When students can copy a model answer but struggle to adapt it

A student can reproduce every line of your break-even calculation, substitute the numbers correctly and arrive at the expected answer. Then the assessment asks for the sales volume needed to achieve a target profit, and the procedure suddenly falls apart.

This is a familiar gap between reproduction and transfer. Transfer occurs when learners apply what they have learned to a meaningfully different problem or context, rather than simply repeating the task they practised. It is also where a model answer can reveal its limitations. Showing students how to complete one problem does not necessarily help them recognise the underlying structure when the next problem looks different.

An integrative review in Educational Psychology Review examines how worked examples might be designed to support this more adaptable use of knowledge. The authors screened 2,644 records and included 85 empirical reports representing 127 studies or experiments. Their qualitative synthesis covers several disciplines and educational levels, although mathematics and science dominate the evidence base.

Accounting-specific evidence is limited. The review identifies one accounting study, by Sithole and colleagues, in which worked examples included prompts that encouraged students to monitor and manage where they directed their attention. Students using these prompts showed greater transfer gains than students in the study's comparison conditions. This is a narrow finding. It supports guided self-management of attention, not subgoal labelling, fading or the wider framework in accounting courses. The review is therefore best treated as a source of design ideas to test, not as a settled formula for accounting teaching.

Make the structure visible

A conventional worked example normally presents a problem and a complete sequence of solution steps. This can be particularly helpful for novices because it reduces the amount of unproductive searching they must do while still learning the procedure.

The difficulty is that students may remember the sequence without understanding its organization. The review identifies subgoal labelling as one promising response. A subgoal is the purpose served by a group of steps. Labels help learners see a solution as a series of meaningful objectives rather than a collection of calculations.

Consider a simple management accounting example. Assume a single-product business over one year, with volume measured as units sold. Within the relevant range, selling price and variable cost per unit remain constant, total fixed operating costs remain $60,000, and costs can be separated into fixed and variable components. A product sells for $50 per unit and variable cost is $30 per unit. The annual break-even sales volume is 3,000 units. Instead of presenting only the calculation, the solution could be organized like this:

1. Find the contribution generated by each unit

$50 selling price less $30 variable cost equals $20 contribution per unit.

2. Identify the total amount that contribution must recover

At break-even, contribution must recover fixed costs of $60,000.

3. Express the required contribution as a number of units

$60,000 divided by $20 equals 3,000 units.

The labels expose a structure that can survive changes in the question. They also give students language for monitoring their own reasoning. If a later question includes a target profit, students can ask whether the amount to be recovered is still fixed costs alone.

Research on subgoal labelling has produced recurring positive findings, but implementation matters. Labels must be meaningful to learners. Highly abstract terminology may support generalization in principle, yet it can be opaque to novices who do not already understand the structure it describes.

For an introductory class, this suggests beginning with accessible labels such as “identify the amount to recover”. A more advanced class might be ready for concise conceptual language. The important point is not to decorate every line with a heading. It is to show where one intellectual purpose ends and the next begins.

Ask for explanation, but provide somewhere to go

Another possibility is to ask students to explain selected parts of the reasoning themselves, a technique researchers call self-explanation. Useful prompts might include:

• Why is contribution per unit used here rather than selling price?
• What changes if the business wants a profit instead of merely breaking even?
• Which part of the method stays constant when the numbers or context change?

The review finds considerable support for self-explanation across several domains, including evidence of improved performance on changed problems. Yet the quality, timing and level of guidance matter. A prompt alone does not ensure a useful explanation. Novices may produce a superficial or incorrect response because they lack the conceptual knowledge needed to infer the answer.

A lecturer could therefore pair one carefully selected prompt with brief feedback. For example, after studying the break-even solution, students might explain why fixed costs form the numerator. They could then compare their explanation with a short model response before attempting the next problem.

There is also a case for restraint. Providing a full conceptual explanation and simultaneously asking students to generate the same explanation can be redundant. Some studies in the review found that combinations of support reduced rather than improved learning. The practical question is not “How many helpful features can I add?” but “What thinking do I want students to do for themselves?”

Withdraw the worked steps deliberately

If students only study complete solutions, they never have to decide what to do next. The review therefore considers faded examples, in which support is removed progressively, and completion problems, in which learners supply selected missing steps.

The break-even example could become a short sequence:

• Example 1: a complete solution with subgoal labels.
• Example 2: the contribution calculation is supplied, but students determine the amount to recover and calculate the required units.
• Example 3: only the subgoal labels remain.
• Problem 4: students solve a changed task independently.

For the changed task, retain the same contribution-volume-profit assumptions and one-year period, but suppose total fixed operating costs are $72,000, target operating profit is $18,000, selling price is $50 per unit and variable cost is $32 per unit. Students must recognize that the required contribution for the year is now $90,000 and that contribution per unit sold is $18. The required annual sales volume is therefore 5,000 units.

This is more than a numerical substitution because one element of the problem structure has changed. It is still a modest transfer target, however. It does not demonstrate that students could analyse an unfamiliar commercial case, interpret ambiguous evidence or exercise professional judgement.

Findings on fading are mixed, particularly for more distant transfer. The pace appears to matter. Novices may need a slower withdrawal of support, while students with stronger prior knowledge may find fully worked steps redundant. A fixed sequence for everyone can therefore be less useful than checking performance and adjusting the next task accordingly.

A short diagnostic question can help. Before assigning the partially completed example, ask students to state what contribution must recover at break-even. If many cannot answer, the class may need another complete example or a clearer explanation. If most answer accurately, retaining every step may simply delay independent practice.

Assess the change you actually care about

Redesigning the example is only half the task. Lecturers also need to specify what kind of transfer they want to observe.

A test that changes only the figures mainly assesses whether students can repeat a familiar method. A question that adds a target profit requires some adaptation. A case that changes the cost structure, includes irrelevant information or asks students to select among possible approaches creates a more demanding transfer task.

These are not interchangeable outcomes. Before editing a worked solution, it helps to write the intended changed task first. What should students be able to recognize, select or adapt when the familiar cues are removed? The answer should determine which principles and subgoals the example makes visible.

The review cannot tell accounting lecturers which design feature will produce the largest improvement. It is a qualitative synthesis rather than a meta-analysis, so it provides no pooled effect size or quantitative ranking of the different approaches. Transfer was also measured in different ways across the studies, making simple comparisons difficult.

There are important limits on how far the findings can be carried into practice. Most studies were conducted in tightly controlled settings, so their applicability to authentic accounting classrooms and workplaces remains uncertain. Long-term and workplace transfer also remain underexplored. In addition, the original accounting study and the review's supplementary study catalogue were not available for this article, so study-level details such as samples, comparison conditions and transfer measures could not be independently checked.

Those limitations argue for a manageable local trial. Choose one model answer from a topic where students commonly succeed in class but struggle when the wording changes. Define the changed problem you care about. Add one feature, perhaps meaningful subgoal labels or a supported self-explanation prompt, then move to a partially completed example. Compare performance with the materials you currently use.

A worked answer is valuable because it makes expert performance visible. The next design challenge is to make the reasoning portable. That requires more than showing students where the numbers go. It means helping them see what each part of the solution is trying to achieve, then giving them a carefully supported opportunity to carry that structure into a different problem.