Understanding Versus Independent Practice

Separate five different kinds of learning evidence

A correct answer can show recognition, guided performance, explanation, transfer or independent work, depending on the task and the support still present. These observations are related, but they are not interchangeable. A learner may recognise a completed model without being able to construct one. A guided attempt may be accurate because the prompt names the key decision. An explanation may sound clear in the familiar representation but fail to transfer to a changed whole or context.

This five-part framework is original synthetic editorial guidance. It is not an official assessment scale, a diagnosis or a claim that learners move through the stages in one direction. Support can increase when the representation changes. A learner who worked independently yesterday may need a prompt when new vocabulary, notation or conditions are introduced. Independent work is evidence for that task at that time, not a guaranteed permanent state.

Current public review copy describes an intended teaching approach that makes reasoning visible, checks understanding, adapts representation and gradually reduces prompts. It is not proof of a live teaching workflow or reporting backend. The current site does not guarantee tutor availability, response times, grades or outcomes. The categories in this guide help a parent ask what support was present; they do not promise when support will no longer be needed.

Follow one original support-fade example

Consider an original fraction-comparison sequence. It is not a real learner record or lesson report. At recognition, the learner identifies which of two completed models shows three quarters while both labels remain visible. That choice is useful evidence, but the completed models carry much of the structure. Ask what feature proves the choice. If the response depends only on remembering the appearance, the relationship between equal parts, numerator and denominator may still need attention.

The sequence is deliberately specific about support. “No worked example” is different from “no support”: familiar notation, task instructions and normal checking time still shape performance. The changed set model tests whether the meaning survives a change of representation. A later recheck asks whether the strategy can be recalled rather than copied from the immediately preceding attempt. None of these checks sets a deadline for independence.

Notice what one correct answer cannot show

One correct response does not reveal which decision was difficult, what prompt was used or whether the same relationship will be recognised elsewhere. Ask how the learner began, what they checked and whether they could explain the choice. If a tutor supplied the first step, record that support. If the learner corrected an error after a question, the correction is evidence of responsiveness but not the same as an unprompted check.

Before each attempt, I would note the support still present: worked example, verbal prompt, visual cue or normal checking time. Remove only one meaningful support or change one meaningful task feature at once; if performance breaks down, restore the smallest useful support and record exactly what was restored. This makes the change in conditions visible without treating support fade as an all-or-nothing test.

• Recognition: can the learner identify the relevant relationship in completed work?
• Guided performance: which prompts or models make the next decision possible?
• Explanation: can the learner state why the choice fits the relationship?
• Transfer: does the reasoning survive a changed representation or context?
• Independent work: can the learner select, complete and check the method without a prompt?

These questions also prevent an unfair conclusion from difficult transfer work. A learner may appear less fluent because the task now tests a deeper relationship. That is not necessarily lost progress. It may show exactly where support should return. Conversely, repeated success on nearly identical questions may show useful accuracy without yet showing transfer. Describe the evidence rather than ranking the learner by a stage label.

Keep the language equally precise when the evidence is incomplete. “Needed a reminder to compare equal parts” is more useful than “does not understand fractions.” “Explained the model but has not yet tried a fresh set example” leaves the next question open. This wording records the observed boundary without turning one lesson into a forecast. It also gives a future tutor or parent a concrete place to resume the check.

Use a fresh transfer check before calling work independent

A useful transfer check changes one meaningful feature while preserving the relationship. The fraction example moves from an area model to a set of counters. A comprehension check might change the character and setting while retaining the same evidence–inference relationship. A Mathematics check might change the whole or the unknown. Changing everything at once makes it difficult to identify what caused the breakdown.

If the fresh task is not completed independently, return to the smallest useful support: a question, visual cue or comparison rather than the whole worked solution. Then check again. Support fade is not a test of willpower. It is a way to discover which parts of the reasoning are available and which still depend on the teaching environment.

Discuss progress without promising a deadline

A parent update can name the task, the support present, the observed reasoning, the changed example and the next check. That is more informative than saying a learner “understands” without conditions. The current review-copy teaching standards provide internal context for this approach; they are not evidence that a reporting service is live. The current local-only form does not transmit a learning record and should not receive private school documents.

There is no fixed number of lessons after which independent work should be expected, and no lesson count guarantees durable independence. The review site does not guarantee a tutor, response time, grade or outcome. Homework and guided practice can both matter; the framework does not rank one task format universally. Its purpose is narrower: to help families distinguish recognition, guided work, explanation, transfer and independent practice so the next conversation is based on observable evidence.