. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Replace “careless” with the specific behaviour that can be trained. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
28. Student self-corrects as soon as told an answer is wrong
Observation. No explanation is needed; the learner finds the mistake independently. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Underlying knowledge may be present while error detection is weak. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Give an unmarked worked solution containing one planted error. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Train auditing and checking rather than reteaching the topic. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
29. Student cannot correct even when the wrong line is identified
Observation. The exact divergence is shown but the learner cannot explain or repair it. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. A concept, prerequisite or procedure is likely unstable. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Step backward through the dependency chain until independent reasoning returns. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Begin repair at that first unstable point. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
30. Homework strong, tests weak
Observation. Untimed supported work is successful; school assessments are unstable. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Prompt dependence, retrieval, transfer or performance control may be hidden by homework conditions. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Replicate test conditions gradually and observe which capability disappears. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Diagnose the condition change instead of relying on the phrase “exam panic”. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
31. Homework weak, tests comparatively strong
Observation. Practice is messy but formal assessment performance is better. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Effort, focus or checking conditions may differ sharply between settings. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Compare routines and evidence before assuming either sample is unrepresentative. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Use the discrepancy to improve practice quality rather than create unnecessary reteaching. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
32. Same invalid shortcut appears across topics
Observation. A behaviour such as cancelling additive terms or assuming proportionality reappears in several chapters. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. A structural misconception may be travelling through the curriculum. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Build contrast examples across topics showing when the shortcut works and fails. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Repair the general rule rather than each chapter separately. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
33. Many unrelated errors appear across every topic
Observation. There is no obvious single misconception, only broad instability. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Fragmented prerequisites, weak retrieval, overload or poor self-monitoring are possibilities. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Sample number, algebra, representation and problem solving with small clean probes. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Search for a common mechanism before compiling a giant list of weak chapters. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
34. Advanced learner uses a valid but inefficient method
Observation. The answer is correct but the route is long, fragile or difficult to generalise. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. This is strategy refinement, not foundation failure. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Ask for a second method and compare efficiency, robustness and generality. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Extend strategic flexibility without invalidating correct mathematical thinking. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
35. Worked examples are memorised
Observation. Near-identical questions are fluent; small surface changes cause collapse. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Apparent method knowledge may be pattern imitation rather than structural understanding. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Change representation, order and known or unknown quantities while preserving the idea. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Teach recognition of deep structure and fade template dependence. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
36. Concept is clear verbally but weak symbolically
Observation. The learner can explain the relationship in ordinary language but struggles with notation. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Meaning may be present while symbolic fluency is not. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Translate repeatedly between verbal, diagrammatic and symbolic forms with simple cases. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Build notation as a language layer rather than restarting the concept. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
37. Symbols move correctly but cannot be explained
Observation. Routine manipulation produces correct answers, but the learner cannot state why transformations are valid. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Execution is ahead of conceptual meaning and boundary knowledge. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Ask what expressions represent and when the rule would fail. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Add explanation and counterexamples so the procedure can transfer safely. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
38. Only combined-topic questions fail
Observation. Each component skill works in isolation, but integrated problems collapse. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. Coordination and recognition across topic boundaries are the bottleneck. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Pair isolated component questions with one integrated problem. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Train planning and connection rather than reteaching either component. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
39. Old syllabus gaps exist but current work succeeds
Observation. Historical checklists show content the student never mastered formally, yet current dependent tasks are stable. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. A curriculum gap is not automatically an active prerequisite gap. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Probe only earlier capabilities that are genuinely required by current or future work. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Repair active dependencies, not every historical omission. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
40. One bad paper triggers a proposed full foundation rebuild
Observation. A cluster of errors creates understandable concern and a very broad remediation plan. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The scale of the proposed solution exceeds the evidence. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Classify the script, test two likely prerequisites and compare with recent ordinary work. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Choose the smallest repair supported by the probes and widen only if evidence demands it. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
41. Student knows facts but cannot decide what matters
Observation. Definitions and formulas can be recalled, yet unfamiliar problems create paralysis. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The issue is not simple memory but relevance selection and problem representation. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Ask the learner to mark given information as essential, derived or irrelevant before solving. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Train information selection and planning. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
42. Student solves only when the diagram is drawn for them
Observation. Once a correct diagram appears, the rest of the solution is independent. Begin with what is visible. Do not translate the observation immediately into a global label such as “weak foundation” or “careless”. The final answer is a compressed output of several stages, and a useful diagnosis first reopens those stages.
Diagnostic hypothesis. The missing capability is representation construction. Treat this as a candidate explanation, not a verdict. Name the likely point in the reasoning chain—prerequisite, concept, representation, selection, execution, retrieval or control—so that the hypothesis can be tested and rejected if the evidence disagrees.
Clean probe. Give short verbal situations and score only the diagram or model, not the final answer. A good probe removes irrelevant complexity while preserving the suspected dependency. It should be short enough to interpret. If the learner succeeds when the suspected floor is isolated, move back upward; if the same structural failure remains, the lower-level repair has a stronger basis.
Repair decision. Make the bridge into Mathematics the explicit practice target. The discipline is economy: repair the earliest unstable link that is actually carrying the present failure, then reconnect it to the original task. A learner should not live permanently in remediation simply because an earlier gap can be found somewhere in their history.
