A student gets a question wrong. What should happen next? The gold standard of school error analysis turns that mistake into a precise teaching decision instead of another red cross.
Ask the learner to reconstruct the step that produced the answer, test a plausible explanation for the error and choose a repair that addresses that explanation. Then check a different problem independently. Here, “gold standard” is a proposed teaching standard, not an official designation; the worked case and measures below are illustrative.
Fictional diagnostic example: Two learners write 1/3 + 1/4 = 2/7. One says fractions are added by adding both top and bottom numbers. The other says a common denominator is needed but cannot make equivalent fractions. For the first learner, compare the answer with 1/3: adding a positive 1/4 cannot make the result smaller. Then use equal-sized parts to show 4/12 + 3/12 = 7/12. For the second, practise converting each fraction while explaining why its value stays unchanged. The matching wrong answers led to different next teaching steps.
One Wrong Answer Can Have Several Causes
A student writes 1/3 + 1/4 = 2/7. The error may reveal a false fraction rule. Another student might produce the same answer after copying an example without understanding it. The response must investigate the reasoning, not assume every identical answer has the same cause.
Separate Slips From Misconceptions
An arithmetic slip, a missing prerequisite, an incorrect conceptual relationship and a representation problem call for different responses. Ask students to explain their steps before choosing an intervention.
Find the Earliest Broken Connection
If a learner struggles with fractional algebra, check whether the problem begins with equivalent fractions, operations, symbolic notation or equation structure. Repeating advanced exercises without repairing the dependency can waste time.
Use Counterexamples and Representations
A number line, bar model or concrete example can expose why a rule fails. The student should explain the correction in their own words rather than memorise a replacement answer.
Build a Useful Error Record
Record the task, observed reasoning, diagnosed cause, repair and next independent check. Avoid creating an enormous error notebook that students never revisit.
Test Whether the Repair Transfers
Give a new problem with different surface details and revisit the concept later. A corrected original worksheet is not sufficient proof of mastery.
The Break-Test
A student gets the right answer by applying an invalid rule that happens to work in one case. Does the teacher examine the reasoning? A gold-standard system detects false confidence before it compounds.
Connected Knowledge
Explore Misconception Diagnosis, Diagnostic Teaching and the MOE syllabuses.
For the assessment principle, the NSW Education Standards Authority’s Formative assessment guidance describes using evidence of misunderstandings to adapt teaching and determine next steps. Its Evidence and feedback guidance also connects feedback with action and application in new contexts. These are educational references from NSW, not Singapore assessment rules; the fraction tasks and record suggested here are original illustrations.
What to Measure
Use a small error record with an observable claim at each step: the learner’s original reasoning, a tentative cause, a task that tests that cause, the teaching response and a fresh independent check. “Careless” is not an adequate explanation unless the work and follow-up show that the underlying method is secure. Keep an “uncertain” category when the available evidence does not distinguish causes.
Track diagnostic evidence by counting reviewed errors that include a learner explanation or a targeted follow-up, out of all errors selected for review. Do not reward a high diagnosis count if the labels were guessed. When the evidence contradicts the initial interpretation, revise the interpretation and record why.
Track repair and transfer separately. In the fraction case, first ask for 1/2 + 1/5 without copying the model; the answer is 7/10, but the explanation of common units also matters. Later, ask the learner to reason through a differently worded problem requiring the same addition. Record accuracy, reasoning, prompts and the interval since teaching. Immediate success with a worked model does not establish later independence.
For a class review, report how many learners who received a particular repair demonstrated the target reasoning on the follow-up, with the denominator and missing checks shown. Also count recurrence of the same invalid rule at the later check. If only the rehearsed form improves, vary the representation or revisit the diagnosis. Do not treat these classroom observations alone as proof that one intervention caused improvement.
How to Reach the Gold Standard
Finish an error analysis with a teaching decision and a date or lesson for checking it. The teacher verifies the mathematics; the learner explains what changed and tests the revised reasoning on a new task. If the same error returns, reopen the explanation instead of asking for another copied correction. The useful outcome is a more dependable way of thinking that transfers beyond the marked question.