GLOBAL KNOWLEDGE ARK

Learning Science · Practising retrieval to support learning

Checking later retention and application

Check whether knowledge remains accessible later and whether it supports a changed task.

Use the examples to connect this lesson with the earlier topics. No specialist equipment is needed.

Back to the chapter

After reading, you can

  • Distinguish retention from transfer.
  • Design a later check that separates remembered wording from usable understanding.
01

Two outcomes worth separating

Retention means that learned knowledge or skill remains available after an interval. Transfer means using learning in a different task or context. A learner may remember a formula yet fail to recognise when it applies, or solve a similar problem while forgetting the exact wording of a definition. These outcomes overlap but require different evidence.

An immediate repeat of a practice item mostly checks current performance under familiar cues. A later question with the source hidden provides evidence about durability. A changed case tests whether the learner can select or adapt the relationship. State the delay and the kind of change so that "learned it" does not hide several different claims.

Authored teaching comparison: separating achievements in the percentage example
CheckConditionWhat it samples
Immediate guidedWorked solution nearbyCurrent execution with support
Later relatedNew values; source hiddenRetention and familiar use
Changed wholeReverse the reductionSelection/adaptation under changed conditions
02

Change what matters to the goal

For a percentage lesson, retain the relationship but change numbers and context. Then include a case where the whole must be identified differently. For an argument lesson, present a new passage and ask for its evidence–claim relationship. Avoid introducing unrelated specialist knowledge that makes failure impossible to interpret.

Transfer can require new learning as well as retrieval. If a task changes the notation, environment or relevant assumptions, explain what preparation is available. A failure does not prove that all earlier learning was useless. Diagnose whether the problem is inaccessible knowledge, failed recognition, missing prerequisites or a genuinely new demand.

03

Keep claims proportional to the check

Research has found retrieval-related transfer benefits in specific prose-learning tasks. A later study showed that focused exposure to key information could account for a substantial part of a comparison with rereading. This makes the comparison condition important and cautions against attributing all transfer to retrieval alone.

Use results to choose the next activity. Lost details may need targeted review and another spaced retrieval. Failed method selection may need contrasting cases. An explanation that survives a familiar question but fails a changed assumption may need a clearer boundary. The check becomes useful when it identifies which part of learning to strengthen.

Worked example

Three checks of percentage learning

Why does a 15% reduction require about a 17.65% increase to reverse it?

  1. On the practice day, a learner calculates a 20% reduction of 50 with a worked example nearby. A week later, they calculate a 15% reduction of 80 without the example. This gives evidence about later use of a familiar relationship. A third task asks what percentage increase returns the reduced price to 80.
  2. The third question changes the whole: after a 15% reduction the value is 68, and returning to 80 requires 12 ÷ 68, about 17.65%. Success requires recognising the new base. The three results should be reported separately, because immediate guided execution, retention and adaptation are different achievements.

The reduction is 12, leaving 68. The reverse increase uses 68 as its whole, so 12/68 × 100 ≈ 17.65%. Report this adaptation separately from familiar guided execution.

The same amount is a different percentage

The same amount is a different percentageThe original 80-unit bar has width 480. The reduced 68-unit bar has width 408; the restored 12-unit segment has width 72. Losing 12 from 80 is 15%; restoring 12 to 68 is about 17.65%.Reversing a reduction changes the wholeOriginal: 80Reduced: 68Restore 12: 12/68 × 100 ≈ 17.65%The reduction used 80 as the whole.The reverse increase uses the smaller whole, 68.
The original 80-unit bar has width 480. The reduced 68-unit bar has width 408; the restored 12-unit segment has width 72. Losing 12 from 80 is 15%; restoring 12 to 68 is about 17.65%.

Try it yourself

Think through each question before opening the answer.

  1. Why does changing only the number sometimes provide limited transfer evidence?

    Show the answer and reasoning

    The same surface cues may still reveal the method.

    Include meaningful changes in selection or conditions when the goal requires flexible use.

  2. A learner recalls the formula but chooses it incorrectly. What should practice target?

    Show the answer and reasoning

    Practise recognising its conditions and contrasting cases requiring different approaches, rather than only repeating the formula.

    A method is usable only when its conditions can be recognised; further formula recitation leaves that choice supplied.

  3. What must be stated when reporting a successful retention check?

    Show the answer and reasoning

    State what was checked, after what interval and with which cues or tools.

    The conclusion should match those conditions.

What to carry forward

  • Check durability and changed use separately.
  • Interpret transfer in relation to the task, supports and comparison.

References

Sources supporting the explanations and learning checks in this lesson.

  1. The Testing Effect and Far Transfer: The Role of Exposure to Key Information (2016)Gerdien G. van Eersel, Peter P. J. L. Verkoeijen, Migle Povilenaite, Remy RikersResearch abstractpubmed.ncbi.nlm.nih.govRead source