Learning Science · When difficulty helps and when it does not

Checking learning benefits and the costs of a challenge

Assess whether a challenging activity produces worthwhile learning for the time, effort and access it requires.

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

After reading, you can

Choose relevant benefit and cost measures.
Revise a challenge based on later evidence rather than effort alone.

Benefit needs an outcome

A new practice activity takes twice as long and produces more errors. These observations establish increased demand, but do not yet show a learning benefit. Ask what should improve: later recall, method selection, explanation quality, transfer or movement control. Use a task that actually measures that outcome after an appropriate interval.

Cost includes time, effort, need for help, frustration that prevents engagement, and obstacles to access. Costs are not always reasons to reject an activity; substantial learning can require work. They are reasons to compare whether the demand serves the goal and whether another arrangement could achieve the outcome more effectively.

Compare complete arrangements

A challenge includes instructions, preparation, practice, feedback and follow-up. If one group gets better feedback or more time, the outcome cannot be attributed to difficulty alone. For personal study, a perfectly controlled experiment may be impractical, but recording major differences still prevents confident overinterpretation.

Look beyond averages where possible. An activity may help prepared learners while leaving others unable to start. Inspect error patterns and support needs, then adjust entry tasks or guidance. A single label such as desirable difficulty should not conceal who could engage with the activity and what they actually learned.

Make the next decision concrete

If later performance improves at acceptable cost, retain the useful demand and test it in other relevant cases. If benefits are unclear, inspect whether the measure matched the goal and whether preparation was adequate. If both learning and engagement are poor, simplify or replace the activity. Persistence should be directed by evidence about the task.

Feedback research finds substantial variation across interventions; learning research likewise requires attention to conditions rather than universal prescriptions. Use that caution constructively. State a bounded conclusion, choose one practical change and plan another later check. A reasonable decision can be provisional without pretending that uncertainty means no action is possible.

Authored personal comparison; several changed features prevent a unique causal attribution.
Later evidence/cost	What it suggests	Adjustment to inspect
Better method selection	The selection demand may be useful	Retain some mixed cases
Calculation errors remain	Execution needs attention	Focused calculation work
Explanations consume much time	Cost may exceed need for familiar choices	Full reasons for uncertain choices; brief checks elsewhere

Worked example

Evaluating a mixed-practice revision

What should be retained and changed after the mixed-practice comparison?

A learner changes from ten same-type problems to a mixed set with explanations. Practice takes longer and initial accuracy falls. A later unlabelled set shows better method selection but continued calculation errors. The evidence suggests preserving some selection practice while adding a brief focused calculation component.
The learner also notes that the explanations consumed much of the session. They keep full reasons for uncertain choices and use brief checks for familiar ones, then compare later performance again. This adjustment responds to observed benefit and cost. It does not conclude that all mixing or all explanation must always be retained at the same intensity.

Retain useful method-selection practice, address calculation errors separately and limit full explanations to uncertain choices. Check later learning and time again.

Try it yourself

Think through each question before opening the answer.

Why is effort an insufficient benefit measure?
Show the answer and reasoning
It measures demand rather than learning.
The same effort can support useful reconstruction or unproductive confusion; later relevant performance distinguishes them.
What should be recorded when comparing two practice arrangements?
Show the answer and reasoning
Record content, time, support, feedback, sequence and later test conditions.
Major differences limit what can be attributed to the challenge itself.
A challenge improves selection but harms completion because calculation is weak. What is a reasonable change?
Show the answer and reasoning
Keep useful selection comparisons while adding targeted calculation support or practice.
Diagnose the components rather than accepting or rejecting the whole arrangement by its label.

What to carry forward

Measure later learning and practical costs.
Keep, revise or replace a challenge according to the pattern of evidence.

References

Sources supporting the explanations and learning checks in this lesson.