10. The three-question prerequisite probe
When an earlier floor is suspected, three well-chosen questions can be more useful than a long test. First, ask a direct prerequisite question with accessible values. Second, place the same prerequisite inside a simple current-level context. Third, change the surface while preserving the relationship. The sequence shows whether knowledge is absent, available only in isolation, or transferable.
If algebraic fractions are failing, for example, use a numerical fraction operation, then a simple algebraic fraction, then a changed factorisation form. A student who fails the first needs a different repair from one who passes the first but fails when variables and factors arrive.
The probe is not intended to produce a score. It exists to choose the next act of teaching.
11. The method-selection probe
To test selection, temporarily remove calculation. Show several questions and ask the learner to identify the structure, nominate a method and justify the choice. A student who chooses well but later executes badly needs execution work. A student who chooses poorly but executes accurately when told needs recognition and strategy work.
This separation is especially useful in Additional Mathematics and JC Mathematics, where students can accumulate many techniques. The difficulty is often not possession of a method but knowing which method the structure invites.
12. Diagnostic language that helps
Replace “weak foundation” with “fraction equivalence is unstable and is causing algebraic-fraction errors”. Replace “careless” with “sign errors occur during bracket expansion and are not caught during checking”. Replace “doesn’t understand trigonometry” with “the ratio is selected correctly, but rearranging it is unreliable”.
Mechanism language makes the problem finite. It gives the student a route from unstable to stable and the tutor a way to retest the exact change. Identity language does neither.
13. Different diagnoses need different practice
Prerequisite problems need lower-floor repair and rapid reconnection. Concept problems need explanation, examples, counterexamples and multiple representations. Method-selection problems need mixed and contrasting tasks. Execution problems need focused transformation practice. Retrieval problems need spaced reactivation. Control problems need timed application, checking and paper-management routines.
“More practice” is therefore incomplete advice. Practice works when it rehearses the process that needs to change. Twenty more simultaneous-equation exercises will not necessarily fix a student who can execute both methods but cannot recognise a simultaneous-equation structure inside a word problem.
Precision also protects time. Students have other subjects, sleep, families and lives. The best repair is not the largest one; it is the smallest one that reliably changes the failure.
14. Three-student tuition and differential diagnosis
In a three-student Mathematics tutorial, three wrong answers to the same question can arise from three different places. One learner may misread a condition, another may choose a poor method, and a third may execute correctly but omit a required unit. The surface outcome is identical; the teaching should not be.
The value of a very small group is diagnostic resolution. The tutor can inspect working, ask one targeted question and run a short prerequisite or method-selection probe without converting the lesson into a long individual assessment. Peer solutions can also expose contrasting routes and common misconceptions.
The destination is still independence. Diagnosis exists to shorten the path to appropriate teaching, not to make the learner permanently dependent on an expert interpreter.
15. Research and evidence notes
The US What Works Clearinghouse guide Teaching Strategies for Improving Algebra Knowledge in Middle and High School Students recommends analysing solved problems, attending to algebraic structure and choosing among alternative strategies. These recommendations support the distinction on this page between owning a procedure and recognising when and why to use it.
The What Works Clearinghouse guides on assisting students struggling with Mathematics in the elementary grades and Response to Intervention for elementary and middle schools emphasise systematic instruction, mathematical language, representations, word-problem structures and progress monitoring. Their age ranges differ from much of BTT’s Secondary work, so the principles are used here carefully rather than presented as direct Singapore Secondary prescriptions.
The guide Improving Mathematical Problem Solving in Grades 4 Through 8 includes monitoring the problem-solving process and using visual representations. The broader diagnostic lesson is durable: inspect the process, not only the answer.
16. Frequently asked questions
How do I know if the problem is really foundational?
Remove the advanced surface and test the suspected prerequisite directly. If it fails cleanly, the lower floor is implicated. If it is secure in isolation but fails only in the current context, investigate representation, transfer, selection or control before prescribing a broad rebuild.
Should a weak student always return to basics?
Return only as far as active dependencies require. Some learners need substantial reconstruction; others need one narrow prerequisite. The evidence should set the depth.
What if the student knows the method but cannot recognise the question?
Treat that primarily as representation or selection. Use mixed tasks, contrasting examples and classification before calculation rather than more blocks of the named procedure.
Are careless mistakes foundation problems?
Sometimes, but often not. Replace “careless” with the exact behaviour—copied value, sign change, unit, calculator bracket, skipped condition, unchecked answer—and test whether the Mathematics underneath is secure.
When is a full diagnostic assessment useful?
When the learner has a long history of gaps, errors are broad and contradictory, or several current topics appear unstable. For one recurring error, a short targeted probe is often more efficient.
17. The shortest useful summary
A wrong answer does not reveal where learning failed. Trace the reasoning chain to the earliest invalid decision. Test suspected prerequisites separately. Separate method selection from execution. Change the representation when you need to test transfer. Remove time pressure when you need to distinguish knowledge from control.
Then repair only as far back as necessary and return to the current Mathematics. A foundation matters because it carries something above it. Diagnosis is complete only when the repaired capability works again in the task that exposed the weakness.
For dependency logic, continue to the Lower-Floor Law. For recurring errors, use the Fracture and Repair Map. For a broad learner-state route, use Find My Mathematics State.